Delft University of Technology Personalized gamification to … · 2020-03-11 · Personalized gamification to enhance implementation of eHealth therapy in youth mental healthcare

Delft University of Technology

Personalized gamification to enhance implementation of eHealth therapy in youth mentalhealthcare

van Dooren, Marierose

DOI10.4233/uuid:9a8d3973-f5b5-4812-97ed-27e5c14afc34Publication date2020Document VersionFinal published versionCitation (APA)van Dooren, M. (2020). Personalized gamification to enhance implementation of eHealth therapy in youthmental healthcare. https://doi.org/10.4233/uuid:9a8d3973-f5b5-4812-97ed-27e5c14afc34

Important noteTo cite this publication, please use the final published version (if applicable).Please check the document version above.

CopyrightOther than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consentof the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Takedown policyPlease contact us and provide details if you believe this document breaches copyrights.We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.For technical reasons the number of authors shown on this cover page is limited to a maximum of 10.

https://doi.org/10.4233/uuid:9a8d3973-f5b5-4812-97ed-27e5c14afc34

https://doi.org/10.4233/uuid:9a8d3973-f5b5-4812-97ed-27e5c14afc34

PersonalizedgamificationtoenhanceimplementationofeHealththerapyinyouthmentalhealthcare

Proefschrift

terverkrijgingvandegraadvandoctor

aandeTechnischeUniversiteitDelft,

opgezagvandeRectorMagnificus,prof.dr.ir.Prof.dr.ir.T.H.J.J.vanderHagen

voorzittervanhetCollegevoorPromoties

inhetopenbaarteverdedigenop

vrijdag3april,2020om10.00uur

door

MarieroseMargotMabelleVANDOOREN

MasterofScienceUniversiteitLeiden,Nederland

geborenteBreda

2

Ditproefschriftisgoedgekeurddoordepromotorenencopromotor.

Samenstellingpromotiecommissiebestaatuit:RectorMagnificus VoorzitterProf.dr.irR.H.M.Goossens TechnischeUniversiteitDelft,

promotorProf.dr.V.M.Hendriks ParnassiaGroep;LeidsUniversitair

MedischCentrum,promotorDr.V.T.Visch TechnischeUniversiteitDelft,

copromotor

Onafhankelijkeleden:Prof.dr.H.deRidderDr.M.vanOrdenProf.dr.B.A.M.SchoutenProf. dr. ir. G.D.S. Ludden Prof. dr. ir. D.J. van Eijk

TechnischeUniversiteitDelft ParnassiaGroep,DenHaag TechnischeUniversiteitEindhovenUniversiteitTwenteTechnischeUniversiteitDelft,reservelid

MarieroseM. [email protected]

Coverby:MargotM.M.deBeerPrintedby:ProefschriftMaken

©MarieroseM. M.vanDooren,2020Allrightsreserved.Nopartofthisbookmaybereproducedortransmittedinanyformorbyanymeanswithoutpermissionoftheauthor.

3

CONTENTS1. 7

INTRODUCTION 7

1.1. YOUTH MENTAL HEALTHCARE 8

1.2. DESIGN TO IMPROVE PSYCHOLOGICAL TREATMENT 9

1.2.1. Blended eHealth 9

1.2.2. Personalization in design for mental healthcare 11

1.2.3. Gamification design in mental healthcare 12

1.2.4. Personalized gamification in eHealth for adolescents 14

2. 17

PERSONALIZATION PROCESS IN GAME DESIGN FOR HEALTHCARE 17

2.1. INTRODUCTION 18

2.1.1. Different definitions of personalization 19

2.1.2. Games for health 22

2.2. METHODS 23

2.3. RESULTS OF THE LITERATURE REVIEW 24

2.3.1. A general overview of papers involved in the Personalized Design Process 25

2.3.2. Overview healthcare effects 30

2.3.3. Combining the healthcare effects with the Personalized Design Process 32

2.3.4. Validating the influence of games involved in Personalized Design Process 41

2.3.5. Applied game‐elements in reviewed papers 43

2.4. DISCUSSION AND CONCLUSION 45

2.5. FUTURE WORK AND CONCLUSIONS 50

3. 53

THERAPY PROTOCOLS AND EHEALTH DESIGN: A FOCUS GROUP STUDY 53


3.2. METHOD 55

3.3.1. Therapy protocol 55

3.3.2. Procedure 55

3.3.3. Participants 56

3.3.4. Data analysis 57

3.3. RESULTS 58

3.3.1. Part 1: Focus group sessions with therapists 58

4

3.3.2. Part 1: Focus group sessions with patients 61

3.3.3. Part 2: Focus group sessions with both patients and therapists 64

3.4. DISCUSSION 67

3.5. CONCLUSION 72

4A. 75

GAME DESIGN RELEVANCE OF PERSONALIZATION IN YOUTH MENTAL HEALTHCARE 75

4A.1. INTRODUCTION 76

4A.2. METHODS 76

4A.3. RESULTS 77

4A.4. DISCUSSION 79

4B. 81

THE DESIGN AND APPLICATION OF GAME REWARDS IN YOUTH ADDICTION CARE 81

4B.1. INTRODUCTION 82

4B.2. METHOD 85

4B.2.1. Ethics 85

4B.2.2. Participants 85

4B.2.3. Design 86

4B.2.4. Variables and measures 89

4B.2.5. Procedure 89

4B.3. RESULTS 89

4B.3.1. Strategy of analysis 89

4B.3.2. Manipulation checks 90

4B.3.3. Difference in playing time according to reward types between substance

dependent and non‐dependent adolescents 90

4B.3.4. Difference in reward evaluation according to reward types between substance

dependent and non‐dependent adolescents 91

4B.3.5. General results 91

4B.4. DISCUSSION 92

4B.5. REWARDS IN PERSUASIVE GAME DESIGN: IMPLICATIONS 94

4B.6. REWARDS IN PERSUASIVE GAME DESIGN: CASE STUDY 98

4B.7. CONCLUSIONS AND FUTURE RESEARCH 99

5. 103

REFLECTIONS ON THE DESIGN, IMPLEMENTATION, AND ADOPTION OF A GAMIFIED EHEALTH

APPLICATION IN YOUTH MENTAL HEALTHCARE 103


5

5.2. RESEARCH THROUGH DESIGN: THE READYSETGOALS AS A DESIGN CASE STUDY FOR GAMIFICATION OF COGNITIVE

BEHAVIORAL THERAPY 106

5.2.1. Define Transfer Effect 108

5.2.2. Explore User Context 109

5.2.3. Design Gamification 110

5.2.4. Evaluation Gamification 113

5.3. PRELIMINARY TESTING WITH TARGET AUDIENCE 115

5.3.1. Results and suggestions for iterative improvement 116

5.4. FINAL ITERATION OF THE READYSETGOALS GAMIFICATION 118

5.4.3. The readysetgoals gamification 118

5.5. IMPLEMENTATION OF THE LUCA APP AND EVALUATION BY THERAPISTS 124

5.5.1. Qualitative mid‐stage evaluation 125

5.5.2. Ethical considerations 125

5.5.3. Results 126

5.7. DISCUSSION: REFLECTION AND LEARNINGS FOR EHEALTH DESIGN 134

5.7.1. Managing stakeholder expectations through framing 134

5.7.2. Integrating therapeutic aspects in a game world experience 136

5.7.3. The value of personalization in youth mental healthcare gamification 137

5.8. CONCLUSION 138

6. 143

GENERAL DISCUSSION AND IMPLICATIONS 143

6.1. IMPLICATIONS 147

6.1.1. Implications of personalized gamification in eHealth 147

6.1.2. Implications for the set‐up of future eHealth evaluation methods within this field 154

6.1.3. Implications and recommendations for future research and eHealth development156

REFERENCES 161

SUMMARY 189

SAMENVATTING 195

ABOUT THE AUTHOR 205

LIST OF PUBLICATIONS 206

6

7

1. INTRODUCTION

Chapter1|Introduction

8

1.1. YOUTH MENTAL HEALTHCARE Mental health disorders are the leading cause of disability in children andadolescents. Around 30% of children and adolescents suffer from a mentalhealthdisorder[1]and70%ofmentaldisordershavetheironsetpriortotheage of 25 [2]. Adolescence is a period in life during which essentialdevelopments occur in biological, psychological, emotional, cognitive, andsocial domains [3‐7]. Mental health disorders, including substance usedisorders, in adolescence have a negative impact on these domains [8‐12]during adolescence and adulthood. For example, mental health disordersduring adolescence increase the risk of educational underachievement andlatermentalhealthdisordersduringadulthood[8,9].Toreducementalhealthproblems and limit their negative effects, prevention, early recognition andeffectivetreatmentareneeded.

With a few exceptions, the majority of evidence based therapies foradolescents with mental disorders include psychotherapy which can bedefinedasa treatmentmodality “inwhich the therapistandpatient(s)worktogether to ameliorate psychopathologic conditions and functionalimpairment through focuson the therapeutic relationship” [13]. From thesetherapies,family‐basedtreatmentandcognitivebehavioraltherapy(CBT)aremost often used and have shown to be effective in reducing mental healthproblemsinchildrenandadolescentsforarangeofdisorderssuchasanxietyanddepression[14‐16].

Although psychosocial therapies are effective in reducing psychiatricsymptoms in adolescents with mental disorders, there is still room forimprovement. For example, a recent meta‐analysis of more than 400randomized controlled trials on children and adolescents receivingpsychological therapies found a mean post‐treatment effect‐size of 0.46(“mediumeffect”),andthismeaneffect‐sizedroppedto0.36(“smalleffect”)atanaverageofoneyear follow‐up.Highesteffect‐sizeswerefoundintreatingchildren and adolescents with an anxiety disorder (mean 0.61), and lowesteffect‐sizes were found among those with multiple disorders (mean 0.15)[17]. Hence, a considerable proportion of treatment‐seeking children and


9

adolescents do not (sufficiently) benefit from treatment, and it is largelyunknown which individual patients benefit most from which type(s) oftreatment.

Factors that reduce the effect of mental health therapy include prematureterminationoftreatment,poorattendanceoftreatment‐sessionsandalowornon‐adherence to homework assignments (e.g. [18‐22]). For example, about28‐75% of children and adolescents in mental healthcare drop out oftreatment [23]. Drop‐out and poor attendance can limit the amount of timethat a patient actually spends “in therapy”, consequently decreasing theimpact of the therapy on the patient’s functioning in everyday. Given thesuboptimal effectiveness of psychological treatment, when focusing on theaforementioned examples of therapy adherence, there is probably room forimprovement in the design of therapy. One area from which designmodifications can be derived is the field of new Information andCommunicationTechnologies.Apartfromthepatientsthemselves,twotargetgroups of stakeholders can be distinguished which should be involved inimproving the ‘design of therapy’, and which are likely to benefit from theworkinthisthesis:designresearchers(i.e.healthcaredesignresearchersandgame design researchers) and mental healthcare professionals (i.e. therapydevelopers and therapists). How improvements can be achieved, will bediscussedinthefollowingparagraphs.

1.2. DESIGN TO IMPROVE PSYCHOLOGICAL

TREATMENT

1.2.1. BLENDEDEHEALTHOne great potential of improving psychological treatment is the use ofInformation and Communication Technologies in the delivery of mentalhealthcare [24‐26]. Combining these technologies with current face‐to‐facepsychological therapy with a therapist is also called “blended eHealth”.Blended eHealth can extend the reach of psychological therapy beyond theclinicalsetting,astechnologiescanbeusedanytimeandanywhere(e.g. [27,28]inadults).Itisespeciallysuitableforadolescents,astheyaretypicallythe


10

earlyadaptersofnewtechnologies andaround99%of theyoungstersandadolescentsaged12‐25yearsoldownasmartphone[29].

Patientsoftendonotworkontheirtherapyoutsideatherapyroom,andthusdo not practice thatwhat they learn during face‐to‐face therapy sessions intheir daily life (e.g. [18]). Blended eHealth has the potential to lower non‐adherence, one of the main reasons for therapy failure and poor therapyoutcomesingeneralmentalhealthcare[19,30,31].Forexample,byenhancingthe patients’ motivation and engagement to go to therapy sessions andresponsibility and possibility to do their homework (e.g. [32‐34], with nospecific focus on adult or youth mental healthcare). Besides, with blendedcare, therapists still have the opportunity to build up a collaborativerelationship–alsoreferredtoasa ‘therapeuticalliance’‐withtheirpatients[33]. According toprevious findings, a strong therapeutic alliancehasbeenassociatedwithpositivetherapeuticoutcomes(e.g.[35]forpsychotherapyingeneraland[36]foradolescentandchildpsychotherapy).

ResearchfocusingontheeffectofeHealthinmentalhealthcareislimited,butexistingmeta‐analysessuggestoverallsmall tomediumeffectsizes inadults(e.g. [37‐39] in reducing substance use or improvements in anxiety anddepression symptoms) and adolescents and children (e.g. [40] in treatinganxiety symptomsand [41] in treatingdepressionandanxiety symptomsordisorders). Moreover, research suggests that blended eHealth is moreeffective regarding mental health symptoms compared with fully onlineeHealthwithout therapist contact (e.g. depressionandanxiety in adults [42,43]). However, research also suggests that blended eHealth is not moreeffectivecomparedtostandardface‐to‐facetherapy[27].Kenteretal.(2015)even suggested that blended eHealth in adult mental healthcare seemed totakemore therapy sessions (face‐to‐face plus online sessions) compared tothosewhoreceivedonly face‐to‐face sessions, resulting inhigher costs [44].Patients do not optimally use the onlinemodules of blended eHealth and amajoritydoesnotcompletetheentiretreatmentprogram(e.g. [45]focusingon adherence of adults and [46] of children and adolescents). Even thougheHealth adherence‐rates have not often been compared with face‐to‐face


11

therapy (e.g. [27, 47]), systematic reviews indicate that non‐adherence ineHealthiscomparabletoface‐to‐facetherapy.

Therefore, the modules of current therapy should either be improved orpatientsshouldbemotivatedtostartandcontinuetousetheonlinemodulesfortherapy‐relatedactivities[48],especiallywhentheyhavetoperformtheseonlinemodulesintheirownenvironmentandtime.Whentheonlinemodulesof blended eHealth are aligned to the patient and his/her context ofapplication,patientscanbemoremotivated tostartandcontinue touse theonlinemodules,consequentlyimprovingthetherapeuticeffects.

1.2.2. PERSONALIZATIONINDESIGNFORMENTALHEALTHCAREToenhancethemotivationofanindividualpatienttostartandcontinuetousea specific product or adhere to therapy, it can be “personalized” toaccommodate the individual characteristics,needsandwishesof individualsorgroupsofindividuals[49].Thiscanbedonebyinvolvingstakeholders(e.g.theusersofaproduct)inthedesignprocess[50],consequentlyenhancingthechance that the product is being used [51‐54]. A lot of research involves“personalization”, but clear and shared concepts of what personalizationentails are lacking. Stakeholders can be involved in different phases of adesignprocess.Atthestartofaprocess,theproblemandfocusforthe‘to‐bedesignedproduct’isidentified,establishedandanalysed.Thisisfollowedbyaphasewherepossiblesolutionsaredeveloped, tested,evaluated,and furtherimproved. Lastly, the process ends in a personalisable product that can betailored to the needs of individual users of the product, derived from theearlierdesignphases.

Current face‐to‐face psychological therapy is often protocolized to structureface‐to‐facetherapysessions.Thisistoensureevidence‐basedpractices,andconsequentlytoincreasethelikelihoodthatthepursuedtreatmentoutcomesareattained[55].Eventhoughitisrecommendedtofollowandapplytherapyprotocols as much as possible, as it plays a large role in the success ofevidence‐based therapies [56], both therapists and patients can have goodreasons to changeorpartlyapplya therapyprotocol in therapeuticpractice[57‐61]. Generally, the therapy protocols that form the basis for the


12

implementationofevidence‐basedface‐to‐facetherapiesarealsousedforthedesign of eHealth [62]. A possible cause for the limited effect of eHealth inmentalhealthcareisamarginallevelofengagementfromusers(e.g.therapistsand patients) during the design process of eHealth. For example, bydigitalizing the full therapyprotocol that isactuallyonlypartiallyappliedorusedintherapeuticpractice.EHealthshouldbepersonalized,asitisexpectedthat personalizationmotivates them to continue to use eHealth [53, 63‐69].Consequently, thiswouldpositively influence the implementationprocessoftheeHealthproductintheindividual’sdailylifeandwiththat,thechancethatthe health related transfer effects are achieved [70, 71]. However,personalization practices are insufficiently described and there is a lack ofsystematic studies on the added value of personalization. Therefore, it isimportant to examine the conditions for a successful implementation ofpersonalizationineHealthforclinicalpractice.

1.2.3. GAMIFICATIONDESIGNINMENTALHEALTHCARECurrenteHealthinterventionsinmentalhealthcareareoftenfocusedonthetherapeuticcontentandprovidelimitedinteractionmotivationforthepatient.A design technique that aims to enhance the motivation of patients to useeHealthbymakingitmoreappealingistheapplicationofgame‐elementsfromentertainment games. Game‐elements such as rewards, challenge andcompetition, generate engaging experiences such as pleasure, and surprise[72] or feelings of flow, a rewarding state of pleasure users can havewhenplaying a game that matches their skills [73]. These experiences in turndirectly fulfil basic motivational behavioral needs [74]: the need forcompetence, autonomy, and social relatedness [51, 75]. However, theexperiencescanbemoreandlesspreferredexperiencesbyusers,dependingon the users’ intrinsic needs, values and goals. Preferred experiences canimprove their satisfaction [76] and increase usage frequency. If thegamification design is thus personalized to the users, this can enhance theengagement and motivation of the user to interact with the gamificationdesignevenmore[68]andconsequently improvethe implementationof thegamifiedproduct.


13

Theuseofgame‐elementsfromentertainmentgamesinnon‐gamecontextsislabelled as “gamification design”. Gamification design aims to change thebehaviorofauserintherealworldbycreatingagameworldexperience[63]that is more engaging, free and enjoyable [63] compared to a real worldexperience (see Figure 1). Gamification design in healthcare and mentalhealthcare has shown potential [71, 77, 78], e.g. by improving healthybehavior, well‐being, and/or positively influencing the knowledge andattitudeofindividualstowardshealthybehavior[78‐89].Gamificationdesignseems especially relevant for youth mental healthcare, as millions ofadolescentsplaycomputergamesasaleisureactivity[90].Therefore,onecanassume that the motivating and rewarding experiences of games are moreimbeddedinthelivesofadolescentscomparedtothelivesofadults.Arecentstudy of Deacon and O’Farrell (2016) focused on serious games andgamificationsforadolescentswithchronicdiseasesandfoundpositiveresults,especiallyforbehavioralinterventionsthatpromotedself‐carebehaviors[91].However,moreresearchisneededtostudyhowgamificationcanbedesignedin the most effective way before implementing it in practice. For example,there are only a few independent trials and direct comparisons betweengamifiedandnon‐gamifiedinterventionsarelacking[91,92].

Figure 1. PersuasiveGameDesign(PGD)modelofVischetal.[63]


14

1.2.4. PERSONALIZED GAMIFICATION IN EHEALTH FOR

ADOLESCENTSEven though gamified eHealth seems successful in mental healthcare andyouth mental healthcare contexts [71, 77, 78], there is a lack of validationresearch on the added effect of gamification and crucial aspects for asuccessful implementationofgamifiedeHealthinterventions.Personalizationhasbeensuggestedasadesigntechniqueforasuccessfulimplementation,butitisunclearhowithasbeenappliedandwhattheeffectsareonhealthrelatedoutcomes.Therefore,theaimofthisdissertationistostudytheaddedvalueofpersonalizedgamificationasafactortoenhance implementationpotentialofeHealthinterventionsinyouthmentalhealthcare.Asnotedinthefirstsectionofthischapter,apartfromthepatients,theresultsofthisthesisarerelevantfor two groups of primary stakeholders: design researchers and mentalhealthcareprofessionals.

Designresearchers

Personalization is often applied by design researchers, since it has beensuggestedasadesigntechniqueforasuccessfulimplementationofthedesign.However, it is unclear how personalization has been applied and what theeffects are on health‐related outcomes. This dissertation can help designresearcherstoknowhowtheycanpersonalizeagamifiedeHealthapplication,whichenhancesthechancethattheproductmatchesthetherapeuticpractice.In this dissertation, we propose a uniform definition of personalization ingame design to execute a literature study on how personalization has beenapplied in game design for health (Chapter 2). This uniform definition ofpersonalizationwillhelpdesign researchers tostructure thepersonalizationprocesses of their designs. Other information that is important for designresearchers is thealignmentofadesigntoyouthmentalhealthcare.Wewillstudyhowadesigncanbealignedtoyouthmentalhealthcarebyinvestigatingthe amount of and reasons for therapy protocol application (Chapter 3). Inaddition,we focuson thegamedesignrelevanceofpersonalization inyouthmental healthcare. Firstly, by describing a game design method using aspecific personalization technique (Chapter 4A), followed by the potential


15

implementation of a game element (i.e. rewards as one of the most oftenappliedgameelement)(Chapter4B)inyouthaddictioncare.Thisinformationisrelevantfordesignresearchers,sincetheyneedtoknowifthisgamedesignmethod is suitable to usewithin this context, or if there are other/multipledesignmethodsneededinthedesignprocess.Inaddition,theyneedtoknowif the game element is suitable to apply in youth addiction carewhen theywant to use game elements to motivate patients to engage and remain intreatment.Thisdissertation concludeswith thedescriptionof an exemplarydesign implementation case, and describes our learnings for designresearchers(Chapter5).

Healthcareprofessionals

Nexttodesignresearchers,thestudiesdescribedinthisdissertationarealsorelevant formentalhealthcareprofessionals.Theamountofandreasons fortherapy protocol application (Chapter 3) can be used by healthcareprofessionals to improve therapeutic practice, e.g. by updating therapyguidelines, providing training and/ormore supervision to ensure evidence‐based therapeutic practice. In addition, they can help design researchers tobetteraligneHealthtobothevidence‐basedtherapyprotocolsandtherapeuticapplication of these therapy protocols. Besides, we provide healthcareprofessionalswithinformationregardingthepotentialusageofrewardsasamotivational element in an addiction treatment context, since substancedependent youngsters may be less motivated by non‐drug‐related rewardsdue to differences in the motivational system (Chapter 4B). Lastly, thisdissertation provides case‐study driven learnings for the development andimplementationofgamifiedeHealthwithinyouthmentalhealthcare(Chapter5).

16

Basedon:vanDooren,M.M.M.,Visch,V.T.,Spijkerman,R.,Goossens,R.H.M.,&Hendriks,V.M.(2016).PersonalizationinGameDesignforHealthcare:aliteraturereviewonitsdefinitionsandeffects.InternationalJournalofSeriousGames,3(4),3‐28.

2. PERSONALIZATIONPROCESSINGAME

DESIGNFORHEALTHCARE

Stakeholders have increasingly been involved in game design, to enhance thealignmentofagametotheend‐user. Inahealthcarecontext,thisalignment isexpected to enhance the end‐user’s motivation to interact with the game,therebyenhancingthegames’healthrelatedtransfereffect.However,thenatureand effectof this involvementhaveneverbeen systematically studied,makingassumptions regarding the benefits of personalization ungrounded. In thisliterature study,weaim toprovide1)anoverviewof existingpersonalizationdesign theory and description of our Personalized Design Process (PDP),consisting of the phases Problem Definition‐, Product Design‐ and TailoringPhase, and 2) a systematic review on the applications of the PDP phases inempiricalstudiesandeffectsacrossthesephases.

Chapter2|Personalizationingamedesignforhealthcare

18

2.1. INTRODUCTIONGames are designed to motivate end‐users to play. Especially in serious‐games, that are typically a bit less entertaining than pure entertainmentgames, it is important forthegamedesigntooptimallyengagetheend‐user.Research has suggested that involving stakeholders (like end‐users anddomain experts) in the design process enhances the engagement andmotivation of the user to interact with the product [68] and consequentlyimprovesthegame’simplementationintheuser’sdailylife.Suchstakeholderinvolvement is often called co‐design, where end‐users are enabled toinfluencethedesign[50].

Currently, a lot of games for health are designed that involve“personalization”, but clear and shared concepts of what personalization ingame design entails are lacking. Besides, it is not sure if personalizationcontributes to the targetedhealth‐effectofagame.Since theoryonapplyingpersonalization in game design is lacking, we will use theory frompersonalized design methods and propose a theory on “Personalized GameDesign”.This“PersonalizedGameDesign”(PDP)‐modelwillbeusedtostudyifandhowpersonalization ingamedesign iseffective in thecontextofhealth.BasedonourPDP,weproposetodefinepersonalizationastheinvolvementofstakeholders in at least one of the three PDP phases (Problem Definition‐,Product Design‐ and Tailoring Phase). Stakeholders that can be involvedacrossthephasesofthePDPare:“designers”,“domainexperts”(therapistandcare staff), “end‐users” (typically patients), or “family/relatives” (of thepatient). Some PDP phases are better suited to these four specific types ofstakeholdersthanothers.Forexample,designersanddomainexpertstypicallypartake in the first Problem Definition Phase, by defining the problem andrecommendations for focus of the design [93]. During the Product DesignPhase, all stakeholders can contribute to provide design suggestions andfeedback [94‐96]. Finally, in the TailoringPhase, the end‐users are typicallyinvolved,forinstancebyselectingapersonalavatar[97].


19

2.1.1. DIFFERENTDEFINITIONSOFPERSONALIZATIONCurrently,many definitions are in use for the concept of personalization. Inthis section we first describe these concepts of personalization based ongeneraldesign literature.Thiswillbe followedbyourproposedPDP‐model,consistingofthreephasesinthedesignprocessinwhichpersonalizationcantakeplace.Thesephaseswillbeusedtostructuretheliteratureresults.

User‐centered design process. Defined as: Any act during the designprocesswheretheusercanbeseenasasubjectinsteadofapartner[53].

Inuser‐centereddesign,thefocusliesondesigningforend‐users[98]wherethese end‐users have a passive role. Insights for designing a product aregeneratedthroughinterviews,observationsandtheory.Anexampleofthisisthedesignof anexercisegame forolderadultswithhelp from focusgroupsandusertesting[99].Aproductisnotcreatedtogetherwiththeuser,butheorsheonlyreflectsonanidea,prototype,orisinvolvedintheproduct’sfinalusertest[53].Therefore,theuser’sinfluenceontheproductislimited.

Co‐creationprocess.Defined as:Any act of collective creativityduring[53].

Co‐creation builds on the tradition of user‐centered design. The term ‘co‐creation’ is often used interchangeably with ‘co‐design’, although they havedifferentdefinitions.Co‐creationreferstoatemporaryexchangeofideasandexperiences,andconsistsof“specificpartswithinthedesignprocess”[50].

Co‐designprocess.Definedas:Anyactduringdesigninwhichusersareconsideredasexpertoftheirexperiences.

Co‐creation takes place within a co‐design process, where the end‐user “isgiven thepositionof ‘expertof his/her experience’ andplaysa large role indevelopingknowledge, ideasandconcepts” [53].Thedesigner facilitates theend‐users,sotheyparticipate inawaythat ismostsuitable totheirabilities[53].Wehave adopted the co‐designdefinition ofMattelmäki and SleeswijkVisser (2011), who viewed it as “a process and tools of collaborativeengagement”[50].Thedesignresponsibilityiskepttothedesigners,becausetheyareexpertsindesign.Itshouldbenotedthatco‐designisalsooftencalled


20

participatorydesign, as both concepts enable the end‐users to influence thedesign [50]. However, with co‐design a designer only wants to collaboratewith end‐users [50]. and in participatory design, more weight is placed onend‐userempowerment.

Tailoring.Definedas:Theadaptationof thedesignedproductby itself,bytheend‐usersorbyothers.

If auserexplicitly changesaspectsofaproductdesign, suchas itsesthetics,we propose the term “User Controlled Customization” [54]. End‐users canthuspartlydeterminetheappearanceorfunctionalityofaproduct[100].Ifasystem tailors itself to the user and the behavior of the user, we term this“Use‐DependentAdaptation”[54,101].Inthiscase,theproductchangeswhiletheuser interactswith it, for example, bykeeping thedifficultyof the gamealignedtotheusers’(health)improvements.

Mugge, Schoormans and Schifferstein (2009) found seven options fortailoring. In one option, theMental Effort, users are creatively involved, forexample,ado‐it‐yourselflampthathasametalsheetwhichcanbescratchedto customize it [100]. These dimensions can generate different tailoringoptionsfortheproduct’sdesign,ofwhichsomecanbemoreorlessfavorableforspecifictargetgroups.Therefore, it is importanttounderstandthetargetgroupandtoknowwhichofthesedimensionsaremoreorlessfavorablefortheend‐user.

Personalized Design Process. Defined as: Stakeholder involvement inProblemDefinition,ProductDesignandTailoringPhasesofaproduct.

Commonusageof“tailoring”and“personalization”isoftennon‐consistentandcan therefore be confusing. For example, some studies refer to individualcharacteristics(e.g.,thenameoftheuser)aspersonalization[102,103]orasatailoring variable [104] and some studies that saw personalization as amechanism of tailoring [102, 105]. We aim to avoid this confusion, byreferring to the involvement of stakeholders across the design process as“personalization”.Wetermthiscompleteprocess,asthePersonalizedDesign


21

Process (PDP).Aswill be shown, personalization can takeplace at differentphasesinthePDP.

The PDP consists of three phases: Problem Definition, Product Design andTailoring. In the Problem Definition Phase, information is generated byconsulting stakeholders, in order to identify, establish and analyze theproblemandgeneraterelatedideas.Thissetsthefocusforthe‘to‐bedesigned’product,andcanbeseenasthebasisforthewholedesignprocess[106].Thenext Product Design Phase includes both Ideation and Embodiment. InIdeation, the firstpossiblesolutionsareproduced,resulting inproduct ideasor design proposal(s). In Embodiment, these are tested and evaluated byusers, and further improved through iterations [107]. In the last TailoringPhase, the finalproductcanbetailoredto theneedsof individualend‐users.Tailoringaproductcanbedonebyanend‐user,othersorautomatically(seeFigure1), forexample,inthecaseofthegame’sdifficultylevelautomaticallyadapting to the user’s skill level. In this review we do not differentiatebetweenIdeationandEmbodimentoftheProductDesignPhasesbecausebothconsider the actual physical design of a product. The PDP thus consists ofdifferentphasesinwhichstakeholderscanbeinvolved.ThelastphaseistheTailoring Phase, which consists of two types of Tailoring: “User ControlledCustomization”and“Use‐DependedAdaptation”.

Although our model shows considerable overlap with earlier models thatdescribe co‐design processes and include stakeholder involvement in theProblemDefinition‐ and Product Design Phase, it differswith regard to theTailoringPhase,whichisnotpresentintheseearliermodels(e.g.,[53,108]).For example, comparing our PDPwith the process previously conceived byZebeko and Tan [108], there is a large overlap between our ProblemDefinition and their Diagnostic phase, where information about anorganizationorcommunity is collected, inorder tounderstand thesituationand challenges [108]. There is also an overlap between our Product DesignPhase and their Design and Develop and Test phases, where the mostappropriatestakeholdersdevelopandprototypetogether[108].However,ourPDPgoesfurther,byincludingaTailoringPhase,toensurethatproductsarealigned to individual end‐users within a target group. This is important,


22

because, even in a coherent target group, there are always individualdifferences that need to be taken into account when designing a suitableproduct.

Figure 2. ThePhasesofthePersonalizedDesignProcess

2.1.2. GAMESFORHEALTHGamesaredesignedtobeenjoyableandimmersive,andcanhelptomotivateorpersuadeend‐userstocontinuingplayingthegame[63].Gamescanalsobeused to facilitate the realization of health‐oriented goals of the user (e.g.,[109]).Amainadvantageofthesekindsofgame‐interventionsisthattheyarealwaysavailable,comparedtoface‐to‐faceinterventions,andofteneffectiveinsupportinghealthrelatedchangesofbehaviors[110‐112].

Whendesigninggameswithstakeholders,thealignmentofthegamewiththeend‐user’s preferences, needs and competences can increase [53, 64, 65],whichinturncanmotivatetheend‐usertointeractwiththeproduct[68,69],thereby enhancing the persuasive feature of a product [70]. This is becausestakeholderswithdifferentexpertise(e.g., indesign,thehealthcontext,orintheir own preferences) have different point of views and can providemorecomplete insights into what the product should consists of and focus on.Stakeholder involvement in the design process of games is more likely to


23

generateahealthrelatedtransfereffect.Healthrelatedtransfereffectsaretheeffects a product is aiming to achieve, for example: effects on knowledge,mood,health,etc.Studieshavefocusedonenhancingthesetransfereffects,byinvolvingstakeholderinthedesignprocess[113,114].However,theeffectofstakeholderinvolvementwhendesigninggamesforhealthacrossthedifferentphasesofthePDPhasnotyetbeenstudiedinasystematicway.Therefore,thisstudyaimstoanswerthefollowingresearchquestion:HowarePersonalizedDesignApproachesapplied indesigninggames forhealth, andhoweffectivearetheyonhealthrelatedoutcomes?”

2.2. METHODSIn order to answer the research question, we conducted a literature study,withtheaimofcategorizingthedesignmethodsusedinpublishedempiricalstudiesbasedonstakeholderinvolvement,asshowninFigure1.Wesearchedthefollowingdatabases:IEEInspec,ELSEVIERScopus,Psychinfo,PubMedandWeb of science. Keywords that served as basis for the search terms weredivided into four groups: methodology, object, context of appliance andresearch method (see Table 1). Only empirical studies were included; thefollowing types of articles were excluded: book reviews, technical studies,theoreticalstudies,reflections,reviews,withdrawnarticles,editorials,studieswithafocusonalgorithmsandarticlesnotrelatedtohealth.Wefirstscreenedthe abstracts and titles in order to deselect studies based on the exclusioncriteria. The remaining articles were then scanned based on theinclusion/exclusion criteria in order to make a second selection. Lastly, wecarefullyanalyzedthefulltextsoftheremainingarticles.

Table1.Researchkeywords,dividedinfourgroups.A

MethodologyBObject(Games)

CContextofappliance(health) DResearchmethod

Co‐creat* Game* Therapy Behavior Experiment*

cocreat* Gami* Disease Illness Random*

Customi* Persuasive Health Wellbeing RCT

Co‐design* Care Hospital Evidence*

Participatory Clinical Training Trial*


24

Codesign* Disorder Therapists Empirical*

Collaborat* Patient Life

Co‐develop* Medical Healthstatus

Codevelop* Psychology Fitness

Co‐product* Rehabilitation Physical

Coproduct* Physiologic* Diseasecourse

Personalize* Lifestyle Healthattitudes

Personalization HealthKnowledge Psychological

Personalise* Medicine (behavio*)

Personalisation Telemedicine

self‐creat* Treatment

Self‐made diseasemanagement

Self‐product*

2.3. RESULTSOFTHELITERATUREREVIEWWeretrievedliteraturefromtheabovementioneddatabasesusingthesearchterms in Table 1 from the start of electronic records until April 2015. Thisresulted in a total of 2579papers: 705 studies fromWebof Science, 497 ofINSPEC, 704of SCOPUS, 326ofPsychinfo and347ofPUBMED.Of the 2579papers,62wereselectedtodeterminehowpersonalizationapproacheswereadopted in research on game interventions for changing health relatedbehavior. To answer the research question (How are Personalized DesignApproachesappliedindesigninggamesforhealth,andhoweffectivearetheyon health related outcomes?), we investigated in what way the reviewedstudiesinvolvedthefourstakeholders(designers,domainexperts,end‐users,andfamily/relativesoftheend‐users)intheirdesignprocess.Becausetheirinvolvement occurred in different PDP phases in the design process, wepresenttheircombinations.WefirstdescribestakeholderinvolvementintheProblemDefinition‐,ProductDesign‐,andTailoringPhaseseparately,followedbythecluster‐combinationofstakeholderinvolvementacrossthePDPphases.In 3.1 we discuss the stakeholder involvement, followed by the generalhealthcareandproducteffects insection3.2.This isspecifiedtostakeholder


25

involvement across the PDP in section 3.3, which ends in a conclusionregardingthequalityofthevalidationpapersin3.3.7.Theclosingsection3.4describestheinvolvedgame‐elementsacrossthePDP.

2.3.1. AGENERALOVERVIEWOFPAPERSINVOLVEDINTHE

PERSONALIZEDDESIGNPROCESS Inthissection,wedescribestakeholderinvolvementinthedifferentphasesofthePDP(ProblemDefinition,ProductDesign(bothIdeationandEmbodiment)andTailoring).

Problem Definition

Table 2. StakeholderinvolvementinProblemDefinitionPhase

ProblemDefinition

D X U

[115] X X

[116] X X

[93] X X

Footnote:D=Designer;X=Domainexperts;U=End‐user

Threeofthe62studiesinvolvedstakeholders(designers,domainexpertsandend‐users) only in the Problem Definition Phase, of which two studiesinvolved both designers and end‐users, but not domain experts [115, 116],andoneinvolveddesignersanddomainexperts,butnoend‐users[93].

Product Design

Table 3. StakeholderinvolvementinProductDesignPhase

ProductDesign

Ideation Embodiment

D X U F D X U

[117] X X

[118] X

[119] X X X X X X FX


26

[120] X X X

[94] X X X X

Footnote:D=Designer;X=Domainexperts;U=End‐user;F=Family

FivestudiesinvolvedstakeholdersintheProductDesignPhase.Threeofthemonly involvedstakeholders in Ideationof theProductDesignPhase,whereaproductwas generated based on their comments, suggestions or guidelines[117,118,120].TwootherstudiesinvolvedstakeholdersinbothIdeationandEmbodiment of the Product Design Phase [94, 119]. In the first study,designerswereonlyinvolvedinIdeation,whereotherstakeholdersprovidedsuggestions for improvement [119]. In the other study, components of aproductwereextensivelypretested,andaftertheproductwasinstalleditwasalsopreviewedbyothers[94].

Combining Problem Definition and Product Design

Table 4. StakeholderinvolvementinbothProblemDefinition‐andProductDesignPhase

Footnote:D=Designer;X=Domainexperts;U=End‐user;F=Family

ProblemDefinition

ProductDesign

Ideation Embodiment

D X U F D X U F D X U

[121] X X X X X

[122] X X X X X X X

[123] X X X X X

[124] X X X X X X

[125] X X X X X X

[95] X X X X

[96] X X X X X

[126] X X X X X X X X X X


27

Of the 62 studies, eight studies involved domain experts during both theProblemDefinition‐andProductDesignPhase.Fourstudies involvedmainlydesignersanddomainexpertsasstakeholdersintheProblemDefinitionPhaseanddesigners,domainexpertsandend‐users inEmbodimentof theProductDesignPhase[121,123‐125],forexample,byobservingend‐usersandgivingculturalprobes(ambiguousstimuliandassignmentsthatbringinspirationtodesign) to domain experts and relatives [125]. Two other studies includedstakeholders,mainlydesignersanddomainexperts,intheProblemDefinitionPhaseandIdeationoftheProductDesignPhase[95,96],bylettingend‐userstest game scenarios that were designed by domain experts and designers.Lastly, two studies involved stakeholders in Problem Definition Phase andIdeationandEmbodimentofProductDesignPhase[122,126].

Tailoring

Table 5. StakeholderinvolvementinTailoringPhase Tailoring

UserControlledCustomization

Use‐DependentAdaptation

Task

Virtualself

Task Virtual/feedback/textual

S 3 I R X

[127] X

[128] X

[129] X

[130] X

[97] X X

[131] X

[132] X X

[133] X

[134] X

[135] X

[136] X


28

[137] X X

[138] X X

[139] X

[140] X

[141] X

[142] X

[143] X

[144] X X

[145] X X

[146] X X

[147] X XX

[148] X X

[149] X X

[150] X X X

[151] X X X

[152] X X

[153] X X

[154] X X X

[155] X X

Footnote;I=ideal;R=real/realistic;x=ideal/real;S=self;3=thirdperson;

With “Use‐Dependent Adaptation”, a Kinect device was often used to givevisual tailored feedback about the performance or remaining time the end‐userhad[138‐144]and/orbytailoringthedifficultyofthetaskstoend‐userinput, like performance [135‐137]. In “User Controlled Customization”, end‐userstailoredavatarsthatcouldrepresentanidealizedselforactualself[97,127, 129, 132, 134] and others defined the objectives, difficulty level orspecifiedthestimuliofaproduct[128,130,131,133].

Eleven studies involved both Tailoring types,where the end‐users providedinputused incombinationwithgivingTailoredvisualperformance feedback[150,151],givingfeedbackbasedonthenameofanend‐user[145,146],orby


29

adapting the difficulty level based on information provided by a userbeforehand,e.g.,throughbaselinemeasurements[147‐149].However,others(mostlydomainexperts)couldalsoTailor,bydefiningobjectivesanddifficultylevels, which was combined with giving automatic feedback aboutperformance[153‐155].

Combinations of Problem Definition, Product Design and Tailoring

Table 6. StakeholderinvolvementinProblemDefinition,ProductDesignandTailoringPhase

ProblemDefinition

ProductDesign Tailoring

User ControlledCustomization

Use‐DependentAdaptation

Ideation Embodiment

Task Virtualself

D U F D X U F D X U X S 3 I R X T F

[156] X X X X X

[157] X X X X X

[158] X X X X X

[159] X X X X X X

[160] X X X X

[161] X X X X X

[162] X X X X X X X

[163] X X X X X

[164] X X X X

[165] X X 3

[166] X X X X X X

[167] X X X X X

[168] X X X

[169] X X X X

[170] X X X X X X

[171] X X X

Footnote:D=Designer;X=Domainexperts;U=End‐user;F=Family;A=automatic;S=self;3=thirdperson;x=ideal/real;T=task,f=feedback


30

Ofthe62studies,seveninvolvedstakeholdersintheEmbodimentofProductDesign‐ and Tailoring Phase. The studies were mainly conducted in thedomainsofrehabilitationandphysicalhealth[164,167,168,170,171].Inonestudy, end‐users with Autism Spectrum Disorders (ASD) played a therapygamewitharobot,ofwhichthebehavioralandexperienceresultswereusedtoimprovethenexttheexperimentbydomainandroboticexperts[166].Fourstudies involved stakeholders (mainly end‐users and/or domain experts) inIdeation of the Product Design Phase and Tailoring Phase. One study usedearlier guidelines thatwere combinedwith interviews and evaluationswithend‐users[160].InvolvingstakeholdersinboththeTailoring‐andtheProductDesign Phase was thus quite common, however involving end‐users inIdeationwas leastcommonlycombinedwith theTailoringPhase.Only threestudies involved stakeholders in all the phases; these focused on physicalhealth[161],mentalhealthcare[162,163],and/orwhereparentsordomainexpertscouldtailorthetool[161,162].

2.3.2. OVERVIEWHEALTHCAREEFFECTSStudies involving stakeholders in the PDP havemainly focused on teachingend‐users about health related topics [96, 115‐119, 122, 126], aiming atbehavioralchangeoradherence[93,95,97,121,123‐125,128,131,132,135,137‐141,143‐145,147‐157,160,161,164‐171],oratattitudinalchange[94,120, 127, 129, 130, 133, 134, 136, 142, 146, 158, 159, 162, 163]. Currently,researchers are optimistic that personalized games in a health context willgenerateapositive influenceon interactionexperience, interactionbehaviorand health related transfer effects. Interaction experience focusses on thesubjectiveexperienceindividualshavewheninteractingwithaproduct[172,173]andconsistsofexperiencesregardingaesthetics,meaningandemotions[173],forexample,whentheend‐userlikestheappearanceofaproduct.Thefocusofinteractionbehaviorliesonallformsofend‐userbehaviorwhentheend‐user interacts with the product, for example, on forming, altering orreinforcingself‐initiatedbehavior[174].Forexample,ifandhoweasyitistouse the product. Lastly, health related transfer effects are the effects on“forming, altering, or reinforcing user‐compliance, ‐behavior, or –attitude”,andcanberegardedasatransfereffectofgame‐worldrelatedexperiencesto


31

a user’s subsequent behavior in the real‐world (c.f., [63] for a theoreticalmodelof experienceeffects).Health related transfer effects thus reflects theintendedbehavioral change of the end‐user in thedaily life of the end‐user,e.g., the complianceof an end‐user to themedication schedule [175‐177] orenhancing daily physical activity [63]. However, the optimism that theseaspectsarepositivelyinfluenced,isnotsupportedbyagreatdealofevidence.

Theeffectsofthestudies,combinedwithstakeholderinvolvementacrossthePDP are described below in more detail where we evaluate the effects oninteraction experience, interaction behavior and health related transfereffects.Studiesinthisliteraturereviewfocusoneitherone,acombinationof,or all three of these outcome variables and of the 62 reviewed studies, amajority (N = 46) focused on interaction experiences. The following fiveaspectswereusedtoratethemethodologicalqualityofthestudies:pre‐postmeasurement,comparisonorcontrolgroup,(blind)randomization,numberofparticipants and valid and reliable measurements. A higher methodologicalqualitymeansthatatleastacomparisonorcontrolgroupwaspresentinthestudy. Most studies included small samples (25 or less participants), and,hence, had insufficient statistical power to draw firm conclusions about theeffects of involving stakeholders in the PDP. A majority of the studiesgeneratedinformationaboutinteractionexperiences,byusingquestionnaires(N=24),interviews(N=17)orobservations(e.g.,toseetheend‐users’facialexpressionswhileinteractingwiththeproduct)(N=13).Atotalof28studiesfocusedoninteractionbehavior,oftenmeasuredbyobservations(N=15)orby using hardware data derived from the tool itself (N = 14). A total of 40studiesfocusedonhealthrelatedtransfereffects,whichwasoftenassessedbyquestionnairesand tests (N=26)andsometimesbyphysiologicalmeasures(e.g.,heartrate)(N=7).Ingeneral,datawasobtainedatpre‐post[119,123,135, 145, 162] or during and after interactionwith the product [127, 136].Becausethedurationofthestudieswereheterogeneous[128,137,156,166],it is hard to compare these results.Aminority of the studiesused a controlgroup (N= 17), ofwhich eighthad small to average study samples, rangingfrom8to57participants,andfivehadlargestudysamples,rangingfrom95to121participants.Atotalof9studiesrandomlyassignedtheirparticipantsto


32

eitherthecontrolorexperimentalgroupgroup[97,132,140,144,146,148,157,162,169].Whenmeasuringtheeffectofaproduct,experimentalgroupsare comparedwith control groups. Ideally, both groupsare equal except forthe independent variable (e.g., when comparing a product with a tailoredproduct,andtheonlydifferenceisthetailoring).Thiswouldmakeitpossibletodrawconclusionsontheeffectoftheindependentvariable[178].Becausevalidationresearchinthecontextofgamesforhealthis limited,wenotonlytook intoaccount thecontrolgroupsthatreceivedanon‐personalizedgame,butalsotreatmentasusual(e.g.,nogame‐intervention),orcontrolgroupsthatconsistedofotheruser‐groups(e.g.,healthyend‐users[121,125]).

There are many different methods for measuring study quality. As gameresearch is a young and developing domain, we did not use these, sinceapplying a strict index is not appropriate. A minority of the studies in thisreviewinvolvedarandomizedcontrolleddesign,andamajorityofthestudiesusedqualitativemeasurementsincludingasmallsamplesize.Inaddition,theresults were mostly founded on outcomes of questionnaires that were notvalidated.Thismeansthatifaquestionnaireisnotvalidated,itisunclearifitmeasures what it claims to measure. Both qualitative measures and smallsample size indicate a ‘low quality’ ranking of the studies included in thisreview.

2.3.3. COMBININGTHEHEALTHCAREEFFECTSWITHTHE

PERSONALIZEDDESIGNPROCESS This section focusses on the healthcare effects of studies that involvedstakeholders in the PDP. The tables consist of a) Problem Definition‐ orProduct Design Phase only, followed by b) both Problem Definition‐ andProduct Design Phase, c) either User Controlled Customization or Use‐DependentAdaptation of theTailoringPhase, d) both typesof theTailoringPhase,e)combiningProductDesignandTailoringPhases,and lastly f)othercombinationofphases.Nostudyreportedapower‐analysis.


33

Studies which involvedstakeholders only in either the Problem Definition or Product Design Phase

Table 7. CharacteristicstoanalyzethequalityofstudiesinvolvedinProblemDefinitionPhase

Pre‐postmeasurement

Numberofparticipants

Validandreliable

measurementsYes No N Yes No

[115] X 807 X

[116] X 23 X X

[93] X 5 X

Studies that involved stakeholders in the ProblemDefinition Phasewere oflow methodological quality. Results suggested improvements regardingknowledge and awareness about health, more specifically regarding (raw)milk andHIV,which could lead to behavioral changes [115, 116].A gradualneed to collaborate and enhanced social interactionwas found in end‐usersinvolvedincollaborationsessionswithamulti‐touchgame[93],beneficialtothehealthprobleminquestion.

Table 8. CharacteristicstoanalyzethequalityofstudiesinvolvedinProductDesignPhase



Validandreliable

measurementsYes No N Yes No

[117] X 33 X

[118] X 41 X

[119] X 3829 X

[120] X 1 X

[94] X 45 X

Studies that involved stakeholders in theProductDesignPhasewereof lowmethodologicalquality.Results suggestedenhancedknowledgeabouthealth(e.g., aboutAIDS) [118, 119]. A study that focused ondiscussions regardingobesity suggested a doubled discussion time between domain experts, end‐


34

usersand family, and improvedself‐efficacyofdomainexperts indoing this[117]. Other enhancements were found in social interaction andcommunication, combined with less stereotype behavior in a child with anASD [120]. Lastly, feedback from end‐users suggested that a product wasfeasibleandacceptablewithregardtowhatitaimedtoachieve[94].

Studies which involvedstakeholders in both the ProblemDefinition and Product DesignPhases

Table 9. CharacteristicstoanalysethequalityofstudiesinvolvedinProblemDefinition‐andProductDesignPhase


Comparisonorcontrolgroup


Validandreliablemeasurements

Yes No Yes No N Yes No

[121] X X ? X

[122] X X 53&36 X

[123] X X 5 X X

[124] X X 5 X X

[125] X X 10 X X

[96] X X 4 X

[126] X X 165 X

[95] X X X 14 X X

Most studies that involvedstakeholders inboth theProblemDefinition‐andProductDesignPhasehadmethodological limitations.Forexample,onlytwostudies had a control group [121, 122], most studies had a general lownumberofparticipants(5studiesinvolvedbetween4and14)andaminorityof the studies applied validated and reliable measurements, e.g., usedobservationaldata.Resultsofthestudiesshowedthatduetothegames,end‐usersimprovedinvariousoutcomesmeasurements,liketheirphysicalhealth.Examples regarding the improvements that were found on physical health,weresomeimprovementsinshouldermuscleactivity[124]andinmotivatedparticipants that played the game often, of which one even improvedmovements and use of the impaired limb [123]. Other results showed thatend‐users improved their knowledge regarding diabetes [96],


35

cardiopulmonary resuscitation [122], and obesity and nutrition [126].Significant correlations were found between physiological responses tostressful experiences and subjective evaluations on stress in PTSS (Post‐Traumatic Stress Syndrome) patients, and a clear correlation betweendiagnosticPTSDseverityandskinconductanceresponses[95],whichcouldbeimportant for stress inoculation training. End‐users with ASD and healthycontrolsmatchedonIQ,genderandage,showeddifficultiesinrespectingthepersonalspaceofvirtualothers,butacknowledgedthatbehavinginavirtualenvironmentwasdifferentfromdailylife[121].Lastly,inastudywhereend‐usersparticipatedwithbothagameandatraditionalleisureactivityproduct,results suggested that some participants improved social behavior duringsessionswith the game, but that the control productmade the user answermorequestionsinsentencesandhandletheobjectmore[125].

Studies which involvedstakeholders in either User Controlled Customization or Use‐DependentAdaptationof the Tailoring Phase

Table10.CharacteristicstoanalysethequalityofstudiesinvolvedinUserControlledCustomization



(blind)randomization


Validandreliable

measurement Yes No Yes No Yes No N Yes No

[127] X X X 30 X X

[128] X X X X 12 X

[129] X X X ?? X

[130] X X X 40 X

[97] X X X 130 X X

[131] X X X 8 X

[132] X X X 95 X X

[133] X X X >200 X

[134] X X X 2 X

Studies that involved stakeholders in User Controlled Customization hadlimitedmethodologicalweaknesses.Thestudieswereeffectivewithregardto


36

variousoutcomesmeasurements.Onestudyindicatedthatend‐usersshowedphysiological indicators (by levels of skin conductance) of emotions duringgameplay, and that they had the feeling they were part of the game [127].Besides,end‐usersweremoremotivatedtoplay,andexperiencedfeelingsofcompetition and understandability of the product [133], [50], and showedbehaviorsandexperiencesoncooperationandplayability[129].].Feelingsoftogethernessandmentalstimulationwereenhancedinavirtualenvironment[134], as well as a reduced agitation and improved mood during aninterventionwithAlzheimerpatients compared to controls [130].End‐usersthat participated in all conditions had more social behaviors in “enforcedcollaboration” than in “free play” [128]. Studies with control conditionsreportedend‐usersbeingmore“aggressive”afterplayingaviolentgamewithacustomizedavatarcomparedtoanon‐violentgameandgenericavatar[97];they also found that an ideal‐self avatar significantly influenced prevention‐focusedbehaviortokeepthisidealappearanceinreallife,butan“actualself”wasrelated topromotion‐ focusedbehavior [132].Lastly, results suggestedthat controls had significantly higher progression on cognitive functionscomparedtotheexperimentalgroup[131].

Table11.CharacteristicstoanalysethequalityofstudiesinvolvedinUse‐DependentAdaptation





Validandreliable


[135] X X X 2 X X

[136] X X X 6&5 X

[137] X X X X 21&20 X X

[138] X X X X 19 X X

[139] X X X 6 X

[140] X X X 8 X

[141] X X X 6 X

[142] X X X 14 X

[143] X X X 16 X X


37

[144] X X X 37&34 X

Studies that involved stakeholders in Use‐Dependent Adaptation also hadsome methodological flaws and were effective with regard to variousoutcomesmeasurements. For example, two studies suggested that theusershad positive subjective experienceswhile playing the game (e.g., enjoymentandasenseofaccomplishment)andthattheyweremotivatedbythetailoringaspect of the activity [136, 142]. Regarding physical health, the physicalperformanceimprovedsignificantly[143],[60],whichwasasignificant[135]orapercentageimprovementinmotorandsensoryimpairments[139].End‐users that participated in both conditions rated the experimental game asmoreenjoyable[138],andafterplayinganimitativecollaborativegamewitharobot, childrenwithASDplayedmorewitheachother [141]. Studieswithacontrol group found significant improvements in symptoms and balancefunctions,withlongerin‐patientstayinthecontrolcondition[144]andthataproductwasusable,acceptableanditofferedpersonalizedarm‐training[137].A study that only focused on the experimental group, found that amajorityincreased their health awareness, connection with the nurse, but alsoexperiencedusefrustration[140].

Studies which involvedstakeholders in both Use‐Dependent Adaptationand User Controlled Customization of the Tailoring Phase

Table12.CharacteristicstoanalysethequalityofstudiesinbothtypesoftheTailoringPhase





Validandreliable


[145] X X X 23 X X

[146] X X X 57&15 X

[147] X X X 10 X

[148] X X X 155 X

[149] X X X 30 X

[150] X X X X 17 X


38

[151] X X X 19&17 X X

[152] X X X 9 X X

[153] X X X 9or8 X

[154] X X X 15 X

[155] X X X 3 X X

Studies that involved stakeholders in both types of the Tailoring Phase hadsomemethodologicalflaws.Resultssuggestedpositiveresultswithregardtovarious outcomemeasures, like positive experiences while interacting withthe product, sometimes with suggestions for improvement (e.g., worries offallingwhileusingtheproduct)[149,150,153].Physicalhealthwaspositivelyinfluenced (e.g., postural stability) [154, 155]. A study that focused onsmokingcessation,showedthatatfollow‐uponly14.3%oftheend‐usershadnotsmokedinthepast7days,andthatproductusedeclinedovertime[145].Two studies let end‐users participate in all conditions, which resulted in alonger playing‐time than allocated, higher than expected speechimprovementscomparedtonaturalconversation[152],andthathealthyend‐users significantly increased successful pointing tasks and challengedexperiences in the tailored session, compared to random adaptation butwithout differences in experiences (difficulty, frustration and fatigue) in apost‐stroketherapygame[147].Studieswithacontrolgroupfoundsignificanteffectsintheinterventiongroupregardingadecreaseinfatmass,weightandBMI (Body Mass Index)[151], better arithmetic skills, higher intrinsicmotivation,feelingsofself‐competencyandattention[146],andenhancementinall8domainsofcognitiveperformancescomparedto4inadherenceonly,or6inintent‐to‐treatofthecontrolgroup[148].


39

Combining the Product Designand Tailoring Phase

Table13.CharacteristicstoanalysethequalityofstudiesinProductDesignandTailoringPhase




Numberofparticipant

Validandreliable


[157] X X2 X1 X2 X1 X1=7X2=16 X X

[158] X X X 19&7 X

[159] X X X X 1 X X

[160] X X X 19 X

[164] X X X 115 X X

[165] X X X 1 X

[166] X X X X 6 X

[167] X X X X 8 X X

[168] X X X 7 X X

[169] X X X 18 X X

[171] X X X 7&1 X

[170] X X X 3&1 X X

Note:X1isstudy1,X2isstudy2.

StudiesthatinvolvedstakeholdersinProductDesign‐andTailoringPhasehadsome methodological limitations. The studies were effective with regard tovarious outcomes measurements. For example, end‐users had positiveexperienceswhileinteractingwiththeprototypegame,e.g.,thattheprototypewasappealingbutalsohadsomeimprovementsforthedesign[158],orthattheproductwaseasytousebytheend‐users,whoalsoexperiencedahigherperceived wellbeing [159]. Studies with a control condition, showedimprovements in physical health in the experimental condition on physicalhealth, for example, a significant increase in steps per week, but also anincrease in weight, BMI and percentage body fat in both experimental andcontrol condition [164], and an improved medication adherence accuracyfrom 43% to 56% in end‐users interested in games [169]. A paper that


40

involved two studies reported significant improvements in the upper limbmotorfunctioninbothstudies,aswellasimprovementsinglobalfunctioninthe first study [157]. A study where end‐users participated in all studyconditions,withdifferenttailoringalgorithms,reportedthatsevenoftheeightparticipants could interact with the product, of which six reached therecommendedenergyexpenditure levels, and that thealgorithms influencedscoresandexperiences[167].OtherstudiesfoundthattheBMIdecreasedforoverweight/obeseparticipants, increased inanunderweightparticipant,andwasmaintainedinhealthyparticipants[160]andthattailoringgame‐levelstothe abilities and performance positively affected body movements duringtherapy[168].End‐usersconnectedwiththeiravatar(itrepresentedtheminperformance), and this had (in)significant positive effects on upper‐limbstrokerehabilitation[170].StudiesthatinvolvedparticipantswithASDfoundsome engagement with a robot through interaction flow and self‐initiationbehavior,butwithroomforimprovement[165],andmoresocialengagementand lessplayingalone,butonlywhen interactingwitha robot [166].Lastly,five out of seven active duty soldiers and one veteran with PTSS weresuccessfully treated by the use of a tailored Virtual Reality, but one did notbenefitandtwootherparticipantsdiscontinuedthetherapy[171].

Other combinations of involving stakeholders in the design Phases

Table14.CharacteristicstoanalysethequalityofstudiesinvolvedindifferentcombinationofthePersonalizedDesignprocessPhases





Validandreliable


[161] X X X X 18 X X

[162] X X X 10 X X

[163] X X X 8 X

[156] X X X 6 X

StudiesthatinvolvedstakeholdersindifferentcombinationofthePDPphaseshad somemethodological limitations. Studies noted that a product could be


41

more than an icebreaker as it could improve client‐patient relationship, butthere were also some engagement concerns, according to therapists [163].Additionally, studies with a control group resulted in enhancements inselective and sustained attention and in overall visual motor abilities,combined with future design requirements [162], but also in designrequirementsforaproducttomotivatephysicalactivityinadolescents[161].Lastly, involving stakeholders in Problem Definition‐ and Tailoring Phaseresulted in effects on playability, where only one user improved control ofgestures[156].

2.3.4. VALIDATINGTHEINFLUENCEOFGAMES INVOLVEDINPERSONALIZEDDESIGNPROCESS

Normally, effects are quantitatively measured by studies that compare anexperimental conditionwith a control group. However,most studies in thisliteraturereviewwereoflowmethodologicalqualityduetothelownumberofparticipants,absenceofcontrolgroup(s)oruseofqualitativemeasurements(e.g.,[96])ornon‐validatedquestionnaires.Validatedquestionnairesthatareoftenusedtoquantitativelymeasuretheusabilityorexperienceofgames,aretheUserExperienceQuestionnaire[179]andtheSystemUsabilityScale[180].Other validated questionnaires that canmeasure the health related transfereffect alsoexist (e.g.,ChildDepressionRatingScaleRevised (CDRS‐R) [181],whenmeasuringdepression).Aminorityof 17 studiesdid includea controlgroup,whereend‐usersonlyparticipatedineitheranexperimentalorcontrolcondition. All found positive results in health related transfer effect,interactionexperienceandbehavior.Ofthesestudies,amajorityof9studiesrandomly assigned their participants to either the control or to theexperimentalgroup[[97,132,140,144,146,148,157,162,169].

Generally,thereviewedstudiescomparedexperimentalgroupswhoreceivedaseriousgame,thatwasdesignedviastakeholderinvolvementinphasesoraphaseofthePDP,withgroupsthatreceivedTreatmentAsUsualwithoutsucha game. Other studies compared groups of patients with groups of healthyend‐users.Suchacomparisoncanclearlyshowthehealtheffectoftheseriousgame,butmakesit impossibletoshowtheeffectofpersonalization.Inorderto test the effect of personalization, a comparison should have been made


42

between aproduct thatwasdevelopedwith stakeholder involvement in thePDP and a product thatwas developedwithout stakeholder involvement inthePDP.Amajorityofthestudieswithacontrolgroupinvolvedstakeholdersin theTailoringPhase [97,130‐132,137,140,144,146,148,151],ofwhichsome studied the effect of personalization by comparing tailored vs. non‐tailoredinterventions[97,132],orapersonalizedinterventionwithalikewisenon‐personalizedintervention[130].OtherstudiesthatinvolvedstakeholdersintheTailoringPhase,comparedactivitieswithapersonalizedgameproducttoastandardactivitythatusedapaperandpenmethod[131],amotivationalandtailoredlearningmethodwiththesamelearningmethodbutwithoutthemotivationalandtailoredvariables [146],andpatientswithhealthycontrols[137].Studiesalsocomparedapersonalizedgamewithtwoelectedexercises[144], a tailored training with a computer game [148], a training withadditionalatailoreddietgame[151],ordidnotstudyorfurtherdescribethecontrol group in the paper [140]. Involving stakeholders in the TailoringPhase seems to have positive effects on the end‐users regarding theinteractionexperiences(e.g.,thataproductwasusableandacceptable[137]),interaction behavior (e.g. reduced agitation and improved mood during anintervention with Alzheimer patients [130]), and health related transfereffects(e.g.significanteffects intheinterventiongroupregardingadecreasein fat mass, weight and Body Mass Index, [151]). For a more detaileddescriptionoftheeffects,seetheresultsection3.3.BecausesomestudiesthatinvolvedstakeholdersintheTailoringPhaseusedanexperimentalset‐up,thatcomparedatailoredwithanon‐tailoredgroup,wecanonlydrawconclusionsregardingtheadditionaleffectofstakeholderinvolvementinthelastphaseofthePDPandrecommendgamedesignerstoinvolvestakeholdersinthisphase.

StudiesthatfocusedonthehealthrelatedeffectofstakeholderinvolvementintheProblemDefinition‐andProductDesignPhaseofthePDPdidnotfocusontheeffectofthispersonalization.Therefore,itisdifficulttoconcludethatthereported outcomes of the studies were due to stakeholder involvement inthese phases. However, some studies did attempt to study the effects of apersonalizedgame.Forinstance,thetwostudiesthatinvolvedstakeholdersinboth Problem Definition‐ and Product Design Phase [121, 122], compared


43

participants with autism spectrum disorders (ASD) to participants withoutASD or two experiments to see if there was improvement on the outcomemeasures between the two. These studies found mixed results regardinginteraction behavior [121] and positive self‐assessed health related transferoutcomesregardinglearningaboutahealthaspect[122]. ThreestudiesthatinvolvedstakeholdersinbothProductDesign‐andTailoringPhase[157,164,169], compared an exercising game with exercises in laboratory sessions[164],twogroupsthathadthesameappofwhichonealsoconsistedofagame[169], and therapy alone with the same therapy that also consisted of apersonalized game [157]. The games that were designed by stakeholderinvolvementinboththeProductDesign‐andTailoringPhaseseemedtoresultin positive [157, 169] and mixed results regarding interaction experiencepositive [164], positive results regarding interaction behavior (e.g., how theend‐user used the product [169]), and health related transfer effects (e.g.,regarding physical functioning and improvement in medication adherence[157, 169]) that were not always fully positive [164]. Only two studiesinvolved stakeholders in all the PDP phases [161, 162], and compared atreatment group (that received extra sessions with games) with a controlgroup(whodidnotreceivethese),orletgroupsusedifferentkindoflikewisetools. This seemed to result in different but mostly positive interactionexperiences [161], and in interaction behavior and health related transfereffects [162]. For amore detailed description of the effects of studies thatfocusedon theeffectof these combinationsofphases, see the result section3.3. To study the additional effect of stakeholder involvement across the(other) phases of the PDP, future studies need robust experimental designsthatcomparepersonalizedversusnon‐personalizedgames.

2.3.5. APPLIEDGAME‐ELEMENTS INREVIEWEDPAPERSTable15.Game‐elementsinthePersonalizedDesignPhases.Game‐element PersonalizedDesignPhases

ProblemDefinition

ProductDesign

Tailoring

Avatar [96,115,121,162]

[95,121,157,158,160,162,

[97,127,129,131,132,134‐137,143,144,154,155,157,158,160,162,167,


44

167,170] 170]

Rewardpoints [123,156,161,162]

[118,119,123,160‐162,164,167‐170]

[137‐139,143,144,149,153‐156,160‐162,164,167‐170]

Progression/level [93,96,115,123,124,156,161,162]

[118,123,124,157,158,160‐162,164,168,170]

[130,131,133,135‐142,144‐149,151,153,155‐158,160‐162,164,168,170]

Social(e.g.,leaderboards)

[96,116,121,124,125,161]

[118,121,124,125,159‐161,164,169]

[128‐130,132,141,143,159‐161,164‐166,169,170]

Puzzle,cards,quiz [96,115,116,122,123,125,126]

[94,117‐119,122,123,125,126,158,159,164‐166]

[128,130,150,151,158,159,164‐166]

Userassignmentsinreallife

[93,96,116,122,125,161]

[118,120,122,125,161,164]

[133,135,137,140,142,143,145,161,

164]

Others [163]

[95,121,157,159,163,164,168,169,171]

[133,135,137,138,141,143,144,146,147,149,150,152‐155,157,159,163,164,168,169]

The games that were described in the reviewed papers contained specificgame‐elements. Game‐elements are the elements that are found in games[182],thatmotivatetheplayerforspecificbehavior[183].Inpapersthatfocuson personalization, the game‐elements have a more abstract role whenstakeholders are involved in theProblemDefinitionPhase, compared to theProductDesign‐andTailoringPhase.Generally,game‐elementsareshapedintheProductDesignPhase,ofwhichsomestudiestestthesegame‐elementsinthe Tailoring Phase. Therefore, game‐elementsweremore “visible”, becausethey are better described and tested in the Product Design‐ and TailoringPhase(Table15).


45

When stakeholderswere involved in the Tailoring Phase, gameswere usedthat mostly contained game‐elements like “progression”, “levels”, “rewardpoints”and“avatars”.Especiallyanavatarwasoftenpresent,sincetheycouldbetweakedtothepreferenceorbehavioroftheuser(e.g.,letend‐userstailoranavatarbygiving them freedomtodoso [97]). “Points”and “progression”wereoftencombined,becauseprogressionorlevelscanmakethepointsmoremeaningful:byreceivingpoints,theend‐userscanseehowtheyprogressandeventually reach higher levels. When stakeholders were involved in theProductDesignPhase,“progression”and“points”werealsooftenpresent,butmostly combined with social game‐elements, puzzles, and quizzes. In onestudy, end‐users could tweak a game‐element themselves, e.g., by adjustingthe difficulty level of the game [146]. The type of tailoring (e.g., tailoringavatars) was studied in isolation. However, if more game‐elements werepresentinaproduct,theeffectsofthesegame‐elementswerenotmeasuredinisolation,butasa“blackbox”.

2.4. DISCUSSIONANDCONCLUSIONTo our knowledge, this is the first systematic literature review onpersonalization in games for health. Existing literature from personalizeddesign methods were applied to propose the Personalized Design Processmodelandtoinvestigateifandhowpersonalizationingamedesigniseffectivein the context of health. The aim of the PDP‐model,which consists of threedifferentdesignphasesinwhichpersonalizationcanbeapplied,istoprovideinsights in when personalization can be applied in the design process. Theeffect of this personalization is aligning a product to the needs andpreferencesoftheend‐users[53,184].Thiscanincreasethesatisfactionwithand thevalueof theproduct [66,67], the interaction timewith theproduct,andconsequentlyitcanpositivelyinfluencethehealthrelatedtransfereffect.Itcanbeconcludedthatstakeholders(mostlyend‐usersanddomainexperts)wereofteninvolvedintheTailoring‐andProductDesignPhaseonly,andnotin the Problem Definition Phase of the PDP. The problem was often pre‐defined, for example, by the government or principal, without any checkwhether it was the correct problem to target [66]. However, it would bepreferable toalso involvestakeholders inProblemDefinitionPhase,because


46

this can serve as the basis of thewhole design process and provide amoreholisticpictureofproblemsandaspectstofocuson[185].

Iftheproblemthatthegamewilltackleisnotcheckedwiththestakeholders(especiallytheend‐usersanddomainexperts), it ispossiblethatthegameisdesignedforaproblemthatdoesnotexistorisnotpossibletoimprove.Thiswouldmakeitdifficultorevenimpossibletoobtainpositiveresults,especiallyregardinghealthrelatedtransfereffects.

Within the Product Design Phase, we observed different stakeholderinvolvement, where stakeholders were most often only involved inEmbodiment. They could provide comments, suggestions or guidelines, butthey could also design products, help designers, and give feedback. It isprobablethatitwouldhavetakentoomuchtime,beentooexpensive,ornotfoundnecessarytoalsoinvolvestakeholdersinshortpassivetestsinIdeation(for example, about the usability) [186]. A majority of the studies onlyinvolvedstakeholdersintheTailoringPhase.Possibly,thisphaseisthemostimportant phase for stakeholder feedback in the designprocess, or perhapsstakeholderscanbe involvedmoreeasily in thisphase.StudiesoftenusedaKinect device to give tailored visual feedback to the user about theirperformanceor remaining time, or to tailor thedifficultyof the tasks to theend‐user input (e.g., in‐game performance). The end‐user’s name was alsooftenusedwhenfeedbackwasgiven,inordertomakethefeedbackorcontentmore personally relevant and thus more motivating and persuasive for theend‐user. End‐users actively tailored avatars, which gave them theopportunity to connect to the avatar, as if it was a representation ofthemselves. Unexpectedly, the objectives, difficulty level or the stimuli of aproductwereoftentailoredbyothers(mostlydomainexperts).Therewasanexpectation that both a domain expert and end‐user would be involved intailoring these assignments together, e.g., by letting domain experts tailorassignments of which an end‐user could choose from, because this wouldoptimallycombinetheexpertiseofboththeend‐users(aboutpreferencesandneeds) and the domain experts (about theoretical proven assignments or


47

therapy aspects). In addition, it would also give the end‐users a sense ofchoice,whichcouldpromoteengagement[187].

The studies generally found positive effects on interaction experience,interactionbehaviorandhealthrelatedtransfereffects.However,becausetheduration(intime)ofthestudieswasheterogeneous[128,137,156,166]andamajority of the validation methods were not methodologically sound (i.e. alow number of participants or use ofmeasurements thatwere not valid orreliable), it is hard to compare the results of these studies and to warrantconclusionsontheeffectsofstakeholder involvementacrossthePDP.Of the62studiesincludedinourliteraturereview,amajority(50)hadasmallstudysample (N=< 50). Only 17 studies used a control group, of which ninerandomlyassignedtheirparticipantstoeitherthecontrolortheexperimentalgroup [97,132, 140,144,146,148,157, 162,169].Taking these limitationsand results into account, it can only be suggested that it is important toinvolve stakeholders across the Product Design‐ and Tailoring Phase for amore effective design of games for health. It should be mentioned thatinvolvingstakeholdersinthePDPcanalsohavesomedisadvantages.Ittakestime,money, andeffort to let stakeholdersparticipate in thedesignprocess[186]. However, if involving stakeholders in the PDP results in betteroutcomes regarding the experience, behavior and health related transfereffects, this can be seen asmore important compared to the disadvantages.Balancing the amount of personalization to the expected outcomeenhancementshouldbeperformedinadvanceofeachgamedesignprocess.

Itispossiblethatourreviewstrategydidnotresultinretrievingallavailablestudiesontheeffectsofstakeholder’ involvementinthePersonalizedDesignphases, because of the different definitions of personalization that currentlyexist [188]. We attempted to minimize this, by brainstorming the searchstrategyandselectionofkeywordswithexpert researchers fromthe fieldofco‐designandpersonalization,andfrompsychologyandgamedesign.WealsoproposedaPDPmodel,whichwouldmakeitpossibletoextendthepotentialofpersonalizationtowardsabetterdesign in thecontextofhealth,and limitthe confusion within this field. It should be noted that we did notsystematically take user‐centered design into account, because according to


48

ourdefinition,itcanbepartofco‐design.Itisstandardtoatleast(iteratively)test a product once with possible end‐users, and to check whether theyunderstandandcanusetheproduct[53,66].

Within the PDP model, we used the general term ‘Tailoring’ for both UserControlled Customization and Use‐Dependent Adaptation. Some studiesfocusedonthetechnicalchallengesregardingtailoring,andnotontheeffectsoftailoring.Thismadeitdifficulttostudytheeffectsofthedesignoutcomesinterms of personalization. It should be noted that in addition to our twotailoring types, we also found another type of tailoring that we termed“Context Dependent Adaptation”, where a product is tailored based on thespecificcontextoftheend‐user.However,withthistype,theend‐userhasnoactive role in tailoring. Therefore, we did not focus on Context DependentAdaptation in this review.Examplesof thiskindof tailoringarestudies thatpersonalizeda game to the contextof end‐user (treatmentof burnwounds)[189], that let designers make suitable levels for end‐user context withoutinfluence of end‐users or other stakeholders [190], or where tailoring wasdonebasedongamertypes,aimingtomotivatebehaviorforeachgamertype[191].

In addition to the definition of personalization, the involvement ofstakeholdersinatleastoneofthethreedesignphasesofthePDP,theremayhavebeenconfusion regarding thedefinitionof gamesandgamificationandthus some studies could have mistakenly been excluded. Generally,gamification is defined as “the use of game design elements in non‐gamecontexts”[192]andgameswithaseriousaimcanbeconsideredasgamesthatdonotprimarilyhaveanentertainmentpurpose,butaimtowardssomething“serious”, like influencingknowledge [182]. It is interesting that inonly onestudy, end‐users could actively tweak the game‐elements, e.g., the difficultylevel [146]. Some game‐elements were commonly involved across thereviewed studies. Points and progression were often combined, e.g., bydesigningathresholdofanumberofpointsbeforetheusercouldproceedtothenextlevel,andavatarswereoftenappliedtorepresenttheuserwithinthegame. However, the effects of these game‐elements were not measured in


49

isolation,butasa“blackbox”,makingitimpossibletomeasuretheinfluenceof separate game‐elements. Only the ‘way of tailoring’, e.g., tailoring thedifficulty level to the user or an avatar in different ways, was studied inisolation. It would be preferable to study the separate influence of specificgame‐elements, to knowwhich specific game‐element influencesmotivationandeffect,andhowthisoccurs.

Regarding the different stakeholder involvement when designing games forhealth, Baranowski et al. (2016) [193] divided the stakeholders into “thosewho (a) are interested in using games for health to advance their or theirorganization’s agenda, (b) may benefit from playing the games, (c) creategames for health for profit, and (d) conduct research on games for health.”[193].ThePDPstakeholderscanbedividedaccordingtotheseroles.Domainexperts are those “interested in using games for health to advance their ortheirorganization’sagenda”,end‐users, familyanddomainexpertsarethosethat “may benefit from playing the games” and designers and experts arethosethat“creategamesforhealthforprofit”,aswellasthosewho“conductresearch on games for health.”. We thus agree that involving stakeholdersacrossthecreationofagameinthecontextofhealthisimportant,inordertoensurethat thegamemeetstheirneeds,expectationsandpreferences[193].However,werecognizeadifferencebetweenBaranowski’sstakeholders.Themain difference is that our PDP model also explicitly takes the ProblemDefinition‐andTailoringPhaseintoaccount,andthatstakeholdersshouldbeinvolved across all the PDP phases when designing games that aim topositivelyinfluencehealthaspectsofend‐users.

Accordingtoameta‐analysisbyDeSmetandcolleagues(2014),gamesshouldbe dynamically tailored to both behavior change needs and socio‐demographic information(e.g., tailorthedifficulty level towhattheusercanmaster). This is already present in our PDP, by involving stakeholders thatknow about theories of games and behavior change (domain experts anddesigners) in the whole PDP [83]. We defined ‘tailoring based on theperformance of the user’ as “Use‐Dependent Adaptation”, and ‘tailoring tosocio‐demographic aspects of the end‐user’ as “Context DependentAdaptation”.


50

In ameta‐analysisof61 studies focusingon stakeholder involvement (morespecificallyend‐users)inseriousdigitalgamesforhealthylifestylepromotion,DeSmet and colleagues (2016) found other results when end‐users wereinvolved in designing games for healthy lifestyle promotion. If they wereinvolved in “participatory design”, where end‐users were involved asinformants,behaviorwaschangedsignificantlylesseffectivelythanwhentheywere involved in pilot‐testing [52]. Participatory designwas also related tolowereffectsonself‐efficacythanwhenend‐userswerenotinvolvedingamedesignorinpilot‐testing.Wheninvolvedinco‐design,strongerhealtheffectswerenotedwheninvolvedingamechallenges,butweakerhealtheffectswheninvolved in character and game world creation. This suggests that howstakeholdersare involved inpersonalizationcan influencethehealtheffects,andthataspecifictypeofstakeholder(end‐users)shouldbeequalpartnersindesign instead of being only informants. However, it should be noted thattheir‐andourdefinitiondiffers:accordingtoDeSmetandcolleagues(2016),participatory design represents stakeholder involvement as informants(wheretheygiveinputandfeedback)orasco‐designers(wheretheyareequalpartners). Co‐design thus has a specific role that end‐users can havewithinparticipatory design. However, according to our definition, participatorydesignandco‐designdifferfromeachother,sinceco‐designputslessweightontheemphasisofuserempowerment,asisthecaseinparticipatorydesign[50]. In addition, DeSmet and colleagues (2016) state that users are equalpartners in the design process when co‐designing. We believe that it isimportant togive theuser thepositionof ‘expertofhis/herexperience’,butthat the design responsibility belongs to the design team, since that is theirfieldofexpertise.Besides,wealsotakeintoaccountotherstakeholder,insteadofonlyend‐usersasisthecaseinthestudyofDeSmetandcolleagues(2016).

2.5. FUTUREWORKANDCONCLUSIONSTo conclude, the results of our literature review do not yet allow definiteconclusions about whether and when involving stakeholders in the PDP(ProblemDefinition‐,ProductDesign‐andTailoringPhase)hasaddedvalueintermsofeffect.Therefore,thecurrentmotivationtoinvolvestakeholderscanbe seen as a theoretically driven concept rather than an empirically driven


51

concept. Our findings suggest that stakeholders should be involved in theTailoring and Product Design Phases. However, a majority of the reviewedstudieswerehamperedbysmall samplesize, lackof control conditions, andother methodological weaknesses. Future studies thus warrant solidevaluation and design strategies for personalization, which may lead toempirically founded conclusions that personalization really enhancesbehavior, experiences or the health related transfer effect in the context ofgamesforhealth.However,thesefuturestudiesshouldchoosetheirresearchmethodwithcare,becausethe“goldenstandard”forexperimentalvalidation,a placebo‐controlleddouble‐blind study, is complicatedusing serious gamesforhealthasinterventioninsteadofamedicalpillyoucanswallow.Designingproducts with stakeholder involvement takes a lot of time, and designingcontrol groups in this context is difficult. This is because almost all seriousgamesmixtheseriouscontentwiththegamecontentandbecauseitisalmostimpossiblecreatea ‘placebo’‐or ‘control’‐gamebyremovingonecomponentwithout changing the other. Besides, end‐users may have differentexperiencesandbedifferentlyaffectedbyagame,andparticipantscannotbeblinded [193]. In addition, the ‘black‐box’of game‐elements shouldbemademorevisible,bystudyingtheeffectsofseparateandcombinedgame‐elementswithinthiscontext.

In this literature study, we have defined personalization and how it can beapplied within the games for health design process. We recommend thatfuturestudiesnotonlyfocusoninvolvingstakeholdersintheProductDesign‐andTailoringPhaseofaPDP,butalso tomethodologically testwhether thisstakeholderinvolvementinthePDPresultsinbetteroutcomesonexperience,behavior and health related transfer effect, by the use of suitable controlgroups[178].

52

Basedon:vanDooren,M.M.M.,Visch,V.T.,Spijkerman,R.,Goossens,R.H.M.,&Hendriks,V.M.(2020),TherapyProtocolsandEHealthDesign:aFocusGroupStudy,JMIRFormativeResearch(forthcoming)

3. THERAPYPROTOCOLSANDEHEALTH

DESIGN:AFOCUSGROUPSTUDY

Personalization is not only often applied in game design for healthcare, toenhance positive effects on interaction experience, interaction behavior andhealthrelatedtransfereffects,butalsoinatherapeuticprocess.Whenaimingtoredesignatherapywithpersonalizedgamification,tofacilitateefficiency,accessand quality of therapy, it is important to know how a therapy protocol ispersonalized.IfeHealthdesignersdonottakethis intoaccount,theredesignedtherapy might not optimally fit the therapeutic practice and impedeimplementation. Therefore, we wanted to generate information about theproportion,typeandreasonsforpersonalizationofagiventherapyprotocolbytherapistsandpatients.Weconductedtworoundsoffocus‐groupdiscussionsonhow a Cognitive Behavioral Therapy protocol in youth addiction care wasappliedinpracticebypatientsandtherapists.

Chapter3|DesigningeHealth:consideringthebalancebetweenprotocolapplicationandpersonalization

54

3.1. INTRODUCTION To reduce the adolescents’ risk of developing adverse consequences due tomental diseases (e.g. [9, 194]), adequate treatment is needed. Therapyprotocols contribute to the implementation of evidence‐based therapeuticpractice and help therapists to structure their face‐to‐face therapy sessions[55]. Although psychosocial therapies are effective in reducing psychiatricsymptomsinadolescentswithmentaldisorders,theavailabletherapiesshowmodesteffectsandnotalladolescentsbenefit[23,195].IncludingeHealth,theuse of information and communication technologies in the delivery of(mental) healthcare [26], in the therapeuticpractice is a promisingmean toimprove the patient engagement and therapeutic effectiveness (e.g. [38, 39,43]).

The therapy protocols that form the basis for face‐to‐face therapies aretypically used as a basis for the design of eHealth as well [62]. Therapyprotocolsplayalargeroleinthesuccessofevidence‐basedtherapies[56]anditisrecommendedtoapplytherapyprotocolsasmuchaspossible.However,both therapists and patients can personalize or only partly apply a therapyprotocol in therapeutic practice (e.g. [57, 59‐61, 196, 197]). For example,therapistscanbelievethatfollowingtherapyprotocolsgoesattheexpenseofastrongtherapeuticalliance[198],thetrustbetweenapatientandtherapistthat allows them towork together in an effectiveway and an indicator forpositivetherapyoutcomes[35].

The difference between therapy protocols and therapeutic practice hasserious consequences for eHealth design. If personalization possibilities intherapeutic practices are not taken into account in the design of eHealth,eHealth may not suit current therapy practice which limits its’implementation.Forexample,wheneHealthdoesnotsuithowtherapistsusethe therapy protocol or if therapists have negative expectations about thebenefitsofeHealthcomparedtoface‐to‐facetherapy[199‐201].ManyeHealthinterventionshavefailedtointegratepersonalizationtotheindividualuserinthedesign[202]


55

ToaligneHealthtotherapeuticpractice,itisimportanttoknowthecontentofthe existing therapy protocols and how these protocols are applied intherapeuticpracticebytherapistsandpatients.DesignersofeHealthcanusethis information to ensure that eHealth matches the therapeutic practice,consequently improving the quality and enhancing the implementationpotentialofeHealth.Therefore,thisexplorativestudyaimstogaininsightintopersonalizationpractices in amental healthcare context and concludeswithrecommendations to eHealth designers on how they can access and involvethe need for protocol personalisation in eHealth design. To achieve this,weexamined therapists’ and clients’ perceptions of protocol application in ayouth addiction treatment facility as a case study by generating bothquantitative and qualitative data. Firstly, we wanted to know how muchtherapists and patients personalized and applied. We generated thisinformationbyaskingthemtoestimatetheamountoftherapyapplicationandpersonalization in therapeutic practice. Secondly, we wanted to know howand why they applied and personalized the therapy and generated thisinformationbyusingthequantitativedataasinputfordiscussion.

3.2. METHOD

3.3.1. THERAPYPROTOCOLThe commonly applied protocol for Cognitive Behavioral Therapy inadolescent addiction care was used as a case protocol [203]. The protocolconsistsofninesessions,followedbyfour“sessionsofchoice”(selectedfromsevenoptionalsessionsinconsensuswithpatients).Ineachsession,patientssetspecificshort‐termgoalswithregardtothetherapeutichomework.Partofthetherapyprotocolisatherapyworkbook,thatpatientscanbringhomeandto therapy sessions. The activities that are described in the workbookcorrespondtothecontenttherapysessions.

3.3.2. PROCEDUREWe conducted semi‐structured focus group sessions in two phases at twolocations of one large out‐patient treatment facility centre for adolescentaddictioncareintheNetherlands(seeTable1).Theaimofthefirstphasewas


56

to investigate therapists’ and clients’ estimations of the amount and type ofprotocolpersonalization.

Therapists estimated how much of their therapy consisted of a strictlyfollowed and adapted therapy protocol and patients indicated howmuch ofthe provided therapy by their therapist they strictly followed and adapted.Important to note here is that clients thus could receive a personalizedtherapyprotocolinpractice.Boththerapistsandpatientsalsoindicatedhowmuchother(non‐protocol)therapeuticpartstheyadded.Theseamountswererepresentedbypercentages that addedup to a total of100%, i.e. thewholetherapy. The secondpart evaluated,with other therapists andpatients thanthosewhopartook inthefirstpart, theresultsof the firstpartandaimedtogain insight into the reasons for personalization. Participantswere broughttogetherwithamoderator(thefirstauthor)foradiscussionlastingonehour.Beforestartingthegroupdiscussions,wereceivedinformedconsentfromtheparticipants and explained the concept of personalization, i.e. changing adesigned end‐product, like a therapy protocol, to match the needs andcapacitiesoftheend‐userandenhancetheeffectivityoftheproduct[204].

3.3.3. PARTICIPANTSWeinvitedexperiencedtherapists,whoreceivedtrainingintheCBTprotocol,toparticipateinfocusgroupsessions.Patients,whowereatleast18yearsold,received CBT,were recruited by their therapists to participate in the study.Therapistsreceivedan information leaflet to informtheirpatientsabout thestudy.A therapist informedus ifapatientwantedtoparticipate. In turn,wecontacted the patient to schedule an appointment for the focus groupdiscussion.Inthefirstpart,sixtherapists(N=3female,N=3male)andfivepatients (N = 1 female,N = 4 male) participated. In the second part, threetherapists (N=1male,N=2 female)and fourpatients (N=1 female,N=3male) participated.All interviews tookplace at the youthmental healthcarefacility.


57

Table1.Set‐upofthefocusgroupsessionswiththerapistsandpatientsonprotocolapplicationandpersonalization

Goal Part Location Participants N

Generateinformationwithseparategroupsofeithertherapistsorpatientsabouthowmuchoftheirtherapyconsistedofastrictlyfollowedtherapyprotocol,adapted

therapyprotocol,andaddedtherapeutic

parts.

P1 A Onegroupdiscussionwiththerapists 3

A Onegroupdiscussionwithpatients 2

B Onegroupdiscussionwiththerapists 3

B Onegroupdiscussionwithpatients 2

B Oneinterviewwithapatient 1

JointevaluationoftheresultsfromP1withcombinedgroupsofbothpatientsand

therapists.

P2 A Onemixedgroupdiscussionwithatherapistandtwopatients

3

B Onemixedgroupdiscussionwithtwotherapistsandtwopatients

4

3.3.4. DATAANALYSISThedataarepresentedaccordingly to thestandardsof reportingqualitativeresearch byO’Brien et al, 2014 [205].We used thematic analysis instead ofgrounded theory to analyse the data. With grounded theory, one wants togenerate an exploratory andoverarching framework or theory [206],whichwewerenot interested in.With thematic analysis, themes arederived fromthedata[207‐209]inwhichwewereinterestedsincethesecoulddirectlygoto theguidelines foreHealthdesigners.We focusedon thephases thatweredescribed in thepaperofBraunandClarke(2006),whenanalysingthedata[207].Theresearcher,alsothefirstauthorofthispaper,wasaPhDcandidateandhad twomasters in clinical andhealth psychology. Shewas qualified indoingqualitativeinterviews,hadneitherassumptionsaboutnorarelationshipwiththeparticipantspriortothediscussions.Allinterviewstookplaceattheyouthmentalhealthcarefacilitiesofthetherapistsandpatients.Experiencedtherapistswere invitedtoparticipate inthestudy.TheyhadtobetrainedinthenewCBTprotocolthestudywasfocusingon.Patients,whohadtoreceive


58

CBT, were informed about and invited to participate in the focus groupdiscussions by their therapists. In this way, more patients were informedabout the study which enhanced the chance that more patients would bewillingtoparticipate.Afterthepatients’consenttoparticipate,theresearcherwould contact the patient by phone to ensure everything was clear to thepatientandtomakeanappointmentforthefocusgroupdiscussion.

We received formal ethics approval from the Human Research EthicsCommittee of the Delft University of Technology in the Netherlands. Alldiscussions were recorded with an audio recording device, after receivingverbalconsentfromparticipants.Quantitativedatawassavedwithonlyalinkto the type of participant (i.e. being either a therapist of patient). All focusgroupsessionstookonehoureachandwereaudiotapedandtranscribedbythefirstauthor.Interviewguideswereusedduringthediscussionsandduringandafterthediscussions,fieldnotesweremade.

Aftertranscribingthedata,thefirstauthorwentthroughallrecordingsagainto check the accuracy of the transcripts in line with [205]. Next, the firstauthorwentthroughallthetranscriptsmultipletimesbeforecodingthedata.This ensured that the themes generated from the codeswere not based ononlyafewexamples.Similarthemesweregroupedtogetherintohigherlevelthemes.Whenanalyzingthedata,themeswerelinkedtoeachother,ensuringacoherentstory.Enoughtimewasallocatedtoanalyzethedataadequately.

ThefourthtopicofO’Brienetal(2014)focussesontheresults(topic16and17)thataredescribedinthefollowingsection[205].Supportivequoteswerechosentosubstantiateanalyticfindings.Thisisfollowedbythefifthtopicthatdescribes the discussion section (topic 18 and 19) and last “other” topic,whereconflictsofinterestsandfundingarementioned(topic20and21).

3.3. RESULTS

3.3.1. PART1: FOCUSGROUPSESSIONSWITHTHERAPISTSTherapistsindicatedthattheystrictlyapplied30–75%(mean48%)ofthetherapyprotocolandadaptedbetween10–50%(mean30%)ofthetherapyprotocol. They further reported to add 10 – 33 % (mean 22 %) of non‐


59

protocol therapeutic parts. Thepercentages of one therapistwere excluded,becausethepercentagesofstrictapplicationandpersonalizationoverlapped(seeFigure1).

Figure1.Therapists’meanestimatedpercentagesofCBTprotocolusage.

Note:Thepercentagesofonetherapistwereexcluded,becausethepercentagesofstrictapplicationandpersonalizationoverlapped.This

isimpossibleaccordingtoourdefinition,becausethecategoriesofprotocolapplicationareexclusivecomponents.

Wefirstwentthroughthequotesseveraltimesandgeneratedcodesfromthequotes that focused on reasons for therapists to personalize the therapyprotocol. These codes referred to therapists who personalized the protocolbased on what the patient needed “Corking works the best, adapting [thetherapy]towherethepatientis” (Therapist1A),what they thoughtwouldbemorebeneficial for thepatient“Itisrelatedtoiftheyhavestopped[withusingsubstances], than focusingon cravings.Also if theyalreadywent into therapybefore,thanremovingelementsofwhichyouthinkarearepetition”(Therapist2A), and because therapists knew other therapy protocols that could helppatientswith different problems at the same time “IalsogivegrouptherapyandsomeelementsofwhichInoticethatwork[duringgrouptherapy]Ialsouseduring individual therapy” (Therapist 1B). Besides, codes mentioned thattherapistspersonalizedtoenhancethetherapeuticalliance.“Muchisrelatedtotheconnection,thetherapeuticalliance is importantsoI investa lotoftime inbuildingone”(Therapist1B).Wewentthroughthecodesagainwhichresultedinhigher‐levelthemes.

48%

30%

22%Strict

Adapted

Added


60

Themain theme thatwe derived from the codes of the therapistswas thattheyusedtheprotocolasa“toolbox”,abundleoftherapytoolsthattheycouldchoose from. The code that did not fit this main theme, focused on addingelements from other therapy protocols. However, all therapists mentionedthattheydidnotapplytheorderofthetherapyprotocolinastrictway.Theprotocolwasnotseenasastep‐by‐stepmanualbutamanualthatconsistedofall the possible interventions. “ToolboxCBT, Idonotuse itasa step‐by‐stepmanual but I can choose interventions from the toolbox that I find relevant.”(Therapist1B).Threesub‐themeswerederived:therapistswhopersonalizedbasedonwhattheythoughttheirpatientneeded,ontheirowntherapy‐givingexperiences,orbecausetheythoughtitenhancedthetherapeuticalliance.

The first subtheme to the grandToolbox theme consistedof therapistswhopersonalized the therapyprotocol based onwhat they thought their patientneeded. They thought that by adapting the therapy, their patientswould bebetter prepared to specific situations. This was influenced by the (possibledifficult)situationsthatpatientsexperiencedpriortothetherapysession(e.g.hadanargumentwith theirparents),howthemotivationofpatientswas tochangetheirbehaviourandifpatientsunderstoodallelementsofthetherapyprotocol. For example, therapists tried to enhance the trust of patients thattheycouldachievethegoalstheysetorbymainlyfocussingonthehomeworkthatapatientdidwellinsteadoffocusingonthehomeworkthatapatientdidnot do well. “What is important for patients, such as dealing with socialpressure.Ingeneral,Ifollowthetherapyprotocolbutifyounoticethatpatientshavedifficultieswithit[socialpressure]youfocusonthat”(Therapist3A).

ThesecondsubthemetothegrandToolboxthemeconsistedoftherapistswhopersonalized the therapy protocol based on their own therapy‐givingexperiences. During the discussions they commented to not apply or onlypartly apply the workbook to prevent their patients from experiencingfeelingsoffailure,sincepatientsgenerallyforgottobringittotherapyand/orfill in the homework assignments. Therapists thought that not applying theworkbook prevented their patients from experiencing feelings of failure. “Ialwaysestimateifit[thepatient]isthetypeofpersonthatcandohomeworkat


61

home, if it [the patient] is someonewhowill really do it [the homework, youwant to prevent experiences of failure” (Therapist 3B). In addition, moreexperienced therapists have more knowledge of and experience with otherdifferent therapy protocols. Therefore, more experienced therapists tend toapply elements from other therapy protocols during therapy more oftencomparedtolessexperiencedtherapists.

ThethirdsubthemetothegrandToolboxthemeconsistedof therapistswhopersonalizedthetherapyprotocolbecausetheythoughtitwouldenhancethetherapeutic alliance. “It depends on the connection [between me and thepatient].Thetherapeuticalliance is important,onwhich Ispenda lotoftime”(Therapist1B).Theywouldtrytoworkonthebondtheyhadwithapatientbyfocussing more on the positive steps a patient made, compared to what apatient did not do. In addition, this would also enhance the motivation ofpatientstocontinuewiththerapyandtrytoachievethetaskstheyagreedon.

3.3.2. PART1: FOCUSGROUPSESSIONSWITHPATIENTSPatients indicated that they strictly applied 12 – 65% (mean 29%) of theprovidedtherapybytheirtherapist,adaptedbetween9–64%(mean48%)andaddedbetween18–26%(mean23%).Thepercentagesofonepatientwere excluded, because the percentages of strict application andpersonalizationoverlapped(seeFigure2).


62

Figure2.Patients’meanestimatedpercentagesoftherapyapplication.

Note:Thepercentagesofonepatientwereexcluded,becausethepercentagesofstrictapplicationandpersonalizationoverlapped.

Thisisimpossibleaccordingtoourdefinition,becausethecategoriesofprotocolapplicationareexclusivecomponents.

We first went through the quotes several times and generated codes thatfocused on reasons for patients to personalize their therapy. These codesreferred to patients who personalized how they achieved their homework,because theypreferred topersonalize“ActuallyItrytothinkofsomerulesformyself”(Patient1B),andbecausetheyweresomewhatcarelessnessandforgottheir homework “It isquitehard tokeepupwith itand it isnotreally inmyroutine,likebrushingmyteeth”(Patient2A).Inaddition,thepersonalizationofpatientswasinfluencedbytheconnectiontheyhadwiththeirtherapist“Theconnection you havewith your therapist influences howwell therapyworks”(Patient2A).Wewentthroughthecodesagainwhichresultedinhigher‐levelthemes.

Themainthemethatwasderivedfromthecodesofthepatientswasthattheypersonalizedthetherapybasedontheirownsituation.Thecodethatdidnotmatch the main theme, focused on personalization of therapy by thetherapists.Eventhoughtherapistsandpatientsdecidedonthehomeworkthepatientwouldworkontogether,allpatientsmentionedthattheypersonalizedtheir homework. Two sub‐themes were derived: personalization to better

29%

48%

23%Strict

Adapted

Added


63

matchtherapywiththedaily lifeof thepatientandpersonalizationthatwasinfluencedbythevaryingmotivationofpatients.

Thefirstsubthemetothegrandownsituationthemefocusedonpatientswhomentioned that theypersonalized their therapy tobettermatch the therapywith their daily life. They personalized to match the therapy to their ownsituation, personality and preferences. This lowered their feelings of stressandworry.“Ialwayschangeit[doingthehomeworkassignments]alittlebitsothat it is in linewithmy personality and how Iwant to be seen by others”(Patient2A). “…Itistheintention[todotheworkbookassignments],butIdon’tdo it. I rather tell about it [cravings] than towrite these experiences down”(Patient1B).

Thesecondsubthemetothegrandownsituationthemeconsistedofpatientswhomentioned that the amount of their personalizationwas influenced bytheir varying motivation. They sometimes just did not want to do thehomeworkassignmentsorforgottodothehomeworkassignments.Besides,arelapsecouldinfluencethemotivationtocontinuetherapyineitherapositiveor a negative way. “Sometimes, I just do not feel like doing it [workbookassignments]andIjustdonotdoit”(Patient1A).Onepatient said thatdoingtheworkbookassignmentsforalongerperiodoftimehelpedhimtogenerateinsights in triggers for craving. The therapeutic alliance influenced theirmotivation,mainlybecauseatherapistwouldputthingsintoperspective(alsoifapatienthadarelapse).

Nexttothethematicanalysis,weanalyzedthequantitativedataofthepatientsand therapists.Wecombinedallpercentagesof therapyprotocolapplicationandpersonalizationby therapist andpatients intoone figure (seeFigure3).Thiswasused in the secondpartof the focusgroupdiscussions,withotherpatientsandtherapists.


64

Figure 3. RangeandmedianpercentageofCBTapplicationbytherapistsandpatients.

3.3.3. PART2: FOCUSGROUPSESSIONSWITHBOTHPATIENTS

ANDTHERAPISTS We first went through the quotes that focused on reasons for patients andtherapists to personalize the therapy (protocol) and generated codes fromthesequotes.Thecodesof thetherapistsreferredtopersonalizationtokeeporenhancethemotivationofpatients, toworkonaconnectionof trustwiththeirpatient,andpersonalizationof the therapy inorder toalign therapy tothe problem of the patient. The codes of the patients referred topersonalization by discussing what was happening in their life during atherapy session, and personalization of homework based on how they feltduringtherapyandathome.Wewentthroughthecodesagainwhichresultedinhigher‐levelthemes.

Since therapists reported that they always – in some way‐ personalize thetherapy, most therapists and some patients had expected that therapistspersonalized more. One therapist thought that therapists could also haveinterpreted a strict therapy protocol application as applying the therapyprotocolinaguidelineway,meaningthattherapistsdidnotapplythedetailed


65

andprecisecontentofthetherapyprotocolsessions,butusedthesetoassistthem to make decisions about which elements of the therapy protocolsessions would be appropriate. “I think that the therapists do follow thetherapy protocol as a guideline, but that they noted this [strict application]downasapplyingitinanunchangedway.” (Therapist 5B). Themes thatwerealsoderivedfromthecodesinthesecondpart,focusedontheenhancingthetherapeutic alliance and on personalization based on the experience oftherapists.

Thefirstmainthemeof therapiststhatwasderivedfromthequotesfocusedon enhancing the therapeutic alliance. “Aligningtotheneedoftheother[thepatient]andsmalltalk[withthepatient]contributestothepersonalconnectionwith a therapist, which contributes to a more personal relationship that isneeded to create openness and that a patient accepts help from a therapist”(Therapist 4B). It was seen as crucial, in order for a patient to trust thetherapist and work together to solve the problem of the patient. Two sub‐themeswere derived: personalization based on the individual situation of apatientandkeepingorenhancingthemotivationapatient.

The first subtheme to the grand therapeutic alliance theme, focused on theindividual situation of a patient. In general, therapists first focus on theindividual situation of a patient, followed by the relevant therapy protocolsession that suited the situation best. They could also apply elements fromdifferent therapyprotocolwhenapatienthadotherpsychologicalproblems.Inthisway,patientswerehelpedwiththeirproblemsatthesametime.“Thepatientsoftenhavemultipleproblems,soyouhaveanxietyandmoodprotocols,orotherones”(Therapist4B)

The second sub‐theme to the grand therapeutic alliance theme focused onkeeping or enhancing themotivation of patients. They either did or did notapply the therapy workbook, mainly to prevent experiences of failure andkeep themotivation to adhere to therapy if a patient forgets theworkbook.“You also have to prevent that it [filling in theworkbook] becomes a failingexperience…theycanthink,wellifIcan’tevendothatwell..”(Therapist 4B). Inaddition, therapists applied motivational interviewing to enhance the


66

motivation of patients. “It ispartof theattitudeasa therapist, that youareempathic,youlistenandalign”(Therapist4B).

The therapy protocol helped therapists to structure therapy, but therapistsdiffered in their opinion regarding protocol application. One therapist fromlocationAfollowedthetherapyprotocolasstrictlyaspossible,whileanothertherapistfromlocationBonlyusedthetherapyprotocoltoguidethetherapysessions.ThetherapistfromlocationAmentionedthatthetherapyprotocolshelped to give guidance to the therapy sessions, while the therapist fromlocationBfounditmoreimportanttofocusonthesituationofapatient.

The second grand theme of the therapists focused on the experience oftherapists that influenced the amount of personalization. More experiencedtherapists often have experience with different therapy protocols sincetherapy protocols are often changed or improved over time. This increasedtheir knowledge, preferences and possibilities to personalize therapyprotocolscomparedto lessexperiencedtherapists.Therefore,twotherapistsmentioned that the experience of therapists could also have influencedtherapyprotocolapplicationandpersonalization.

Based on the quotes of the patients,wederived twomain themes. The firstgrand theme focused on personalization based on their own personalsituation.Theypersonalizedtheirhomework,influencedbypossiblerelapsesandhowtheyfelt.“Ihadtodoexposureexercisesonce,foreveryday.ButifIdonotfeelwell,itdoesnotworkandI’mnotgoingtoletmyselffeelworsebydoinganother exercise” (Patient 4B). The second grand theme focused onpersonalization based on the personal preferences of patients. How theyworkedontheirtherapyandpreparedforatherapysessiondifferedbetweenpatientsbasedonwhattheypreferredtodo.Forexample,byshuttingdownthe mobile phone when starting therapy or working on assignments on acomputerinsteadoftheworkbook.


67

Table2.Themesandsub‐themesofthefocusgroupdiscussionswiththerapistsandpatients.

Who Maintheme Sub‐themesTherapistsround1 Useprotocolasatoolbox Personalizationbasedon

whatpatientneeds

Personalizationbasedonowntherapy‐givingexperiences

Personalizationtoenhancetherapeuticalliance

Patientsround1 Personalizationbasedonownsituation

PersonalizationtobettermatchtherapywiththedailylifePersonalizationinfluencedbythevaryingmotivation

Therapistsround2 Personalizationtoenhancetherapeuticalliance

PersonalizationbasedontheindividualsituationofapatientPersonalizationtokeeporenhancethemotivationofpatient

Personalizationbasedonexperience

Patientsround2 Personalizationbasedonownpersonalsituation

Personalizationbasedonpersonalpreferences

3.4. DISCUSSIONExisting research focusing on the effect of eHealth in mental healthcaresuggests overall small to medium effect sizes [37‐41]. Moreover, researchsuggeststhatcombiningeHealthwiththerapistcontact,i.e.blendedeHealth,ismoreeffectivecomparedtofullyonlineeHealthwithouttherapistcontact[42,43].OnemainreasonforeHealthtobemoreeffective,isthatitcanextendthereachofpsychologicaltherapybeyondtheclinicalsetting,astechnologiescanbeusedanytimeandanywhere [27,28].EHealthdesigners typicallyuse thetherapyprotocolsof evidence‐based face‐to‐face therapies as abasis for thedesign of eHealth. However, not all parts of therapy protocols are always


68

appliedintherapeuticpractice[196,210].IfeHealthdesignersdonottakethisinto account, the designed eHealth might not optimally fit the existingtherapeuticpracticeandimpedeimplementationandmotivationtoadopttheeHealthbyboththerapistsandpatients.Inthepresentstudy,westudiedtheproportion, typeandreasons forpersonalizationofagiventherapyprotocolbytherapistsandpatientsinfocusgroupstudies.

Results showed that in clinical practice, the therapy protocol is not fullyapplied but also personalized (see Table 2) (also found in [58‐61]). Theavailable therapy protocol is thus just one factor in a therapeutic process.Other factors that influence the therapeutic process are the personalizationpracticesoftherapistsbasedontheneedsofapatient,motivationofapatient,therapy‐givingexperiencesoftherapists,andthetherapeuticalliancebetweenthetherapistandpatient.Therapistsestimatedthattheyonlystrictlyfollowed48%of the protocol, adapted 30%of the protocol and replaced 22%of theprotocol by other non‐protocol therapeutic parts such as other therapyprotocolelements.Otherpersonalizationpracticesthat influencetheamountoftherapyprotocolapplicationispersonalizationofpatientstobettermatchtherapywith their daily life, personal situation and preferences, and it wasalso influenced by their varying motivation. Patients estimated that theystrictly followed 29%, 48% was adapted and they estimated that theyreplaced23%ofthetherapybyothernon‐therapeuticelements.

Itisimportanttomentionthedifferenceinpersonalizationfortherapistsandpatients.Theestimationsofpatientsandtherapistsregardingtheiramountofpersonalization are not only different because theymay personalize less ormore, but also because of their own share in the personalization process.Therapistsalreadypersonalizeatherapyprotocol,ofwhichtheirpatientsalsopersonalize elements from in their daily life. Therapists know the wholecontent of the therapy protocol and patients do not. Therapists provide thepatientwithapartlypersonalizedtherapy.Therefore,patientscanneverfullyknow the whole possible content of a therapy protocol and have lesspersonalization options of the standard therapy protocol. For example,therapists often mentioned that they did not use the therapy workbook to


69

preventpatients fromexperience feelingsof failure if theyeitherdidnotdothe homework assignments or forgot to bring the workbook to therapy.However,bydoingthistheyalsopreventedpatientsfromtryingtoexecutethehomework assignments in their workbook. Besides, personalization bytherapists can have both positive, no and negative effects [211‐214]. Forexample, itmaybe that theelements thatarepersonalizedbya therapistorhowatherapistpersonalizesspecificprotocolelementsisnotpreferablewiththepreferencesofapatient.Thismayinfluencethealignmentofthetherapytoapatientandmaypossibly lowermotivationofapatient toadhereto thetherapy. In general, most therapists in the second part had expected thattherapists personalized more than was suggested by the estimatedpercentages of protocol application from therapists in the first part. Apreviousstudythatonly focusedonpersonalizationbytherapists foundthattherapistspersonalizemorecomparedtowhatwefoundinourstudy[215].Apossibleexplanationforthisdifferenceisthattheyaimedtoassessalltypesofactivities in the general psychotherapeutic practice of eating disorders,insteadof studyingpersonalizationpracticesof bothpatients and therapistsbyusingaspecificCBTprotocolinyouthaddictioncareasacaseprotocol.

TheresultsofourstudyareimportantforeHealthdesignclientsandeHealthdevelopers,sincetheyneedtoknowwhatprotocolelementsineHealthshouldand shouldnotbe open forpersonalization to facilitate implementation andpatient engagement. Designers can implement the personalization practicesby focussingon the function thatpersonalizationhas in therapeuticpractice(i.e.enhancingthemotivationofpatientstoadheretothetherapy).However,sincepersonalizationmayhavebothpositiveandnegativetherapeuticeffects,it is important to know what elements are crucial elements to apply inpracticetoenhancetherapeuticeffects.Especially,sincedesigncaninfluenceandenhancemotivationtoadhereorexecutespecificbehavior.Forexample,by applying motivating elements from entertainment games (also called“gamification”).Gamificationdesign inhealthcareandmentalhealthcarehasshownpotential[71,77,78],e.g.byimprovinghealthybehavior[78‐89].Basedontheresultsofthisstudy,eHealthdesignersarerecommendedto:a)studyandcopyatleasttheactualappliedpartsofatherapyprotocolineHealth,b)


70

co‐designeHealthwiththerapistsandpatientssotheycanallocatethepartsoftheeHealth thatshouldbeopen forusercustomization,andc) investigate ifparts of the therapy protocol that are not actually applied by therapists orpatients shouldbepartof theeHealth.Otherwise, implementationwouldbenegatively influenced, i.e. because the eHealth does notmatch the habits oftherapists [199] or complexity ofmental problems that patients experience[34]. Besides, validation studies of therapy protocols need to focus on theactual application of these protocols in therapeutic practice, as it can beconsideredasgenerallyoverestimated[198,216,217].Thismayoverestimatethe benefit of therapy protocols to therapeutic effects. In the followingparagraphwewillelaborateonthethreerecommendations.

With regard to the first recommendation, our study showed that therapistsandpatientsdonotfullyapplythetherapyprotocol.Thisinformationshouldbe generated and implemented in the second product design phase of aPersonalized Design Process [204]. In this phase, stakeholders such astherapists, patients and protocol developers can be involved to make thedesign of the product suitable to support the user during therapy and toensure that it is technically possible to use the eHealth application duringtherapy and that the design of eHealth suits the therapeutic practice of atreatment centre. In this phase, the information of the applied therapyprotocol elements by therapists and patients is generated so eHealthdesignerscanatleastcopythesepartsineHealth.TheeHealthdesignerscangenerate this information by, for example, recording therapy sessions ofpatients with therapists. Therapy protocol developers can listen to theserecordings and rate what parts of a therapy protocol are applied intherapeuticpractice.

Asasecondrecommendation,theresultsofourstudyshowedthattherapistsandpatientspersonalizedthetherapyprotocolbyadjustingspecificprotocolparts and adding other (non‐protocol) therapeutic parts. Why and howtherapists and patients personalize is important information for eHealthdesigners, to select those parts in eHealth that should be open topersonalisationfortherapistsandpatients.Thisinformationcanbegenerated


71

andimplementedinthelasttailoringphaseofaPersonalizedDesignProcess[204]withpatientsandmoreand lessexperienced therapists. In thisphase,thedesignedproductistailoredtotheindividualuser.Itconsistsoftwotypesof tailoring: user controlled customization and use‐dependent adaptation.Withusercontrolledcustomization,auserhimorherselfcantailoraproductto own preferences and needs. Patients noted that they personalized thetherapybasedontheirownpersonalsituationandpersonalpreferences,anditisthusalsoimportanttogivethemtheopportunitytodosointheeHealthproduct.Therapistsmentioned topersonalize the therapyprotocolbasedonthepatientsituationortheirtherapeuticexperiences.Byprovidingtherapiststhepossibility to tailor theelements ineHealth, theycanchoosewhetherornot to use these during therapy with a specific patient. Especially by notforcing them to use all elements of the eHealth application. With use‐dependent adaptation, a product automatically adapts itself to the user. Forexample,bynot showingspecificpartsof a therapyprotocol if the therapistalways skips these in therapeutic practice or by tailoring the momentsreminderspop‐uptoapatientwhoalwaysexperiencescravingsafterdinner.

Asathirdrecommendation,werecommendeHealthdesignerstoinvestigateifthere are parts of the therapy protocol that are not actually applied bytherapistsorpatientsbutshouldbepartoftheeHealthsincetheyarecrucialfor theeffectof therapy.TheeHealthdesignercangenerate this informationby interviewing therapy protocol developers about the crucial therapyprotocolparts.ThisinformationcanbegeneratedbyinvolvingstakeholdersinthesecondproductdesignphaseofaPersonalizedDesignProcess[204].Forexample, to allow the therapist to use the eHealth application as a toolbox,suchastheyusethetherapyprotocol,butensuringthecrucialelementsnottobetooeasilypersonalizedorskipped.

Our study has two limitations. The first concerns asking therapists andpatients to quantify their own behaviour. It may be challenging for both toquantify this themselves. Other research also found that therapistsoverestimatedtheextentoftherapyprotocolapplication[218]orthat ithadthepoorestreliability[219].Forexample,notallrespondentsunderstoodthe


72

assignmentastheindicatedpercentagesofstricttherapyprotocolapplicationof a patient and therapist overlapped with their other percentages. Thisoverlapping is impossible, e.g. as one cannot strictly follow and change atherapyprotocolatthesametime.However,askingtherapistsandpatientstoquantify their ownbehaviourmay still be a suitable techniquewhen askingthem only to estimate the amount of therapy protocol application andpersonalization. It is thus a suitable technique to generate first insights, butnottosolelybaseresultsonthistechnique.Asecondlimitationisthatwedidnot take the therapeutic experience of the therapists and severity of thepatients’ condition into account [220]. Compared to less experiencedtherapists,moreexperiencedtherapistsgenerallyhavemoreexperiencewithothertherapyprotocols.Thismayinfluencetheirpersonalizationpractices.Inaddition, it is possible that the severity of a patient’s condition could haveinfluencedrecruitmentandresults.Besides,thisstudywasconductedwithalimited amount of participants. This might have enhanced the possibleinfluence of individual preferences regarding protocol application andpersonalization on the results [221]. Future research should take this intoaccount, e.g. by conducting the studywith a larger sample sizewhile takingintoaccountthesebackgroundvariables.Inaddition,whenfutureresearcherswant to design a toolkit they should keep in mind to also involve actualeHealth designers, eHealth design employers and researchers. This isimportant, since the toolkit may otherwise not correspond with currentpractices of these target groups which would negatively influenceimplementationofthetoolkit.

3.5. CONCLUSIONTo optimize eHealth implementation, our study indicated that eHealthdesigners shouldknowwhich therapeuticparts shouldbeduplicated,whichpartsshouldbeopentopersonalizationpossibilities,andwhichpartsthatarenot applied in practice should be part of the eHealth design. In order togenerate this information,we suggest eHealth designers to collaboratewiththerapists, patients, protocol developers, and mental healthcare managersduring the design process of eHealth [204]. Not involving all thesestakeholders enhances the chance that the designed eHealth might not


73

optimally fit the therapeutic practice and impede implementation. Forexample,therapyprotocoldesignersknowwhatprotocolpartsarecrucialforthe therapeutic effect but do not know how protocols are applied andpersonalized in therapeutic practice. Personalization practices can beimplemented by actively co‐designing with patients and more and lessexperienced therapists, to ensure that it is aligned to their preferences andcapacities. Based on the presented research, we expect that theimplementation of eHealth can be facilitated when stakeholderrepresentatives, e.g. patients, therapists, protocol developers and mentalhealthcare managers, are involved in the design process by providing theeHealth developer with their needs and demands of therapy protocolapplicationandpersonalization.

74

Basedon:vanDooren,M.M.M., Spijkerman,R.,Goossens,R.H.M.,Hendriks,V.M.,&Visch,V.T. (2014,October).PLEXas inputandevaluationtool inpersuasivegamedesign:pilotstudy. InProceedingsof thefirstACMSIGCHIannualsymposiumonComputer‐humaninteractioninplay(pp.449‐450).

GAMEDESIGNRELEVANCEOFPERSONALIZATIONINYOUTH

MENTALHEALTHCARE

Tomotivate therapists or patients to use eHealth, game elements are oftenapplied to eHealth aiming to encourage interaction and to facilitate theachievement of aimed‐for real‐world goals such as behavioral change.Personalizedgamificationcanenhancemotivationofusersenhancemotivationtokeepinteractingwithaproduct.Thischapterdescribesadesignmethodusinga specific personalization technique, where cards were used that representplayfulexperiences,toexaminewhetherthe inputofplayfulexperiences isalsoexperiencedbyotherend‐usersfromthesamecontextintheactualdesignitself.

Chapter4A|Gamedesignrelevanceofpersonalizationinyouthmentalhealthcare

76

4A.1. INTRODUCTIONThis pilot study is part of a larger project on implementation guidelines forpersuasivegamedesigninyouthaddictioncare:howtoalignthedesigntotheuser and how to implement it to the user context. Persuasive Game Designaimstocreateauserexperiencedgameworldtochangetheuserbehaviorintherealworld[63].

Inordertoadjusttheuserexperiencebygamedesign,itisnecessarytoalignuser’sgame‐relatedexperientialpreferences.Inordertoverifyifthedesignofagameismatchedtotheuser’spreferences,itisimportanttouseatoolthatgeneratesandevaluatestheinputfromusers.Theplayfulexperiences(PLEX)framework[72]isoneexampleofsuchatool.Itconsistsof22PLEXcardsandhas two proposed design techniques for the user research input phase: a‘brainstorming’‐anda‘scenario’technique[222].Also,PLEXcardshavebeenusedwithoutthesetechniquesintheevaluationphase[223],butnotinbothphasestogether.

Currently, it is not clear yet whether one specific tool can be used forverification in the design process by applying it in both the user researchinput‐ andevaluationphase. In this study,PLEXexperienceswereusedas atooltomapgameexperiencesinbothphases.

4A.2. METHODSThisstudyconsistedofthreephases:userresearchinputphase,designphaseandevaluationphase. In theuserresearch inputphase itwasstudiedwhichPLEX cards motivated and which PLEX cards did not motivate addictedyoungsterstocontinueplayingagameforalongerperiodoftime.Youngsters(N=7), intreatmentatMistraladdictionclinic,participated inthisphaseanddifferedinageandcomorbidproblems.MistralispartofBrijder,anaddiction‐careorganizationintheNetherlands. It isanopenclinicthatprovidesgrouptreatment for approximately 14 youngsters who have stopped usingsubstances. The participants received the PLEX cards, which presentedmotivating game experiences, and were asked to divide the cards into‘motivating’ or ‘not motivating’ cards. In the design phase, a professionalexternalgamedesignerwasaskedtodesigntwoprototypegames:oneaiming


77

toelicitthemostmotivatingPLEXexperiencesandanothertoelicittheleastmotivating PLEX experiences In the evaluation phase another group ofyoungsters (N=5) in treatment atMistral, evaluated both prototypes on thepresenceofPLEXexperiences.

4A.3. RESULTSIn the user research input phase, the most motivating experiences were“competition” and “thrill” and the leastmotivating oneswere “nurture” and“suffering”.Themostmotivatingexperiencesresultedinthepaperprototype“EvolutionBattle”,where10organismsoncards,e.g.,humanandrankedfrom1to10,battledforasurvivalofthefittest.Thegamewasplayedinroundsof10seconds.Thefirstroundwasstartedwithbettingonecoin.Ineachround,aplayer could raise his/her bet, swap the organism or pass. After a pass theotherplayerhadoneroundtoswap,raiseoralsopass. Intheend,thecardswith theorganismswerecompared.The fittest: theplayer thathad thecardwiththehighestnumber,receivedallthebetsplaced(seeFigure1).

Chapter4A|Gamedesignrelevanceofpersonalizationinyouthmentalhealthcare

78

The paper prototype based on the leastmotivating experiences, resulted in“FallingAngels”.Angels fell from the skyonto aworld filledwithplatforms.The platforms contained all kinds of danger which made the angels suffer.Angels had to be guided safely towards heaven. Players could protect theangelswithhelpfromthe“handsofGod”representedbythreesymbols,i.e.afist,anopenhandandapointingfinger(seeFigure1).

To evaluate the game experiences of the paper prototypes, five youngstersplayedandratedbothprototypesbymeansofPLEX.Inlinewithexpectations,our results showed that the prototype based on the most motivating PLEXexperiences was preferred by four out of five participants. However, otherexperiencesthanthosederivedfromtheuserresearchinputphasewerealsoreported. Furthermore, in the Falling Angels prototype, participantsexperiencedotherPLEXmorestronglythantheones fromtheuserresearchinputphase(seeTable1).


79

4A.4. DISCUSSIONOurpreliminary findingssuggestthatusingmotivatingPLEXexperiences forgamedesignresultedinabettertailoredprototype,sincetheprototypebasedonthemostmotivatingPLEXexperienceswaspreferredbyamajorityoftheparticipants. However, the PLEX experiences derived in the user researchinputphasedidnotcorrespondone‐on‐onewiththeexperiencesreportedinthe evaluation phase, problematizing the application of PLEX as a generaldesigntoolforexperience‐basedgamedesign.

A possible explanation could be that the PLEX experiencesmay bemultipleinterpretable and can show overlap [3]. Furthermore, a game designermayinterpretthePLEXexperiencesdifferentlyincomparisontonaïveusersofthegame. Finally, designersmight base their choices in the design process notsolelyoninformationaboutgameexperiencesderivedfromtheuserresearchinputphase.

More research is needed to drawmore in depth conclusions on PLEX as apossible tool in both phases and to verify if the two prototypes differsignificantlyonmotivation.

Chapter4B|Thedesignandapplicationofgamerewardsinyouthaddictioncare

80


Based on: van Dooren, M. M. M., Visch V. T., & Spijkerman R. (2019), The Design and Application of Game Rewards in Youth Addiction care, Information, 10(4), 125.

THEDESIGNANDAPPLICATIONOF

GAMEREWARDSINYOUTHADDICTIONCARE

In the previous chapter, we described a design method and found that theexperiencesderived inthe inputphasedidnotcorrespondone‐on‐onewiththeexperiencesreportedintheevaluationphase.Inthischapterwewilldescribetheeffectofaspecificdesignelement:rewards.Rewardsarethemosttypicallyusedgame‐elements to fostermotivation in entertainment gaming. However, it isunclear whether game‐rewards are also effective in a healthcare context.Especially inourtargetgroup, i.e.youngsterswithsubstance‐relateddisorders,since these typeofpatientsmaybe less sensitive tonon‐drug‐related rewardsthanpatientswithoutasubstanceusedisorder.Therefore,wefirststudyiftherearemotivationaldifferencesbetweendifferenttypesofgame‐reward.Secondly,we study the differences in reward type preferences between youngsters intherapy for substance dependence and youngsters without a substance usedisorder.Concluding, theaimof this chapter is to investigate ifwe canapplygame rewards in this population and if personalization of rewards wouldfacilitateimplementation.


82

4B.1. INTRODUCTIONAn evidence‐based therapeutic strategy to motivate substance dependentindividualstoremainabstinentistoaddmonetary‐basedrewardstoevidenceofsuccessfulbehavioralchange,e.g.,substance‐freeurinetests[224].Rewardscanbeseenasstrongmotivatorstoinfluencebehaviorchange[225‐228]andareacrucialaspectinthedesignofentertainmentvideogamestoenhancenotonly feelings of enjoyment and flow [229] but also feelings of mastery,autonomy,andasenseofbelonging[230].Because it takestimeforuserstointend,start,andmaintainbehaviorchange,gameelements,suchasrewards,haveoftenbeenusedasmotivational ingredients inPersuasiveGameDesign(PGD)[226,231,232].

The aim of PGD is to transport the users’ real‐world experience towards a(partial) game world experience that is more enjoyable and engaging thanreal‐world experiences [63, 233, 234], thereby enhancing persistence ofspecificdesiredbehaviorintherealworld,knownasthetransfer‐effect[235,236].Researchhas shown that applying gameelements in a serious contextcanpositivelyinfluencehealthrelatedproblemsandbehaviors[83,237],suchasanxietymanagement[238],physicaltherapeuticexerciseandfitness[239,240], burn pain management [241], diabetes [242], and asthma [243].However, research has also suggested that applying game elements in aseriouscontextcanreduceoverallengagementandintrinsicmotivation[244]or lead to unintended effects that distract players and lowers the overalleffectivenessof an intervention [245].Most importantly, gameelementsuseextrinsicrewards,suchaslevelsandpoints,toenhanceengagementofusers,while striving to enhance users’ general feelings of competence, autonomy,and a sense of belonging and connectednesswith others [233]. These threeelements form the basic human psychological needs that facilitate users’motivation,bothintrinsicandextrinsic,toexecutespecificbehavior[230].

Thus,PGDseemstobefruitfulforenhancingpositivehealthcareeffects,sinceithelpsplayerstoaimforagiventargetexperienceorbehavior.Crucialinthepersuasive effect of a gamedesign is the choice of the used game elements.Theseare theelementswithinagamethat functionascoremotivators fora


83

play experience, such as a challenge in platform games, social teaming insoccersportgames,orexplorationinrole‐playinggames.Among thesegameelements, rewards are one of themost applied kind of elements. Sometimesrewardsaredesignedasacoregame‐elementinagame,suchasthemonetaryrewards in gambling, and sometimesas a supportivegameelement, suchastheweaponsandpowersyoucanearnasarewardforcompletingachallengeinMMORPGs.Althoughmotivationaleffectsofrewardsindailylifehavebeenstudied extensively in psychological and neurocognitive studies [236, 246‐250],thereissurprisinglylittlefundamentalresearchaboutthemotivationaleffectsofrewardsingames.

Ingames,rewardsaremosttypicallyappliedintheformofmonetaryrewards,virtualpoints,andsocialrewards[228,251].Thesethreerewardtypesdifferin their value of use. Monetary rewards have a dominant value in the realworldoutsidethegame.Incontrast,virtualpointshavetheirdominantvaluewithin the game world, and social rewards, such as received complimentsaboutyourgameplaybyyourplaymates,haveavalueinboththerealworldandthegameworld[252].Monetaryrewardsconsistofatangibleamountofmoney that a player receives for a specific performance [253, 254]. Virtualpointsareusedasascoringsystemorasawaytobuyvirtualgoodsthatareusable in the game (e.g., better weapons). Scoring systems based on theearned player points are often a symbolic way of reflecting the players’progression, performance, achievement, and competence [253]. In socialrewards,playersgiveandreceivecomplimentstoandfromotherplayers,ortheyinviteandareinvitedtojoinspecificplayergroups.Thistypeofrewardincludes positive incentives related to the general human need of feelingrelatedtoothers[253]andreceivingsocialrecognitionforspecificbehaviors[255, 256]. From a neurocognitive perspective, preliminary findings fromFunctional Magnetic Resonance Imaging (fMRI) research suggest that thesethree reward‐types may activate specific areas in the brain [257]. Forexample, brain areas that have been linked to the processing of self‐relatedand social information showed more activation when social rewards weregainedthanmonetaryrewardsorperformancefeedback,suchaspoints.


84

In the present paper we will investigate, for the purpose of serious gamedesign,themotivationaldifferenceofthethreebasictypesofgamerewards:monetaryrewards,socialrewards,andvirtualpoints.Theapplicationofgameelements (such as game rewards) in a non‐entertainment (i.e., “serious”)contextiscalled“gamification”[182].Inordertostudytheapplicationvalueofgame rewards in serious contexts comprising specific user groups, weinvolved (a) adolescent patients with substance use disorders from asubstanceaddictioncarecontextand(b)asame‐agedcontrolgroupofhigh‐school students without substance use disorders. The context of substanceaddiction therapy might benefit from the study of persuasive game designinvolvingrewards,sincereward‐basedgamebehaviorandsubstance‐usebothderive their motivation from shared neurological dopamine systems. Morespecifically, video gaming is associated with dopamine release, and alladdictive substances trigger increases in dopamine in a key region of thereward (limbic) system in the brain [258, 259]. Additionally, adding game‐elementstoanaddictiontherapymightmakethetherapymoreengagingforpatients, and hence enhance the therapeutic adherence [260]. While game‐rewardsmaybeparticularlymotivatingforadolescents,itisnotclearwhetherthis also holds for adolescents with substance use disorders. Neurologicalfindingssuggestthat—comparedtonon‐dependentpersons—theapplicationofrewardsmayhavelessimpactonsubstancedependentindividualsduetoahyperactive dopamine system for psychoactive substances (alcohol,amphetamine, opiates, or marijuana) and a decreased sensitivity to stimulithat are not related to these substances [259, 261‐266]. This “dampened”effect of non‐substance related rewards in substance dependent personsinforms our hypothesis that game‐rewards may have a lower motivationaleffect in this population than in a non‐dependent high‐school population.Althoughwedohaveevidencethatrewardscanworkintheclinicalpracticeofaddiction treatment—particularlywhenusingmonetary incentives followingan evidence‐based contingency management scheme [267‐270]—neurocognitive findings indicate that natural rewards may have a lowerimpact on this population. It is unclear whether substance dependentindividualswillbesufficientlymotivatedbygamerewards,sincethistypeof


85

individualmaybemorestronglymotivatedbytheexpectedrewardingeffectof substance use. To determine which types of rewards are suitable forpersuasive games aimedatpatients inmental health care, it is important toempirically test the potential impact of game rewards for specific patientgroups, suchas individualswith substanceusedisorders.Thepresent studywill,thus,focusoncomparingtheeffectsofthethreebasicseparatedrewardtypes between a clinical sample of substance dependent adolescents and acontrol group of non‐dependent high‐school students. Because substancedependentindividualsmayshowdecreasedsensitivitytorewards[259,261‐266].

Wehypothesizethatallseparaterewardtypeswillbelessmotivatingforthemcomparedtotheirnon‐substancedependentcounter‐parts.

4B.2. METHOD

4B.2.1. ETHICS TheMedicalEthicalCommitteeoftheLeidenUniversityMedicalCentreintheNetherlandsgrantedexemptionforafullethicalapplication.

4B.2.2. PARTICIPANTS Participants(agedbetween12–24years)wererecruitedfromtwolocationsinthe Netherlands. A total of 32 non‐substance dependent adolescents wererecruited from a secondary school and36 substancedependent adolescentswere recruited from a substance addiction care facility. Due to computerproblems during the test, we had to exclude 23 participants (16 substancedependent and 7 non‐dependent adolescents). Approximately 50% of theseparticipants (N = 11) did not play the game for all three types of rewardsbecauseofsoftwareproblems.Theother50%of theseparticipants(N=12)unwillinglypressed the stop‐buttonwhileplaying, even though theydidnotwant to stop playing the game. At the start of the experiment we clearlyexplainedtoparticipantsthattheycouldpressthestop‐buttoniftheywantedtostopplayingthegame(seeFigure1).Thiswasimportantforouranalysis,since thestop‐buttonwasdirectlyrelatedwith thedependentvariable “playpersistence”. However, since participants pressed the stop‐button even


86

thoughtheydidnotwanttostopplaying,eithertheydidnotunderstandthisexplanationortheypressedthebuttonbyaccident.Whenparticipantsdidnotplaythegameforallthreetypesofrewards,wehadtoexcludethemfromthewholestudyaswecouldnotcomparetheirplaypersistenceforthedifferenttypesofrewardsanymore.

The final study sample consisted of 45 participants, with 20 substancedependent and 25 non‐dependent adolescents. The group of substancedependent adolescents contained fewer females (15%) compared to non‐dependent adolescents (52%), matching the general substance dependencepopulation that also consists of more males [271, 272]. We did not collectpersonal informationregardingthe typeofsubstancedependence,since thiswasnotthefocusofthestudy.Inaddition,itwasoftencomorbidandaskingfor this information might have decreased the participants’ motivation toengage in theplayteststudy.Adolescents inDutchaddictioncaremostoftenreceive therapy for cannabis, alcohol, and gaming. A smaller group receivestherapy for simulants (mainly amphetamine, but also cocaine or ecstasy)[273‐275].Wetriedtomatchtheageofbothsubstancedependentandnon‐dependent groups. The average age of the respondents from the secondaryschoolwas around16 years old (14–18 years old), andpatientswho are intherapyat theyouthaddictioncare clinicaregenerallyaround18yearsold(12–22yearsold)[274].

4B.2.3. DESIGN Pertypeofreward,participantswereabletospendamaximumof40minutesplaying the game. If participants used the maximum playing time with alltypesofrewards,themaximumplayingtimewouldbetwohours.Participantsincurrentstudyplayedthegamewithalltypesofrewardsinatotalof30–60min. The game consisted of a four by four grid with 16 buttons. Of thesebuttons, 8 randomly displayedmultiplications of 2 up to 9, and the other 8displayedpossibleanswersofthemultiplicationproducts.Ofthese8possibleanswers of the products, 6 matched the outcomes and 2 were incorrect.Participantswereinstructedtomatchamultiplicationproduct,andaftereachmatchthescreenrefreshed.


87

Participantsreceived incrementalrewardsafteraspecificnumberofcorrectanswers (after 3, 6, 9, 12, 15, 18, 21, 24, 27, 30, 33, 36, 39, and 42 correctanswers). The screen showed how many correct matches were needed toobtain the next reward and howmany rewards the participant had alreadyearned. Participants could, thus, earn a total of 14 rewards per session andcomplete amaximumnumber of 315 products if they played themaximumplaytimeandalwaysansweredcorrectly.Duringthewholegame,thescreenshowed a “stop‐playing” button at the top of the screen. This provided theparticipantwiththepossibilitytostopplayingthegameatanymomentwhenpreferred(seeFigure1).Afterhitting the“stop‐playing”button,anewgamestarted with similar exercises but with another randomly chosen differentrewardtypeuntiltheplayerhadplayedforallthreerewardtypes.Attheendof the study, all participants received 10 euros for their participation,regardless of their score in the game. The participants were not informedabouttheparticipationfeebeforehand.

Intotal,participantsplayedthreegame‐sessions. Ineachsessiontheywouldplayforoneofthethreerewardtypes:monetaryrewards,virtualpoints,orasocial reward (see Figure 2). Regarding the monetary reward, participantscould receive 50 Eurocent per rewarduntil they reached a total of 7 Euros.Theyreceived thisreward typeafter thestudy.Regarding thevirtualpoints,participantscouldreceive5pointsperrewarduntiltheyreachedamaximumof70points.Thethirdrewardconsistedofasocialreward,whereparticipantssawapop‐uppictureofarandomlyselectedblurryface,withathumbsupandatextualcompliment.Theblurryfacesweretakenfromapoolofportraitsofparticipantsofthestudythatwephotographedbeforestartingthestudy.Forethical considerations we blurred the photographs to the extent that facesknown to the participant were recognized but faces unknown to theparticipant were not. Participants received one compliment per rewardmoment, which could vary according to five different kinds of texts: “Welldone!”,“Wonderful!”,“Howsmart!”,“Calculationtiger!”,“Thumbsup!”.


88

Figure1.Thetablet‐basedgameshowingthegame‐tasktocombinemultiplicationsandoutcomes.Translationoftextinredbuttonupperright:“EndingtheGame”.Textbelowstates:“Getanother5correctcombinationstoreceive5morepoints!You

alreadyhave:25points”.

Figure 2. Examplesofthreetypesofrewards(translationfromtoptobottom:

“Youearned5morepoints”,“Youearned50moreEurocents”,“Natasha,howsmart!”).


89

4B.2.4. VARIABLESANDMEASURES Astheindependentvariableweusedthetypeofreward(monetaryrewards,virtualpoints,andsocialrewards,)andrewardevaluationwasconsideredasthedependentvariable.Weusedtimeinminutesthatusersspentplayingthegame,untiltheyhitthe“stop‐playing”button,asameasureofplaypersistence(timespentplayingasameasureofpersistencewasalsoused inapreviousstudy [276]). Participants could evaluate the reward by answering thefollowing four questions on a five‐point Likert scale (0 (= totally disagree),until (4 = totally agree)): (1) “I did notwant to quitwhilewinning/earning“therewardtype”(onlyfillinifyouhavestoppedbeforetheendofthetest)”;(2) “Iwanted to continueplayingbecause of “the reward type”; (3) “I thinkthat“therewardtype”isagoodreward”;(4)“Iamhappywiththeamountof“therewardtype”Ihavewon”.

4B.2.5. PROCEDUREParticipants first providedwritten informed consent for studyparticipation,afterwhichtheyreceivedaniPadforuseinthepresentstudy.Atthestartofthegame,participantsfilledintheirnameandwereinstructedasapracticetofirst completeasmanymultiplicationsaspossiblewithin twominutes.Afterthat, participants received information about how the game worked. Theywere also given the opportunity to askquestions if anythingwasunclear. Iftherewere no questions or all questionswere answered, the game started.After the third and last game‐session, participants were asked some finalquestions about playing the game in general. For each respondent, all gamesessions took place during one session, in which the order of the types ofrewardswasrandomized.

4B.3. RESULTS

4B.3.1. STRATEGYOFANALYSISAll analyses were conducted in SPSS version 22. Since the data were notnormally distributed, as shown by a Kolmogorov‐Smirnov test, we appliednonparametric tests. Without the first item, Cronbach’s Alphas for theevaluation of monetary rewards, virtual points, and social rewards wererespectively.84,.85,and.82.


90

4B.3.2. MANIPULATIONCHECKS According to the nonparametric independent samples tests there was nostatistically Bignificant difference between our control variable “order ofreward type”and timeparticipantsplayedwiththerewards(allp‐values>=.20).Furthermore, therewas no significantdifferencebetween the “orderofreward type”and rewardevaluationofall typesofrewards(allp‐values>=.29).

4B.3.3. DIFFERENCEINPLAYINGTIMEACCORDINGTOREWARD

TYPESBETWEENSUBSTANCEDEPENDENTANDNON‐DEPENDENTADOLESCENTS

TotestdifferencesinplayingtimeweconductedaGEE‐analysis(GeneralizedEstimating Equations), including playing time as the dependent variable,reward type as thewithin‐subject variable, group (substance dependent vs.non‐dependent) as a factor, and gender as a covariate (see Table 1). Thesignificant effect of reward type (monetary, social, and virtual points)indicatedthatadolescentsplayedlongerformonetaryrewards(M=24.35,SD=11.39)comparedtosocialrewards(M=9.30,SD=10.22)orvirtualpoints(M = 12.06, SD = 11.15). Results further showed significant effects for thefactor group (substance dependent vs. non‐dependent) (X2 = 13.77, p <.05)and the covariate gender (X2= 11.84,p< .05). Regardless of type of rewardandgender,adolescentswithsubstancedependence(M=18.14,SD=14.23)played longer compared to non‐dependent adolescents (M = 12.91, SD =10.84).The significant effectof gender suggested that girls (M=16.83,SD=12.27)playedlongercomparedtoboys(M=14.36,SD=12.89),regardlessofgroupandtypeofreward.

Table1.Regressionestimatesforplayconsistency.

B SE WaldX2(95%CI) SigVirtualpoints −12.29 1.91 41.37(−16.04to−8.55) 0.000Socialrewards −15.05 2.26 44.40(−19.47to‐10.62) 0.000

Monetaryrewards . . . .Substancedependent 7.20 1.94 13.77(3.40to11.01) 0.000

Non‐substancedependent . . . .Gender 5.35 1.55 11.84(2.30to8.40) 0.000


91

4B.3.4. DIFFERENCEINREWARDEVALUATIONACCORDINGTO

REWARDTYPESBETWEENSUBSTANCEDEPENDENTAND

NON‐DEPENDENTADOLESCENTSInasecondGEE‐analysiswithrewardevaluationasadependentvariable,wetestedtheeffectsofrewardtypeandgroupwhilecontrollingforthecovariategender.Thetypeofrewardwastheonlysignificantvariable(X2=30.61,p<0.05). Adolescents evaluated playing for monetary rewards (M= 3.02, SD =1.00)significantlymorepositivelythanplayingforvirtualpoints(M=2.22,SD=1.03)orsocialrewards(M=2.35,SD=1.03)(seeTable2).

Table2.Regressionestimatesforrewardevaluation.

B SE WaldX2(95%CI) SigVirtualpoints −0.88 0.17 28.38(−1.21to−0.56) 0.000Socialrewards −0.66 0.16 16.64(−.97to−0.34) 0.000

Monetaryrewards . . . .Substancedependent −0.02 0.19 0.01(−0.39to0.36) 0.93

Non‐substancedependent . . . .Gender −0.04 0.18 0.05(−0.40to0.31) 0.82

4B.3.5. GENERALRESULTS Resultsshowedthat therewasastatisticallysignificantdifference inplayingtimeaccordingtorewardtypes.Participantsplayedsignificantlylongerwhenthey were playing for monetary rewards compared to the other types ofrewards. In addition, there was a statistically significant difference inparticipants’ reward evaluations of the game according to reward type, andparticipants evaluated playing for money more positively compared to theother types of rewards. When comparing substance dependent and non‐dependentparticipants,resultsshowedthatsubstancedependentparticipantsplayed longer compared with non‐dependent participants. In addition,regardlessof typeof reward, femaleparticipantsplayed longercompared tomaleparticipants.


92

4B.4. DISCUSSIONInthepresentstudywetestediftheeffectsofthreetypesofrewards(social,monetary, andvirtual)ongameplayduration, andgameevaluationdifferedbetween substance dependent versus non‐dependent adolescents.Adolescence isaperiod in life that is characterizedby increasedrisk taking,resulting from an overactive reward system in the brain [277], relative tochildhood and adulthood [278, 279]. Therefore, rewards may have anincreased motivating effect on adolescents and can be used as a usefulincentive.However, itwasunclear if,andwhich,game‐basedrewardswouldwork in a substance addiction therapy context, based on the link foundbetween a hyperactive dopamine system and a decreased sensitivity tonaturalrewardsinsubstancedependentindividuals[259,261‐266].

Our findings suggest that rewards can successfullymotivate both substancedependentandnon‐dependentadolescentstocontinuetheirinteractionwithagame.Whenusersinteractmoreorforlongerwithagame,itismorelikelythat the transfereffectof thegamewillbeachieved.Therefore,our findingsconfirmthatrewardsmaysuccessfullybeappliedinpersuasivegamedesignfor both substance dependent and non‐dependent adolescents to enhancemotivation for tasks (e.g., therapy adherence). However, this study onlyfocusedontheeffectsofrewardsonserioustasksandnottherapeutictasks.Withserioustasksthereisadirectinteractionbetweenrewardsandbehavior,butwith therapeutic tasks thepointof impactgenerally takesmore time. Inaddition, in persuasive game design for therapeutic tasks it is needed tocarefullymatchtherewardswiththedesiredtransfereffectinordertoavoidconfounding conflicts between the two, and to also study contributions tolong‐termtherapyeffects.Thisstudyshowsthatwithserioustasks,rewardsare suitable to enhance motivation to continue interaction with a product.More research is, however, needed to see if rewards are also effective fortherapeutictaskswithamorelong‐termeffect.

Our results further indicate that when receiving rewards, substancedependent adolescents played significantly longer than non‐dependentadolescents.Bothgroupsofadolescentsdidnotdiffer inhowtheyevaluated


93

the reward types. Overall, adolescents evaluated monetary rewards morepositively compared to the other types of rewards. An explanation for thismightbethatgameplaydurationwasevokedbyother(perhapsunconscious)processesortriggersthatwerenotstronglylinkedtotheexplicitevaluationofall three types of game rewards. For example, substance dependentadolescentsmayexperiencetheirclinical“realworld”contextaslessexcitingandplayfulthanhownon‐dependentadolescentsexperiencetheirnon‐clinicalrealworld. In terms of the persuasive gamedesignmodel [63], the startingposition of the participants with substance dependence might, thus, bepositionedmoretowardstherealworldthanthestartingpositionofthehigh‐schoolparticipants.Thisdifferencemightinfluencethemotivationaleffectofthe designed mathematical game in transporting the user’s experiencetowardsagameworld.Themotivationaleffectofagamemightbestrongerina less playful environment than in an already playful environment. Futureresearch has to be conducted to investigate this relationship betweenexperiencedrealworldposition,effectofgame,anditsresultinggameworldexperience.

Thefindingthatparticipantswithsubstancedependenceplayedlongerforthetypes of rewards was contrary to our expectations. We expected thatparticipantswith substance dependencewould play shorter for any rewardtype during the experiment, as research showed that substance dependentindividualshaveanoveralldecreasedrewardsensitivity[259].Thisprevioushypothesis was confirmed in previous research by Kim et al. (2014), whocompared themotivational effects of similar reward types, i.e., performancefeedback, social rewards, andmonetary rewards, between internet addictedadolescents and non‐addicted adolescents. The outcomes of this particularstudydidsuggestadecreasedsensitivitytogamerewardsinparticipantswithan internet addiction compared to non‐addicted participants [265]. Ourfinding on the impact of monetary game rewards are in line with previousresearchshowingthatmonetaryincentiveshavesuccessfullybeenappliedinsubstance abuse therapy [267‐269]. For virtual points and social gamerewards our findings cannot be confirmed by previous clinical research,


94

althoughsome formsofevidence‐based therapiesdoapply to these typesofincentivestoreinforcenon‐drugrelatedactivities.

Thisstudyhassomelimitationsthatneedtobementioned.First,wedidnotdifferentiatethegroupofsubstancedependentadolescentsaccordingtotheirmaintypeofsubstanceproblem,e.g.,alcohol,cannabis,orstimulants,nordidwedifferentiategroupsaccordingtospecificpersonalitycharacteristics.Somestudies have shown that different player groups, i.e., groups with differentpersonality dimensions, canbemore interested in, ormotivatedby, specificgame‐rewards than others [280]. Since studies have found that somepersonality traits aremoreassociatedwith substance addiction thanothers,more research is needed to further explore this topic [281]. Secondly,althoughweknewtheagerangeofpatientsthatwereadmittedtotheyouthaddictioncarefacility,wedidnotrecordtheageofthosewhoparticipatedinour study and could not control for age as a covariate in our analyses. Inaddition,itisimportanttotakeintoaccounttheout‐gamevalueofrewardsforusers.Futurestudiesshouldfocusontheneedforpersonalizingrewardsandwhetherdifferentplayertypes,personalitytraits,andtypesofsubstancesarelinked to reward sensitivity [282, 283]. Secondly, althoughwe tried to keepthe intensity of the three reward types comparable, i.e., either onecompliment,5points,or0.50Eurocentsperreward, it isnotcertainthatwesucceeded in this. It is possible that participants’ reward experiences wereaffected by how the rewards were designed [284]. Future studies couldaddressthisissuebytestingamoresophisticateddifferentiationintypesandintensityofrewards.

4B.5. REWARDSINPERSUASIVEGAMEDESIGN:IMPLICATIONS

The present study investigated if game‐based rewards can be used asmotivating game‐elements in a persuasive game for adolescents with asubstanceusedisorder.Theresultsturnedouttobepositive,sincethetypesof rewards motivated substance dependent adolescents in addiction caremore compared to non‐dependent adolescents in high‐school. Thus, apersuasive game designer developing eHealth for an addiction care context


95

canconsiderusingrewardstomotivatepatients.However,howrewardscanbestbeapplied inapersuasivegamedoesnot follow fromourstudy. In thepresent section, we will provide suggestions for reward inclusion inpersuasivegames.

Inpersuasivegamedesignpractice,thechoiceforamotivatinggameelementisnotmade at the start of aproject. FollowingourPersuasiveGameDesign(PGD)method[260],gamificationprojectsstartbyspecifying therealworldgoalofapersuasivegame, i.e., the“transfereffect”, followedbyinvestigatingthe“usercontext”.Theinformationgatheredinthesetwostagesisusedinthenext stage, the gamification design, which includes choosing and designinggame elements for the game. The choice for the type, form, and interactionschedule of a rewardwill, thus, be influenced by the transfer goal and usercontext,aswewillshowinthissection.

A transfer effect can be specified into four components (effect type, changetype, point of impact, and domain),which all can influence the choice for amotivatinggameelement.Forinstance,ifthedesiredtypeoftransfereffectina persuasive game is to increase the social relatedness of employees on thework floor [285], a game designermight rathermotivate the employees bysocial rewards, e.g., compliments, in the game instead ofmonetary rewards,whichmightleadtoeconomicdisparitiesamongtheemployeesanddecreasesocial relatedness among them. In contrast, when the aim of a persuasivegame is to increaseself‐efficacyamong independent livingelderly,monetaryrewards might be considered as a central game element, since they canincrease a person’s required resources to overcome real‐life obstacles, tomaketheirownchoices,andthusenhanceconfidenceinpersonalcapabilities[286].Othertypesoftransfereffects,likelearning,mightnotfavorrewardsascentral gameelementsbut rathermotivateusersbyprovidingchallengesorexplorationopportunities.

Nexttotransfertype,atransfereffectisspecifiedbyitschangetype(initiating,altering, diminishing, or reinforcing a behavior) and its pointof impact(i.e.,when one expects the transfer effect to occur—during gameplay (e.g.,exergames), directly after gameplay (learning games), or a long time after


96

gameplay(lifestylechange))[260].Theexpectedpointofimpactofatransfereffectwillinfluencedesigndecisionsregardingrewards.Thiswillnotsomuchinfluencewhattypeofreward(social,monetary,points)willfitthepersuasivegame,butratherhowaplayercanobtainareward, i.e., thecontingencyofareward design in a game. Rewards can be linked to the player’s tasks,performance, orengagement [287]. For short‐term initiating transfer effects,such asphysical exercise in an exergame, rewards canbe linked to the task(getarewardwhentheplayerhascompleted10sit‐ups),totheperformance(arewardwhentheplayerdoes10sit‐upsinashorttime),ortoengagement(a rewardwhen theplayerhasplayed thegame for10minutes).Long‐termeffects,suchasalifestylechange,mightfavorengagementcontingentrewards(arewardeveryweektheplayereatshealthyanddoesphysicalexercise).Onemightalsodesigncombinationsofrewardcontingencyrelations.Forinstance,inapersuasivegamewithatransfereffecttoquitsmoking,onemightstarttoearn rewards by completing tasks, e.g., not smoking for one day, applyperformance‐contingencyafteraweek,receivearewardwhentheplayerhasnot smokedandhasbeenactive in sports, anduseengagement‐contingencyafterafewmonthsbyearningarewardwhentheplayerstillhasnotsmoked.

Especiallywhenatransfereffecthasamedium‐orlong‐termpointofimpact,it is crucial to avoid player acclimation [288] of a reward; players mightattribute high value at a reward during the beginning of the gameplay butmightnotbemotivatedbythesamerewardlateroninthegame.Toaccountfor such a decrease ofmotivation by reward, a game designer can vary theprocess of giving the rewards. Variation in rewards to maintain playermotivationcanbeachievedby(1)varyingthecontingencyofthereward(seethe quit smoking example above), (2) the value of the reward (for instanceincreasethevalueofarewardgraduallyorprovideanrewardwithunknownvalue,suchasa“mysterybox”)[288],or(3)insertingvariablereinforcements[289],suchasasuddenrewardsoccurringatunexpectedmomentsduringthegameplay.

Thedesigndecision for the formandplacementof a reward inapersuasivegamedoesdependonthespecifictransfereffect,butitwillalsodependonthe


97

userandcontextofuseofagame.Peoplecandifferintheirgeneralresponsetorewardsortheymaybeespeciallyresponsivetospecifictypesofrewards.For example, compared to adults, adolescents appear particularly “reward‐sensitive”, and hence show stronger neural and behavioral responses torewardingstimuli[284,290,291].Otherstudiessuggestthatresponsivenesstoa specific typeof rewardmaybe linked togender [292,293],personalitytraits,suchasempathyor impulsivity [255],andmentaldisorder[294‐296].TooptimizethedesignofPGD[297]andtodevelopthemostsuitablerewardforaspecificinteractionofaspecificindividual,itiscrucialtoinvestigatethemotivationsanddemographicsofyourtargetgroup.Ausefulmethodtotailorgames to specific personality types is the Hexad framework [282]. Thisframework categorizes users into six types of player personalities looselyrelated to theBig Five personality traits: Disruptors (motivated by change),Socializers (motivated by relatedness), Philanthropists (motivated bypurpose), Free Spirits (motivated by autonomy), Achievers (motivated bycompetence), or Players (motivated by extrinsic rewards). Although such aplayertypeclassificationmightworkwelltodesignentertainmentgames,theseriouscontextofapersuasivegamemightcruciallychangetheplayertype;someone might be a socializer in an entertainment game context but anachieverinaworkingcontext.Investigatingifandhowtheplayingmotivationof a user differs in an entertainment and serious context is, thus, a crucialphaseinthepersuasivegamedesignprocessandwillinfluencedesignchoicesregardingrewards.

In thepresent studyweused threebasic typesof gamerewards (monetary,social,andpointrewards). Ingamepractice,andespecially inentertainmentgames, other reward types are used as well, and they often occur incombinations.Schell(2008)listsasetofninecommonlyusedentertainment‐based in‐gamerewards[288].These includepointsandsocialrewards,suchaspraise,butalsonestedrewardsthatareprovidedwhenaplayerreachesaspecific amount of points, such as prolonged play opportunity, unlocking anewlevel,perceivingajuicyspectacle,orimprovingcharacterpowers.Money,asarewardwithanout‐gamevalue,alsocomesinvariants,e.g.,discountsorgiftcoupons.Justlikethein‐gamerewards,theseout‐gamerewardsoftenare


98

pacedduringthegameplaybypoints—aplayerhastocollectin‐gamepointsandcanonlyexchangeapredefinedamountofpointsintoarewardwithout‐gamevalue.

4B.6. REWARDSINPERSUASIVEGAMEDESIGN:CASESTUDY

In a youth addiction care context, we involved patients and therapists in aPersuasive Game Design process aimed at realizing a transfer effect toenhance a patient’s motivation to set and achieve cognitive behavioraltherapy‐related goals. To understand what game‐experiences patientsexpected to be motivating, we used Playful Experiences (PLEX) cardsrepresenting 22 game experience categories [222]. The most motivatingexperience patients selected was the experience of “thrill” [298]. We thencarried out brainstorm sessions with game designers from a serious gamedesignagency in theNetherlands togenerate the followinggamemechanicsthatwe expected couldmotivate patients in a youth addiction care context[299]: risk taking, progression map system, selfie photograph feedbacksystem,rewardsystem,andpersonalvalues.Thesemechanicswereevaluatedbyninepatientsandeighttherapists,whorankedthembasedontheexpectedmotivational impact for the transfer effect. Interestingly, patients andtherapists differed in their ranking. Patients rated risk taking and personalrewardsasthebestmotivatingmechanics,while therapistsratedrisktakingand external rewards asmost favorable [299]. The preference of therapistsregarding the external rewards seemed to correspondwith current therapytechniques that already apply external rewards to patients by usingcontingency management [267]. However, it is essential for rewards tocorrespond with both the context of application, i.e., the addiction carecontext,andthepreferenceoftheend‐user,i.e.,thepatient.

Inordertooptimizethemotivationaleffectofarewardinapersuasivegame,agamedesignercantailor,asinthePersonalizedDesignProcessmodel[204],therewardasmuchaspossible to thepreference, type[191],orpersonality[300]oftheend‐user.Moreover,itispossibletodesignagameinwhichend‐userscanchooseorgeneratetheirownrewards,ortoletfellowplayerstailor


99

the rewards for them. In our persuasive game design for a therapy context,patientsdidnotfindourpre‐setreward(a3Dprintedtokenofagoatthatwasrelated to the level they achieved) motivating. Therefore, we wanted toprovidethemwitharewardfortheiraccumulatedpointsthatwouldbemorepersonally relevant and motivating. This resulted in giving patients theopportunitytochoosetheirownrewardincollaborationwiththeirtherapist.Inaddition,weaimedtoincreasethepatients’therapeuticinvolvementingoalsetting by usingmechanisms similar to those used in the “shared decision‐making” approach in therapy [301]. The rationale for this adaptation wasbased on patients’ negative evaluations of the pre‐set tasks in setting goals.Accordingtothepatients,thisproceduremadeitmoredifficultforthemtosetgoalswhichweresufficientlychallenging,personallyrelevant,andvaluable.Intheadaptedversion,both the therapistsandpatientscoulddecideonwhichlong‐term therapy‐related goals theywould use together. This ensured thatthese goalswere relevant for the patient’s health objectives and of intrinsicvalueto thepatient. Inaddition,patientscouldtype in theirownshort‐termtasks.Insum,inouriterationweincludedthreeopportunitiestopersonalizethe game: reward, (main) goals, and short‐term tasks. However, it can bedebatedhowmuchpersonalizationwouldbepossibleandpreferableingamedesign. For example, would it be preferable to design one game for eachindividual user, or to design one game that is so open that it can be fullypersonalized toeach individualuser? Inboth situationsonecanask if thesegames would have enough overlap to be considered as the same gameresultinginthesamecomparableeffect.

4B.7. CONCLUSIONSANDFUTURERESEARCHInvolving rewards as a basic game‐element in persuasive game design toredesign psychotherapy has shown potential for youth addiction care, assubstancedependentadolescentsweremoremotivatedbyrewardscomparedto non‐dependent adolescents. In the current study, participants receivedrewardsbasedona fixed reinforcement schedule. Itwouldbe interesting toexplore different schedules for providing rewards, since specific users mayprefer a variable schedule more than a fixed one, which can be used forpersonalization. In addition, it is interesting to study how the motivating


100

effectsofrewardsdifferwhenembeddedinagameandwhenisolatedinshell‐games. The mathematical game that was used in the current study can beconsidered a “shell‐game”, since the rewardswere not integratedwith eachtask(i.e. calculation).Futurestudiescan focusonpossibledifferences in theeffects of rewards in both integrated and shell games. We expect thatmonetaryrewardsaremoreeffective inshellgamescomparedtoembeddedgames,sincetheyhaveanexternalvalueoutsidethegame.

Alignmentofarewardtothetransfereffectanduser‐contextofapersuasivegamewill informdesigndecisionsastothemostoptimalrewardtype, form,and interaction structure for a given player and context. The present paperpresentedastartinfundamentalresearchonthemotivationaleffectofgame‐based rewards in persuasive games. Since rewards are so fundamental forhuman behavior and motivation, and thus for persuasive game research,future research is stronglywelcomed,which on the onehand elaborates onreward design (e.g., reward (sub)types, combinations, and interactivestructure), and on the other hand on users (e.g., personality and context ofuse).


101

102

Basedon:vanDooren,M.M.M.,Siriaraya,P.,Visch,V.T.,Spijkerman,R.,&Bijkerk,L.(2019).Reflectionsonthedesign,implementation,andadoptionofagamifiedeHealthapplicationinyouthmentalhealthcare,EntertainmentComputing,31,1875‐9521.

REFLECTIONSONTHEDESIGN,

IMPLEMENTATION,ANDADOPTIONOFAGAMIFIEDEHEALTH

APPLICATIONINYOUTHMENTAL

HEALTHCAREThe previous chapters provided argumentation that it is important topersonalize both the eHealth product and game elements and involve thestakeholders(e.g.usersofthegamifiedeHealthproduct) inthedesignprocess.Thiswas taken into account in the personalizeddesignprocess ofa gamifiedmoduleofthetherapysupportive“Lucaapp”and thedesignofthe“Lucaapp”itself. Luca has been designed for youthmental healthcare and supports thepatient to alsowork on their therapy at home. In this chapterwewill firstdescribe the overall design process of the gamifiedmodule of the Luca app,followedbyadescriptionoftheothermodulesoftheappthatweredevelopedbythe Luca team. Following a preliminary evaluation,we applied a final designiteration to the app thatwas implemented in therapeutic practice.However,fewertherapiststhanexpectedusedtheappandtheinflowofpatientswaslow.Inaseriesofqualitativeinterviews,weinvestigatedtheirreasonsfornotusingit.BasedontheoutcomesandourdesignprocessweidentifiedrecommendationstoenhanceimplementationandadoptionofeHealthinmentalhealthcare,whichwedescribeinthefinalpartofthischapter.

Chapter5|Reflectionsonthedesign,implementation,andadoptionofagamifiedeHealthapplicationinyouthmentalhealthcare

104

5.1. INTRODUCTIONDigital games have become a popular pastime for most people in today’ssociety.Playersarehappytospendconsiderableamountsoftimeandenergyengaging on tasks and activities within a well‐designed game environment.The term“Gamification”hasbeenused todescribe theapproachofapplyinggame elements to a non‐game context [192]. These game elements aretypicallyused to transform theusual experienceof anon‐game task intoanexperience thatpeople findmoregratifying, thereforemaking the task itselfmoreengaging.Thisincreaseinuserengagementmakesgamificationusefulinavarietyofdomainssuchaslearningandmarketing.

Inthefieldofhealthcare,thereisgrowinginterestinapplyinggamesinorderto improve health related outcomes, especially in the area ofmental health[78]. For a long time, serious games have been developed and used to helpmanage symptoms related to various conditions such as depression, post‐traumaticstressdisorders(PTSD)andautismspectrumdisorder(ASD)[302].Asforgamification,thisapproachcouldbeparticularlyusefulintransformingordinarytherapeuticpracticesintomoregratifyingexperiences,thushelpingtomotivatepatientstobetteradheretothetreatmentprocess.Thiscouldbeparticularlyusefulformanyofthetherapeuticactivitiesusedinfieldssuchaspsychotherapy, where a substantial proportion of individuals with mentalillness drop out of treatment [303]. Prior studies have argued that varioustherapeutic training tasks (such as Cognitive‐Bias Modification tasks (see[304])) could be enhanced through the use of game elements (e.g.motivational feedback or a surrounding shell game system), making themmoreengagingforpatients[245].

Despitethemanypotentialbenefitsofgamificationinmentalhealthcare,itisnotaseasyasitmightfirstappeartodesignanddevelopeffectivegamificationwithinthiscontext.Foranygamificationtobeeffectiveand“meaningful”,thedesign should take into account the needs, interests and capabilities of itsusers and the underlying tasks and usage context [305]. However, in thementalhealthcarecontext,thiscanbeparticularlycomplexastheoutcomeofagamifiedinterventiondependsonmultiplestakeholders,anddifferinglevels


105

ofmotivationaretypicallyrequiredtoachievethetherapeuticobjectives(i.e.:motivationtoachieve long‐termbehaviorchange,motivation toengagewiththe treatmentor to accomplish short term therapyobjectives,motivation touse gamification within the therapeutic procedure, etc.). In addition, it isrequisite that active and constant feedback from these stakeholders isintegratedwithinthegamesystemitself.Forexample, incertaintherapeuticapproaches, the therapeutic staff has to review whether the therapeuticactivitieshavebeenaccomplishedsatisfactorilybeforegamerewardscanbegiven, or before the player can progress to the next level. This requires thedesignertonotonlyconsiderplayer‐gameinteraction,butalsostaff‐gameandplayer‐staff interaction when designing the gamification. Finally, thegamification process itselfwarrants careful consideration as, in some cases,inappropriate integrationof game elements couldhave an adverse effect onthe underlying therapy. When gamifying computerized training tasks forinstance, it was found that adding features such as real‐time feedback toimproveuser engagement could lead to anunintended increase in cognitiveload, thus distracting users and reducing the overall effectiveness of thetraining [245]. Thismeans that it is difficult to find an appropriate balancebetween the health goal and game appeal when developing gamifiedtherapeuticinterventions[306].

Astheapplicationofgamificationinmentalhealthcareisrelativelynew,therehave as yet been few theoretical or practical guidelines established on howsuch issues could be addressed.Most existing gamedesign frameworks andguidelines are based on serious games developed for other specific usagepurposes (such as for the self‐management of diseases [307]): knowledgewhich might not transfer well to designing games for mental healthcareproblems. Overall, there have been few studies that provide case specificknowledge about the processes required to design and implement effectivegamifications that are able tomeet the complex demands of practical usagewithin a mental healthcare context. In mental healthcare, the majority ofevidence‐based therapies include psychotherapy,which can be defined as atreatmentmodality “inwhich the therapist and patient(s)work together toameliorate psychopathologic conditions and functional impairment through


106

focuson the therapeutic relationship” [13].Oneof themost frequentlyusedand studiedpsychotherapies is cognitivebehavioral therapy (CBT) (e.g. [15,308,309]).

5.2. RESEARCHTHROUGHDESIGN:THEREADYSETGOALSASADESIGNCASESTUDYFOR

GAMIFICATIONOFCOGNITIVEBEHAVIORAL

THERAPYTheoverallaimofthispaperistoprovidemorebottom‐updesignknowledgeabout the development of gamifications for CBT. In particular, the learningsand challenges for designing such gamifications are presented from aResearch throughDesign (RtD) perspective [310, 311]. RtDmethods aim togenerate the knowledge to address so‐called “wicked problems”: complexproblems where different context areas overlap. A holistic approach thatintegrates theories fromdifferent domains is required to provide a solution(and as suchwould be appropriate for generating knowledge for designingmentalhealthcaregamifications)[310,311].Thestudyhighlightshowvariousmethodsfromthefieldofdesign(gamedesignanduser‐centereddesignetc.)couldbeusefulinhelpingtoaddressthechallengesencounteredateachstageofthegamificationdesignprocess.Inaddition,theresultsfromapplyingsuchmethodsarediscussedindetailtoprovideacasestudyforgamificationdesigninyouthmentalhealthcare.

As the theoretical basis of the gamification design process, the PersuasiveGameDesign(PGD)model[63]wasusedandplayedakeyroleinthedesignofthe gamification, i.e. the ReadySetGoals application (Figure 1). This modelproposes that games are essentially experience‐defined. In daily life, usersexperiencearealworld.Throughgamificationdesign,itispossibletoshiftthisordinary“real‐world”experiencetowardsamore“game‐like”experience.Byaddinggameelements to real‐world tasks,usersare triggered toexperiencegratifying and motivating game world specific feelings during ordinaryphysical world activities. A transfer effect occurs when the experiencesobtained by users in the game world successfully influence the player’sattitudesorbehaviorinthephysicalworld.


107

Figure1.ThePersuasiveGameDesign(PGD)model.

Using the ReadySetGoals application as a case study, the various questionsthat needed to be addressed when designing and developing mentalhealthcare gamifications were analyzed through the theoretical lens of thePGD model. The processes employed at each step to obtain those answerswere then presented as a structural process, according to the model.Afterwards, the key challenges encounteredduring the design, developmentand implementation of the application were discussed, followed by designlearnings.Overall,theobjectivesofthispaperareto:

1. Presentacasestudyoftheiterativedevelopmentofagamificationformentalhealthcare.

2. Investigate, byResearch throughDesign, a structural process for thedesignofgamificationsthatcanmeetthecomplexdemandsofmentalhealthcare.

3. Evaluate (qualitatively) the support of therapists for the therapeuticapplicationofanewgamification.

4. Extend gamification knowledge in the mental healthcare domain byhighlighting reflected learnings based on the practical experience ofiteratively designing and developing a youth mental healthcaregamification.

The design and development process of the ReadySetGoals applicationconsistsoffourkeystagesbasedonthetheoryofgamificationoutlinedinthe


108

PGD model – c.f. Siriaray et al 2018 (PGD cookbook method) [260]. Thesestagesinclude:

1. DefineTransferEffect2. ExploreUserContext3. Design Gamification (RtD Concept Design Stage, Iterative

Development,UserTestingStage)4. EvaluateGamification

Variousmethodsfromthefieldofuser‐centereddesign,seriousgamedesignandgamedevelopmentwereusedduringeachof these stagesof thedesignand development of the ReadySetGoals application [260]. Feedback fromdiversestakeholderssuchasdomainexperts, endusersandgamedesigners(totalingN=44)was incorporated into thedesignof thegamifiedapplicationby means of the studies carried out at each stage of the process. Table 1providesanoverallsummary.Adetaileddiscussionofeachstageisprovidedinthefollowingsection.

5.2.1. DEFINETRANSFEREFFECT The first stageof thegamificationprocess involves thedeterminationof thedesired transfer effect. In the case of mental healthcare gamification, thisinvolves examining the various procedures, activities and tasks used in thetherapy process and determining which outcomes (cognitive or behavioraletc.)thegamificationshouldaimtoachieve.Twoaspectsinparticularrequireconsideration.Thefirstistoevaluatewhetherthegamificationapproachitselfcanoffersufficientaddedvaluetowardsthedesignatedtransfereffectandtodetermine how useful the achieved transfer effect is towards the overalltherapyobjectivesofCBT.Forinstance,iftheretentionofaspecificprocedureis found to obstruct the therapy outcome, such as in Cognitive BiasModificationtrainingprogramsthatrequireprolongedandrepetitivetraining(see [312]), then the use of gamification, which has shown to be useful inimprovingusermotivation,couldbeparticularlyvaluable.Thesecondaspectwhichshouldbeconsidered iswhether the transfereffectand theunderlingactivity or task used to achieve the transfer effect within the therapeuticcontext is suitable for gamification.Transfer effectswhich rely onoutcomes


109

thatarecomplex,time‐consumingtomeasureorhaveanunclearrelationshipwith their actions can be harder to gamify and therefore might not beadequatelyeffectiveinpractice.

5.2.2. EXPLOREUSERCONTEXT After identifying the desired transfer effects, the next stage involvesexaminingtheuser’sreal‐worldcontext.Inmentalhealthcaregamificationitisparticularly important to situate this analysiswithin the overall therapeuticcontextandtoconsidertherelatedtheoreticalaspectsunderlyingthetherapy.InthespecificcaseoftheReadySetGoalsdesign,weaimedtounderstandmoreabout three aspects of the present real‐world therapeutic context: 1) thecharacteristicsoftheusers(inthiscase,adolescentsubstanceabusepatients),2) the therapeutic context in which the gamification is to be implemented(youthsubstanceabusetreatment)and3)thecontextunderlyingtheactivityto be gamified (goal setting). This involved discussions with addictiontreatmentexperts(N=2)attheclinic.Basedonthis,anumberofcontextualfeaturespertainingtoeachoftheseaspectswereidentified.Thesecontextualfeatureswerelaterusedtoformulatethegamificationconcept.

With regard to 1) the characteristics of adolescent substance abusers, thepersonality characteristics of sensation seeking and impulsivity werecommonlyidentifiedinpeoplewithaddictionproblems(e.g.[313‐317]).Thiswas further confirmedbyanexploratory studywhichwas carriedoutusingPLEX cards. In this study, participants selected the playful experience of“thrill”asthemostmotivatingexperience[298].Anotheraspectsuggestedinthe discussions as being particularly effective in the addiction treatmentcontext was the use of rewards, as rewards can motivate young patientsundergoingaddictiontreatmenttoexhibitspecificbehavior[318].As forthenatureoftheactivityofgoalsettingitself,oneapproachwhichhasbeenusedsuccessfully in therapeutic treatment to encourage the setting of achievablegoals is to divide the tasks into smallsteps, starting from easily achievablesteps to more difficult ones as people tend to perform better and to enjoypursuinggoalswhichprovidesufficientchallengetomatchtheircurrentskills.In addition, the feelingofaccomplishmentderivedfrom completing previous


110

goalscouldbeleveragedwithinafeedbacklooptoencourageuserstosetandpursuenewgoals.

Table1.ThedesignanddevelopmentprocessfortheReadySetGoalsapplication

Stage Objective Method1 Define

TransferEffect‐Discussionsessionwithcarestaff(therapistpractitionersatanaddictiontreatmentcenter)(N=2)(role:domainexperts)

Outcome:Appropriatetransfereffectidentifiedas“Encouraginguserstosetandfollowthroughtherapygoalswhichwouldresultinpositivetherapyoutcomes”

2 ExploreUserContext

‐PLEXcardstudywithpatientsatatreatmentcenterforaddiction(agebetween17‐21years)(N=7)(role:endusers)‐Discussionsessionwithaddictionexperts(N=2)(atherapistpractitionerandacaremanageratanaddictiontreatmentcenter)(role:domainexperts)

Outcome:Sensationseeking,applicationofrewards,smallstepsandfeelingsofaccomplishmentidentifiedasgamificationopportunities

3 DesignGamification

‐Brainstormsessionwithgamedesignersfromaseriousgamedesigncompany(N=2)(role:designexperts)‐Questionnairewithpatientsinaddictioncare(role:endusers)(N=9)andstaff(role:stakeholders)(N=8).Staffmembersincludedcaremanagers,systemtherapistsandtherapistpractitioners.

Outcome:Designofcoregameplayloopbasedontherisk‐takingconcept4 Evaluation

Gamification‐TestingofearlyprototypewithgeneralusersfromaUniversity(age21‐24)(N=5)(role:generalusers)‐Testingofprototypewithpatients(role:endusers)(N=6)andstaff(role:stakeholders)(N=3)fromayouthoutpatienttreatmentcenter

forsubstanceabuse.Outcome:Additionofnarrativemetaphorandimprovementofthegoal‐settingmechanism

throughpersonalization

5.2.3. DESIGNGAMIFICATION Thegamificationconceptdesignstageconsistsoftwoparts.Thefirstpartistoidentify appropriate gamification concepts which align with the contextualfactors identified in the previous stage. The subsequent part involves theconstruction of a core game loop. In the design of ReadySetGoals,brainstormingsessionsforthefirstpartwerecarriedoutincollaborationwithgame designers from &RANJ (a serious game design agency from theNetherlands)togenerateideasaboutthegamificationmechanicsorelementsthatcouldbeusefulinmotivatingplayersonthebasisofthepreviousfactors.

During the brainstorming session, five different game mechanics wereidentified with the potential to engage patients undergoing therapeutic


111

interventionforaddiction.Themechanicofrisk‐takingwasselectedtoappealto the contextual factor of sensation seeking and impulsivity, by providingexperienceswith “thrill factor”. Theprogressionmapsystem was selected toappeal to the contextual factor of small steps and the selfie photographfeedbacksystem toappeal to thecontextual factorofa feelingofsuccessandaccomplishment. These factors were aimed at providing experiences ofachievement and competence. In addition, the tangible reward system andpersonal values reward system were used as they can motivate patients toexhibit specific behavior. Afterwards, these gamemechanicswere expandedinto gamification concepts, with the core gameplay of each concept builtaround the identified game mechanic. For instance, in the “risk‐taking”gamification concept, the main task of the players involves placing an“investment”ontheirsetgoals,basedonhowlikelytheythinktheyaretobeable to complete the goal. Players receive bonus points if they succeed andlosepointsiftheyfail.

Thesegamificationconceptswerethenevaluatedbypatients(N=9)andcarestaff (N= 8). A questionnaire studywas carried outwhereby patientswereaskedtoranktheorderoftheconceptstheythoughtwouldbemostenjoyable(activitiesrepresentingtheconceptswereprovidedasexamples).Meanwhilethecarestaffrankedtheconceptthattheythoughtwouldworkbestwiththepatientsandwouldbepractical in the care setting. Interestingly,our resultsshowedadiscrepancybetweenpatientsandcarestaffintheirpreferenceforgamification concepts, suggesting that it would be useful to involve bothgroupsofstakeholdersinthedesignprocess.Theresultsshowedthattherisk‐taking (mean = 4.10/5.00) and personal rewards (mean = 3.90/5.00)wererated highest by the patients, while care staff rated risk‐taking (mean =3.50/5.00) and external rewards (mean = 3.50/5.00) as the joint mostfavorable. Based on this, the risk‐taking conceptwas selected as the maingamificationconceptinourdesign.

5.2.3.1. CONSTRUCTING THE CORE GAMEPLAY LOOP

Thelaterstageofthegamificationconceptdesigninvolvedtheconstructionofthe core game loop. This element represents the repeating processes thatdrivethecoreactionandthe interactionbetweenplayersandthegame(the


112

main interaction‐feedback loop). At the core of most game systems, theactionsof theplayersareaimedataccomplishingcertain in‐gameobjectivesandinteractionsforwhichthegameprovidesfeedback(see[319]).

While the game loop concept emerged from the traditional field of gamedesign, we found that this concept was equally useful and applicable togamificationdesign. Like gamedesigners, gamificationdesignersmust carryoutthetaskofcreatingastructuredgameplayexperienceforusers.Thecoregameloopisparticularlyusefultohelpachievethis.However,akeydifferenceisthatasthegoalofgamificationistohelpusersachievethedesiredtransfereffect, the game experiences often incorporate real‐world elements duringgameplay. In the caseofmentalhealthcaregamification, the coregame loopcouldbeformulatedbysituatingthecoreactionsandinteractionsneededtoachieve the therapeutic objectiveswithin a game play loop, based upon thechosengamificationconcept.Specifically,theformulationofsuchagameloopallowsdesignerstobetteranalyseswhetherthegamemechanicsusedinthegamificationdesignactuallysupportthekeyactionsorinteractionswhichwillleadtothedesiredtherapeuticoutcomes.Thecoregameloopalsoservesasasharedlanguagewhichfacilitatescommunicationbetweenthegamedesignersandpracticingclinicians,bymakingthetherapeuticprocedureclearerto thegame designers and informing the clinicians of how the game interactionprocesswillworkinpractice.IntheReadySetGoalsapplication,thecoregameloopwascenteredontheactivityofsuccessfullysettingandachievinggoals,and the risk‐taking concept and the resulting game play system wereformulatedonthebasisoftheproposedpremise(Figure2).


113

Figure2.ThecoregameloopoftheReadySetGoals

Afterwards, structural gameplay elements were added to support the coregame loop. Structural gameplay elements represent the wider context ofgameplay(progressionsystems,exploration,anarrativestoryline,highscoresystemsetc.) andcanbeused to enhance the coregame loopexperience, toencouragesustainedengagementacrosstheloops[320].IntheReadySetGoalsapplication design, thesewere not only selected tomatch the design of therisk‐takingconcept,butalsotocoincidewiththecontextualfactorsidentifiedin theprevious stages (i.e. small stepsandapplicationof rewards).Aplayerprogression‐basedrewardsystemwasimplementedtomaketheaccumulatedpointsmoremeaningfulandrewarding.Inaddition,tofacilitatethetransitionfrom easier to more difficult goals, a skill‐tree style difficulty progressionsystem was implemented within which easier tasks had to be completedbeforeproceedingtomoredifficultones.

5.2.4. EVALUATIONGAMIFICATION Aplayableprototypebasedon theproposed coregame loopwas iterativelydesigned, developed and tested with users. The prototype was created anddeployed as an Android mobile application using PhoneGap(http://phonegap.com) (Figure 3). Although the ReadySetGoals applicationwas designed to be domain independent (objectives related to different

Set a new goal by selecting a task to

carry out

Place a wager and set a time limit. A higher wager and lower time limit will increase reward but

is more risky

Users are challanged to

complete the task within the time limit they set

Provide visual feedback of the status of the goal

Take photo as proof of goal completion

Receive points based on their wager and time limit which they can use in their next

wager


114

aspectsoftherapycanbeusedasgoals),whentestingtheapplication,thecarestaffdecidedtofocusonusingtheapplicationtoencourageparticipantstosetgoals to commence and pursue leisure activities. Such activities encourageparticipants to find alternative sources of reward and fulfilment that areinconsistentwithdruguse.Overall,fiveactivitieswereavailableforselection(Reading, Watching Films, Football, Jogging and Photography), eachrepresenting alternative behaviors which are inconsistent with drug use.Altogether, 84 pre‐defined tasks were provided, based on discussions withexpertsineachactivityandfurtherrefinedbythreegamedesigners.Carestaffwasaskedtoreviewthetaskstoensuresuitabilityforthepatients.Toensureanonymity,alltheinformationwasstoredlocallyontheparticipants’phones.

1.Anoverviewofthecurrentgoals.Userscansetnewgoals

byclickingonthe“+”

2.Userscanseetheirprogressiononthemountain.

Greenflagsrepresentachievedgoalsandorangeflagsrepresentcurrentlyset

goals.Thewhiteflagsrepresentopengoalswithwhichuserscanchallenge

themselves

3.Whenusersachieveagoaltheycantakeaphotoasproof

Figure3:ScreenshotsoftheReadySetGoalsApplication

Theapplicationwasfirstusedinausertestingsessionwithfiveparticipants(aged 21‐24) recruited from the Delft University of Technology (twomales,


115

threefemales)toinvestigatehowplayersperceivedtheimplementationofthegamemechanics and to investigate issues related to general accessibility. Inthe testing session, participantswere asked to set and complete goalsusingtheapplication.A“ThinkAloud”approachwasadoptedandparticipantswereasked open‐ended questions about their general experience with theapplication.Thisledtostructuralimprovementstothecontentstructureandnavigation schema. In addition, the narrative metaphor of climbing amountainwasaddedtoimproveconnectednessbetweenaccomplishinggoalsandprogressingwithoveralltherapy.Themountainrepresentedachallengingobstacle,andthegoalsthathadbeenaccomplishedsofarwerevisualisedwithflagsplacedon themountain, each representing a step thatusershad takentowardsovercomingtheirobstacles.

5 .2 .3 .1 . THE IMPROVEDREADYSETGOALS APPLICATION PROTOTYPE

Screenshotsof theReadySetGoalsapplicationcanbeseen inFigure3. In theapplication, participants start by setting a new goal which they want topursue.Theyselectwhichtaskinaspecificleisurecategory(i.e.runXkminYminutes)thattheywouldliketotakeonasachallenge.Afterwardstheyselectatimelimitandusetheiraccumulatedpointstoplaceawageronhowlikelytheythinkitisthattheywillachievethegoal.Placingahigherwagerorsettinga lowertimelimit increasesthepotentialreward,but isriskierasusers losethepointsputupforthewageriftheyfail.Whenuserscompletethegoal,theytake a photowith their phone as proof and receive their rewards in points,based on the risk level and the difficulty of the task. As shown in Figure 3,participants select tasks from the bottom of the mountain andmove up tomore difficult tasks as they progress. Finally, as a reward for accumulatingenough points to reach a higher level rank, they are presentedwith a non‐monetaryrewardattheendofthesession,intheformofa3Dprintedtokendepictingtheirlevel.

5.3. PRELIMINARYTESTINGWITHTARGETAUDIENCEResearchersrecruitedparticipants fromayouthoutpatient treatmentcenterforsubstanceabuseintheNetherlands.Threecarestaffmembersdecidedtoparticipateinthestudyandwereaskedtorecruitpotentialparticipantsfrom


116

theirpatientpopulation.Patientswhoagreedtoparticipateinthestudywereasked to engagewith the application for three to fourweeks. The inclusioncriterionwashavinganandroidphoneandtheexclusioncriteriawerehavinga gaming and/or gambling addiction. During the experiment, participantswereasked touse theapplication freely.Toensure thevalidityof thegoals,carestaffwasasked to review,during their therapymeetings, thegoals thattheirpatientshadaccomplished.

After the evaluation period, each participant was asked to fill outquestionnaires about their general game experiences [321], and semi‐structured interviewswerecarriedout.Participantswerealsoasked to ratevariousfactorssuchasthevalueoftheapplicationinhelpingtomotivatethemto set goals and the value of various game elements such as the narrationtheme, on a 0 to 10 Likert scale. The semi‐structured interviews coveredtopicssuchasthegamemechanics, thepotentialusageoftheapplicationfortherapy and suggestions on how to improve the application. As a reward,participantswerepresentedwithacinemagiftcardworth10euros.

5.3.1. RESULTSANDSUGGESTIONSFORITERATIVEIMPROVEMENT

The results from the preliminary study provided key insights, highlightingmultiple areaswhere theapplication couldbe further improved.Overall, sixparticipants and three care staff participated in the interviews andquestionnaires(onedroppedoutduetorelapseandoneparticipantwasnotable to use the application and was excluded from the study). Participantsreported that the application was moderately useful in motivating them toachieve their goals (Mean Rating: 5.00/10.00, SD = 1.67) and reported amoderate sensation of achievement after achieving a goal by means of theapplication (Mean Rating: 5.50/10.00, SD = 2.59). One aspect that wasperceived asparticularlynegativewas the tasks thatwereused in the goal‐setting process. In particular, the decision to provide pre‐set tasks forparticipantstouseinsettinggoalswasnegativelyperceived.Predefinedtasksallow for amore controlled provision of beneficial health objectives (whichcould be essential if the goals are related to serious therapeutic objectives,suchas training activities thathave rigid rules).However,whenused in the


117

domainofleisureactivitiesitappearedtobelessusefulandrenderedfarlesssatisfaction to the participants, as they felt that their autonomy had beenconstrained. “Iwould feelprouderaboutachieving thegoals if Ihad thoughtthemupmyself,thenIwouldfeelmoremotivatedtoachievethem”(participant2,female).Therestrictivenatureofpre‐settasksalsomeantthatitwasharderforparticipantsto formgoalswhichweresufficientlychallenging,personallyrelevant and valuable. These aspects were frequently criticized during theinterviews.Soparticipantsfeltthattheapplicationwasnotveryhelpfulintheformulationofgoals(MeanRating:3.50/10.00,SD=2.07)andthatitwasnotsufficientlychallenging(MeanRating:0.63/4.00,SD=0.43).

To improve this aspect, mechanisms similar to those used in the “shareddecision‐making” approach in therapy [301] were suggested as a usefuladditiontothegamification.Thisapproachincreasesthepatients’therapeuticinvolvement by allowing them to make collaborative decisions within thetherapy,thusenhancingtheirfeelingofautonomy,whilethetherapistguidestheminkeepingthegoalsrelevanttotheirhealthobjectives.Asimpleexampleofhowthisapproachcanbeappliedtothegamifiedapplicationistoprovideparticipants with “suggested goals” pre‐developed by the care staff and toallowparticipantstomodifytheseintotheirowngoals.

Regardingthegamificationdesigningeneral,participantsparticularlyenjoyedthe narration aspect. Participants experienced the theme and therepresentation of the skill‐tree based progression mechanism as positive(Meanrating:6.33/10.00,SD=2.66)and felt that thenarrativeconceptwasone to which they could relate. The risk‐taking mechanism was ratedmoderatelyintheapplication(MeanRating:5.67/10.00,SD=2.16)andsomeusersperceivedthisfeatureasbeingmotivatinginitself“Ididenjoysettingthepointsand tended togo ‘all in’withmypointsso that there ismoreatstake.”(participant5,male).Tofurtherimprovethismechanism,abetterrewardforthe accumulation of points would need to be provided, such as the socialrewardsgainedfromcomparingperformanceandprogresswithothers.


118

5.4. FINALITERATIONOFTHEREADYSETGOALSGAMIFICATION

5.4.3. THEREADYSETGOALSGAMIFICATIONToimplementtheReadySetGoalsapplication,wecollaboratedwiththe“Luca”appproject.TheLucaappwasdesignedtosupportpatients inyouthmentalhealth or addiction treatment health care in their home environment andincludes a number of therapymodules such as: 1) themedication alarm, inwhichpatientscanregisterthemedicationstheyhavetotakeandcansetanalarm to remind them, 2) the emergency plan, where patients prepare –togetherwith their therapist–aplanofhowtosolvea futurementalhealthcrisis,3)achat,thatallowstherapistsandtheirpatientstoquicklysendeachother short messages in an encrypted and safe manner, 4) an activity list,where patients record positive activities that reinforce their mental andphysicalwell‐being,5)adiary inwhichpatients registerhigh‐risksituations(for example situations that evoke a craving to use substances) and theirmood.Figure4providessomescreenshotsofthediarywhichconsistsofthreeelements: 1) “ups& downs”, 2) “mood‐measurement”, and 3) “registration”.Withthe“ups&downs”,patientscanquicklyregisterhowtheyaredoingbyclicking on one of the three emoticons with either a thumbs up, thumbssidewaysorthumbsdownandgradingthisonascalefrom1to10(screenshot1 and 2 of Figure 4). They can also describe their state‐of‐mind and add apicture.With the “mood‐measurement”, patients can register theirmood byrating the intensity of the emotions (such as fear, happiness, anger, andsadness)whichtheyhaveexperiencedduringtheday(screenshot3ofFigure4). Lastly, patients can register the difficult situationswhich triggered theiremotions, by first noting information such as who they were with andinformationabout theevent thathadoccurred (screenshot4ofFigure4).Atherapistcanthendiscussthecontentofthemodulesduringface‐to‐faceCBTsessions.


119

1.Inthediary,usersrecordhowtheyarecurrentlyfeelingbyclickingononeofthe

emotionicons

2.Inaddition,usersalsoregisterhowstrongtheirfeelingsare,bygradingthem

onaten‐pointscale


120

3.Next,userscangradespecificaspectsoftheirmood,byslidingtheemoticons(scared,happy,angry,sad)tothesuitablegrade

4.Userscanalsodescribespecificsituationsbyrecordingwhattheyweredoingandwherethesituationoccurred.Theycanalsotakeaphotoforextra

clarification

Figure4:ScreenshotsofthediarymoduleoftheLucaapp.

The therapy goals module is one of the Luca app modules (Figure 5). Weimplemented ReadySetGoals as a gamified version of this module.ReadySetGoalswasimproved,basedonearlierfeedbackfromend‐users,by1)enhancing user autonomy, 2) improving the intrinsic value of tasks, and 3)improvingthenarrativemetaphorelement.Asuserswantedmoreautonomyinselectingtaskswhilethetaskshadtorelateto themaintherapyobjectivewe adopted a shared‐decision approach, as commonly applied in healthcare[301],totaskselection.Specifically,patientscouldsettheirtasksthemselvesor in collaboration with their therapist. In addition, patients could set thenumber of points to be rewarded for the task togetherwith their therapist,


121

and were given the option of writing comments on how difficult it was to(attemptto)fulfilthetask.

To improve the intrinsic value of tasks, and to ensure that the long‐termtherapy‐related goalswere relevant for the patient’s health objectives, boththetherapistsandpatientsdecidedonwhichlong‐termtherapy‐relatedgoalstheywouldworktogether.Thetherapistsfirstselectedtherelevantareainlife(e.g.mood / psychiatric problems, school /work / education, or substanceuse) and typed in the specific long‐term goal. Patients could achieve theselong‐term therapy‐related goals by setting and trying to achieve short‐termtherapy‐relatedgoals(i.e. thevarioustasks) intheseareas.Patientstypedintheirown tasks, after selecting the corresponding long‐term therapy‐relatedgoal,andsetremindersofspecificdatesandtimes.Inaddition,patientschosetheir own rewards for the accumulated points in collaboration with theirtherapist. Thismade the rewardsmore personally relevant andmotivating.During each therapy session, therapists evaluated the progress of the long‐termtherapygoalsandsetnewgoalsoncethetaskshadbeenaccomplished.

Lastly,toimprovethenarrativemetaphorelement,weenhancedthe‘climbingamountain’metaphor.Thewagerthatpatientscouldsetwhenselectingataskwasreferredtoasdiamondsinsteadofpoints.Thiswasbecausediamondsareperceivedashavingahigherobjectivevalueandbecausediamondsareoftenfoundinmountainsornearmountainranges.Whenpatientsachievedatask,they could take a photo and this would be displayed on the mountain tohighlighttheprogressof theirclimb,andtheachievementof their long‐termtherapy‐relatedgoal.Toalign theskill‐treebasedprogressionmechanismtothe therapy structure, therapists estimated the amount of therapy sessionsneeded,andthishelpeddeterminetheheightofthemountain.Patientswouldreachthesummitofthemountainwhentheycompletedtherapy.Weexpectedthat this would motivate patients to visit therapy sessions. When patientsattended a therapy session, the therapist evaluated the therapy goals in thetherapy goalsmodule of the app. In this way, patients would automaticallyprogress to the next therapy session and thus climb visibly further up themountain.Notattendinga therapysessionpreventedpatients fromclimbingthemountain.


122

1.Theuserselectsarelevantlong‐termtherapygoalandformulatesashort‐termtask.

Theuserthenbetsdiamondsonthesetgoal.Afterwards,theuserclicksonthe“volgende’(next)

button

2.Theuserthenestimatesthedifficultyofthetaskandproceedsbyclickingon“opslaan”(save)button

3.Theuserthenseesanoverviewofthetaskandthecorrespondinggoal.Itisalsopossibletosetremindersfor

thestep

4.Ifauserwishestosetareminder,heorshecanselectthedateand

time

5.Ifauserhasachievedthetask,heorsheclicksthe“ja”(yes)buttontoproceed

6.Theusercanthentakeaphotoofthegoalorsituation,fordisplayonthemountain


123

7.Theuserwillthenseehowdifficultheorshehadthoughtthetaskwasbeforestartinganddescribes

howdifficultthetaskwasinpractice

8.Anoverviewoftaskssetandagreeduponbypatientandtherapistisshownonthemountainscreen.Thisscreenshotshowsasingletask(inpurple).Byclickingonthe“+”,theusercansetanewtask.Byclickingonthetreasurebox(top

right),theusercanseehowmanydiamondshavebeenearnedandhowmanyhavetobeearnedinordertogainthenextreward.Theusercan

navigatethemountainbyscrolling.Heorshecangobacktothecurrentsessionbyclickingonthe

lowerleftbutton

Figure5.Screenshotsofthegamifiedtherapy‐goalsmodule.

In summary, the process of using the gamified therapy goalsmodule in theLucaapp isas follows:Whenapatientwishes tosetagoal,he/sheselectsalong‐termgoalandtypesintheshort‐termtask(screenshot1ofFigure5),Hethenbetsdiamondsandgivesfeedbackabouthowdifficulthethinksthetaskwillbe(screenshot2ofFigure5).Inaddition,thepatientcansetaremindertime(screenshot3and4ofFigure5).Atthesespecifiedtimes,apop‐upwillappearonthepatient’ssmartphonetoremindhimofthetask.Onceataskhasbeen accomplished, the patient selects “yes” (screenshot 5 of Figure 5), cantakeapicture(screenshot6ofFigure5)andrateshowdifficultthetaskwasinpractice (screenshot 7 of Figure 5). If a patient takes a picture, it will bedisplayed on the mountain. The patient can discuss what task and therapy


124

goalhasorhasnotbeenaccomplishedduringthe followingtherapysession.The last screenshot shows an overview on the mountain of the number oftasks that were set and agreed upon by patient and therapist before thecurrenttherapysession(representedbyapurpledot).

5.5. IMPLEMENTATION OF THE LUCA APP AND

EVALUATIONBYTHERAPISTSWhen writing this paper, we were testing the general effect of thegamification,by conductinganon‐randomizedpre‐post (eightweeks) study.In this studywe contrasted two conditions: one thatwas gamified and onethatwasnotgamified,toevaluatetheLucaappwithagamifiedtherapygoalsmoduleandtheLucaappwithanon‐gamifiedtherapygoalsmoduleinayouthmentalhealthcare setting (bothmentalhealthandaddiction care).Thenon‐randomized prospective study that we planned was a study with an A‐Bdesign among 60 youngsters in youth mental healthcare who used thetherapy‐supportiveLucaappinthecontextofoutpatientcognitivebehavioraltherapy (CBT, their regular therapy). Our exclusion criteria comprised ofpatients with problematic gaming or gambling behavior, who receivedinpatient treatment less than 3months prior to the start of therapy, had a(light) mental disorder, acute suicidal or psychotic complaints, or aninsufficientcommandoftheDutchlanguage.Accordingtoourstudyprotocol,thefirst30youngsterswhoparticipatedinthestudywouldreceive‐besidesCBT ‐ the Luca app with a non‐gamified therapy goals module and thefollowing30youngsterswouldreceivetheLucaappwithagamifiedtherapygoals module.We expected that a higher frequency of use of the Luca appwould be related to better therapy outcomes (i.e. less psychologicalcomplaints/substance use, higher therapy retention, higher motivation fortherapy); that theLuca appwitha gamified therapy goalsmodulewouldbemore frequently used than the Luca appwith a non‐gamified therapy goalsmodule and that youngsters who received the Luca app with a gamifiedtherapy goals module would have better therapy outcomes than theyoungsters who received the Luca app with a non‐gamified therapy goalsmodule.Atthestartofthestudy,thetherapistsweretrainedintheuseoftheLuca app and given a leafletwith screenshots and instructions for the Luca


125

app and both the gamified and non‐gamifiedmodules, and they understoodhowthetherapygoalsmodulewouldbeusedfortherapy.However,theinflowof patients in the Luca study was low, and at the time of writing it wasuncertainwhether the study could be finalized as planned according to ourprotocol.

5.5.1. QUALITATIVEMID‐STAGEEVALUATION Duringthestudy,theparticipantinfluxwasmuchlowerthanexpected,whichmotivated us to conduct a qualitative evaluationwith therapists in order toinvestigate the causes for this low research participation by patients andtherapists. To solicit opinions and receive feedback about the potentialadvantagesordisadvantagesofusinggamificationaspartofthetreatment,wecarried out semi‐structured phone‐interviews of about half an hour withtherapistswhohadparticipatedinthestudy.TheinterviewsfocusedontopicssuchasthedesignoftheLucaapp(expectations,usability,look‐and‐feel)andtheiropinionsaboutitsintegrationintherapyandexperimentation.Atotalofnineparticipatingtherapists(onemaleandeight females)wereinterviewed.They had used the Luca app for therapy but had not yet implemented thegamifiedtherapygoalsmodulewithintherapy.Fivetherapistsworkedattheyouthmentalhealthcareinstitutionandfourattheyouthaddictioncareclinic.At the start of the interview, therapists provided informed consent forrecording the interview and the anonymous use of their data. We usedthematic analysis for the iterative analysis of the data [207, 208]. Secondly,similarcodesweregroupedtogetherintohigherlevelcodesandthemeswerecreated from the recurring codes.Wewent through our data several timesuntilwefeltthatourcodinghadachievedsaturation.Finally, thecodeswererefinedbytworesearcherswhocriticallydiscussedandreviewedthethemes.

5.5.2. ETHICALCONSIDERATIONS Variousstepswere taken toensure theconfidentialityandanonymityof theparticipantsintheinterviewsandtheethicalityofthisstudy.Participationoftherapists was voluntary and consent was verbally obtained. Data wereanonymizedattheearliestpossiblestage.


126

5.5.3. RESULTS Overall, three themes were identified on the basis of the results of thethematicanalysis.Thefirstwasrelatedtothefittingnessofgamificationwithindividual users. Patients in youth mental healthcare have various types ofmental health issues and disorders and it was a common opinion thatgamification might only be beneficial to a certain group of patients. Inaddition, gamification may be more suitable for younger adolescents thanolderadolescents, sinceyoungerpatientsmaynotbe intrinsicallymotivatedfor behavioral change and to actively adhere to therapy and correspondingtherapeutictasks.Therefore,youngerpatientsneedmoreextrinsicmotivationtoadheretotherapy,andthefeelingwasthatgamificationmightplayamoreprominent role with this age group. The second theme was related to thefittingness of eHealth with face‐to‐face practice. Blended eHealth has theadvantageofextendingthereachofpsychologicaltherapybeyondtheclinicalsetting[29].However,integrationwithcurrenttherapeuticpracticesremainscomplexandthepersonalizationofthedesignremainsanimportantaspectoftheimplementationofsuchtechnology.Inaddition,technicalinstabilityissuescanhavehugeimpactoneHealthandgamificationforaseriouscontextsuchasyouthmentalhealthcare,andthiscancausethetherapisttolosetrustinthedigitalsolution.Thelast themewasrelatedtoadistortionofthetherapeuticalliancecausedbyeHealthandgamification.At thestartof therapy,patientsareoftennotmotivated forbehavioral change.We found that the therapistspreferred to first motivate users during face‐to‐face contacts, by applyingmotivational interviewing and building up a therapeutic alliance, beforeemployingagamificationandeHealthapplication.

5 .6 .1 . F ITTINGNESS OF GAMIFICATION TO INDIVIDUAL USERS

The first theme identified from our analysis was related to the therapists’expectationthatthegamificationwouldonlybesuitableforacertaingroupofusers.During the interviews, five therapistsmentioned that thegamificationof the therapy goals module would be more suitable and motivational foryounger patients and less for older adolescent patients. According to thesetherapists,olderadolescentswouldbemoreintrinsicallymotivatedtoadhereto therapy than younger patients. Lower intrinsic motivation in younger


127

relativetoolderindividualsisalsoreflectedinotherstudies[322].Inthecaseofyouthmentalhealthcare,youngerpatientsmaynotseetheimportanceandnecessityofactivelyadheringtotherapyandcorrespondingtherapeutictasks.Therefore, younger patients requiremore extrinsicmotivation to adhere totherapy, which therapists feel might offer a more prominent role togamification.Inaddition,onetherapistfeltthatgamificationwascomparableto other techniques currently used in youth addiction care, such ascontingencymanagement,atypeofinterventionthatusesrewardstoenhancethemotivationofpatientstoadheretotherapy[267].

“Youngerpatientsoftenfinditmoredifficulttounderstandhowtogoaboutit,andgamification canmake itmore interesting.For example, if they can earnsomethingwithitthatcanmakeadifference.Olderpatientshavemoreinternalmotivation and younger ones have to be more extrinsically motivated.”(TherapistG,male)

Patients in youth mental healthcare can have a variety of mental healthdisorders,suchasanxietyandmood,impulsecontrolandaddictiondisorders(some suffer frommultiple disorders simultaneously) (e.g. [323]). AlthoughReadySetGoalswasdesignedforyouthaddictioncare,thetherapistsfeltthatthis gamification could also be applied to othermental healthcare domains.However, gamificationmightbemore suitable for certaingroupsofpatientswith specificmental disorders. One such observationwasmade concerningthegamifiedgoal‐settingmoduleinthisstudy.Twotherapiststhoughtthatthegamificationwouldbeparticularlymotivating forpatientswithanAttentionDeficit Hyperactivity Disorder (ADHD) or Attention Deficit Disorder (ADD).Patients with these conditions need either extra stimulation to becomemotivatedor tendtoachievehyper‐focuswithinspecific tasks.Researchhasshownthatwhengameelementswereaddedtocognitive tasks, therewasaslightalleviationofperformanceproblemsinchildrenwithADHD[324,325].However, in contrast, another therapist felt that the design of the therapygoalsmoduleandgamificationwouldnotbesuitableforpatientswithADHD.Inordertosettherapeutictasks,patientsneededtoclickthroughanumberofscreensandthetherapistfeltthatthiswouldcausepatientswithADHDtoloseattention.


128

“IfIlookatthetargetgroupADDandADHD,theyneedexternalinput.Inagametheycangetsomekindofhyper‐focusthatmakesitfun.”(TherapistF,female)

Anothertherapistmentionedthatgamificationwouldbebeneficialtopatientswithautism,sincetheycanbemorefocusedonthetaskthankstotheorderlyandcleardesignofthegamification.Intheliterature,however,thereislimitedresearch on the beneficial effect of gamification for patients with autism,thoughtherehavebeenpaststudiesthatsuggestthattheuseofgamificationmay help individuals with autism to recognize the emotion and facialexpressions of others [326]. Lastly, another therapist thought thatgamification would be motivating for patients with a game addiction orsymptomsofcompulsivegaming1.Thistherapistalsothoughtthattheappwasnot suitable for patientswith suicidal thoughts (another exclusion criterionfor the study). When patients with suicidal thoughts experience a crisis, atherapistneeds toreact immediately.However, ifa therapistdoesnotchecktheappdaily,andthepatientcontactsatherapistthroughtheapp,thereisachance that these crisis situations are missed. Therefore, it would bepreferablethatthesepatientscallthetherapistratherthanusingtheapp.

“Autismpatientscanbemore focusedonthat[orderandclearnessofgames]”(TherapistC,female)

“…andIthinkthatespeciallygameyoungsters[areeasilymotivatedbyit].It’sashame[thattheyareexcludedfromparticipation]”(TherapistB,female)

Onthesubjectofspecificgameelements,onetherapistfoundtheprogressionfeedback system with the mountain to be very effective, especially whenpatientsdonotaccomplishtheirtherapeutictasks.Shegaveanexampleofapatient with an anxiety disorder who was helped to confront fears usingexposuretherapy.Thispatientfeltasthoughshehadnotachievedanythinginher therapy, after failing to complete a therapeutic task. According to thetherapist, the mountain (the skill‐tree based progression mechanism) mayhelp thesekindsofpatients, since it registers smallprogressionsandmakes

1Patientswithagameadditionorcompulsivegamingwereexcludedfromthestudy,mainlybecausethesetypesofpatientscouldbemoreordifferentlymotivatedbygameelementsthanothers,influencingthe(generalisabilityofthe)results.


129

these progressions visible. In thisway, patients are better able to see someform of progression and achievement when they go to therapy sessions(because even though patients may not have completed a therapeutic task,goingtoatherapysessionisanachievementinitself).

“Thevisualizationofthemountainisveryeffective.”(TherapistD,female)

“Afternothavingachievedthenexttask,anexposurepatientfeltlikeshehadnotachieved anything.Themountain can helpwith this [feeling of not achievinganything].”(TherapistD,female)

5 .6 .2 . F ITTINGNESS OF EHEALTHWITH FACE‐TO‐FACE PRACTICE

MostparticipantsviewedandevaluatedtheapplicationasablendedeHealththerapy tool. The Luca application was designed in such a way that it cansupport face‐to‐face therapy. For example, patients can keep track of theirgoals, emotionsandsubstanceuse,which therapistscanalsoreviewoutsidetherapy sessions. Prior studies have shown that using digital technology insuch a way to deliver mental healthcare (also referred to as eHealth) canimprovepsychosocialtherapy[24‐26]inadults(e.g.[37‐39]),adolescentsandchildren(e.g.[40][41]).

“Thefactthatitissomehowdigital.Normallyyougivelotsofdifferentkindsofstencils onwhich they [patients] have to note things down. This is just theirmobilephone,which theyalreadycarryaround in theirhands.” (Therapist H,female)

Itisunderstandablethatthetherapists’evaluationsplacedastrongemphasison the ease of integration of the application into therapy as a part of thetreatmentprocess.Despitetheirpositiveexpectations,therapistsdidnotusetheappfrequently.Sixtherapistsfounditdifficulttointegratetheapplicationwiththerapeuticpractice,especiallywhentherewasadivergencefromaone‐to‐one translation of the CBT protocol. For instance, the application usesdifferent words for specific assignments than the commonly used therapyprotocol,andthiscausedconfusion. Inaddition, theuseof theLucaappandexplainingtopatientsabouttheLucaexperimentwasnotapartoftheusualroutine of the therapists. Combinedwith their already busywork schedule,


130

this caused some of the participants to attribute a low priority to the Lucaapplication during their therapy work. All therapists mentioned they werebusyatworkandsixmentionedthatasaresulttheyhadgiventheLucaapplowpriorityand/orhadforgottentointroduceLuca.

“TheappshouldbeidenticaltotheCBTprotocol.Nowitissomethingaddedontotheprotocol,withdifferentformulationsanddifferentwaysofwritingthingsdown.Thisconfusesusaswell..”(TherapistsE,female)

“Ihavea lot to thinkabout,and theLucaexperiment isnotmy firstpriority.”(TherapistC,female)

“It’s easier to add some noteswhen using paper. That iswhat I foundmostdifficult.InoticedthatIammoreintothepaperversion.Ithinkyoujusthavetogetusedtoit.Ifyoudoitforalongerperiodoftimeitbecomeseasierandmorefamiliar.”(TherapistH,female)

However,thiswasnottrueforalltheparticipants,astwotherapistsexplicitlymentionedtheywereabletointegratetheappanduseitproductivelywithintheir therapy. For example, one therapist noticed that when using the appduring therapy, the therapy went faster and that the patients were able toworkwiththeappefficiently.

Six therapists also found that there was not enough personalization in thedesignofthetherapygoalsmoduleandthatitwastoostructuredand/ortoodifficulttouse.Thetherapiststhoughtthatitwasnoteasytosetagoalandataskandtheydidnotlikehavingtoselectarelevantareainlifewhensettinglong‐term therapy goals. In addition, they thought that the evaluationof thetherapeutic tasks and goals was not handily implemented. However, theyfoundtheideaofthetherapygoalsmoduleuseful.Forexample,onetherapistmentioned that in face‐to‐face therapy the overview of goals and tasks cangive patients the chance to see a pattern that reveals why some goals andtasks are or are not achieved and whether there are specific situations inwhichachievingthosegoalsandtasksbecomesmoredifficult.


131

“Regardingthetherapygoals:Ididnotapplyit[withpatients]becauseitwassocumbersome,andIthoughtthatifevenIlosetheoverview,thenmypatientswillforsure.”(TherapistG,male)

“Therewasnoroomforpersonalization.Youhadtochoosefromareastodefinea therapygoal.Sometimes I justput thegoal somewhere,even if itwasnotaperfectfit,soIwascreativewithit”(TherapistB,female)

“How goals are done now is quite complicated, it’s not how Iwould like it.”(TherapistG,male)

“Inoticedthattherewasamaximumnumberofcharactersthatyoucouldtype.Sometimesthiswasabittoolimitedforme.Sothenyouarenottryingtothinkofasuitablegoalbutyou’rethinkingofagoalthatfitstheLucaapp.”(TherapistI,female)

Fromthedataoftheinterviews,wealsoobservedmultiple instancesofhowtechnical instabilityissuesdecreasedtheleveloftrustintheapplicationasasupportivetoolfortherapy.Especiallyatthestartoftheexperiment,theLucaapp suffered from technical instabilities such as crashing during use. Inaddition, there were a few instances when the therapists were unable toaccess the applicationdue to internet connectionproblems.Althougherrorsandbugsingamesarequitetolerablewhendesigninggamificationinanon‐mentalhealthcarecontext,theycouldhaveaconsiderableimpactinaseriouscontextsuchasyouthmentalhealthcare.Forexample, twooutof fiveof thetherapists who experienced technical instabilities mentioned that thisimpairedtheirtrustintheappasatherapysupporttoolanddecreasedtheirmotivationtousetheapplication.

“Inoticedtherewerea lotoftechnical issues,whichreducedmymotivationtouseit”(TherapistF,female)

Most of the technical difficultieswere encountered in the early stage of theusage. For some therapists this may have influenced their motivation andwillingnesstousetheapplication.However,tolimitthenegativeinfluenceonmotivation and willingness, technical support was provided promptly toaddresstheseissues.


132

Mostofthebenefitscitedbytheparticipantswererelatedtothemobilityandportability aspects of the application. All the therapists saw benefits in themobilenatureoftheapp,suchashowpatientscanberemindedoftherapeutictasksandgoals.Thetherapistsfeltthatthishelpedthemtomaximizethetimeutility of the therapy, as the app could link the therapy and therapyassignments to the home context of patients. They also thought that it wasmoreinteractivethantheoriginalworkbook.

“Great idea, this isgoing tobeagreatsuccess.Theyeasily forgetaworkbook,but as they always carry their phone around, so that problem is solved.”(TherapistA,female)

“Itwouldworkbetterthantheworkbook,generatemore information, itwouldbeabetterreminder forpatientsregardingwhathave I learned,whatdidwediscuss,whatweremygoals?And that I couldmakeuseof it inbetween [thetherapysessions]”(TherapistE,female)

Thechance forpatients toperformtheir tasks immediatelyandrecord theirprogress was another positive factor that was cited. In the Netherlands,around99%of theyoungstersandadolescentsaged12‐25yearsoldownasmartphone [29]. As most users already use their smartphone for otherleisure‐relatedactivities,theydidnothavetoaddanotherdevicetotheirdailyroutine. This makes it easier for them to perform therapeutic tasks moreregularly.

5 .6 .2 . DISTORTION OF THERAPEUTIC ALLIANCE BY EHEALTHAND

GAMIFICATION

Because the Luca experiment ran at a youthmental healthcare and a youthaddictioncarecenterwithdifferenttypesofpatientsandtherapyprotocols,itwas important that patients in the experiment received CBT at the earlieststage of their therapy. Otherwise, therewere toomany other elements thatcould influence theeffects.We thought that thegamificationwouldenhancethe motivation of patients for behavioral change. However, we observedseveralincidenceswherethisexpectationconflictedwithexistingtherapeuticpractice.


133

Our interview data showed that the therapists prefer to motivate patientsthemselvesbeforeemployingthegamificationandeHealthapplication.Duringtheinterviews,therapistsexplainedthatpatientsareoftennotmotivatedforbehavioralchangeatthestartoftherapy.Therefore,therapistsgenerallystarttherapy by first focusing on enhancing the patients’motivation by applyingmotivationalinterviewing.Inaddition,therapistsworkonbuildinginitialtrustwith their patient, allowing them towork together in an effectiveway. This“therapeuticalliance”isalsoanindicatorforpositivetherapyoutcomes[35].

“Makingcontact is important, Iwouldnotuse theLucaapp immediatelyafterintake.”(TherapistE,female)

“...andontheotherhand, ifIintroduceitduringtheintakeorthefirsttherapysession,they[patients]areabithesitant.SoIapplyitatalaterstage,andstartby enhancing motivation [by personal face‐to‐face therapy].” (Therapist G,male)

Despite the purpose of adding gamification to the application to enhancemotivation, the therapists seem to prefer to initially motivate patients fortherapythemselves,ratherthantorelysolelyonthegamificationaspectofthesystem. In the interviews, four therapists mentioned this issue explicitly.Enhancing the motivation of patients for behavioral change, and loweringambivalence to change, is often performed using motivational interviewingwhich improves therapy engagement and outcome (e.g. [327] for substanceabuse). As designers, we had thought that gamification would enhance thismotivation,butgamificationdesignershavetoberealisticaboutthepotentialandsuitabilityofgamificationintherapeuticpracticeandhavetoaccountforthe preferences of therapists in enhancing the motivation of users (i.e.patients).Itshouldalsobementionedthatsomepatientsremainunmotivatednomatterwhatsystemispresentedtothem.Fourtherapistsmentionedthatcertainpatientscannotbemotivatedtoadheretotheirtherapy.Insomecases,this is because patients are only motivated to adhere to their therapy in alimitedwayandwantto focusonworkingdirectlyontheirproblems,ratherthan engaging in additional activities, even those which can make therapymorefun.


134

“Some of the youngsters thought itwas really a significant achievement thattheywereintherapyatall,anddidnotwanttoinvestanymoretime[usingLuca/participatinginLucaexperiment].”(TherapistB,female)

“…Iwasreallysurprisedthatsomepatientsthoughtthatit[usingtheLucaapp]wasnotforthem.”(TherapistH,female)

5.7. DISCUSSION:REFLECTIONANDLEARNINGSFOREHEALTHDESIGN

Overall,anumberofchallengeswereencounteredthroughout theprocessofdesigning the ReadySetGoals application. Similar to serious games andgamificationsinotherdomains,thereisanexpectationforthedesignedgameto not only be entertaining for users but also tomotivate them to realize aspecific,beneficialreal‐worldeffect[328].Inmentalhealthcaregamifications,these benefits and the gamified underlying activity are generally situatedwithin a therapeutic context where the treatment process involves activeparticipationfromvariousstakeholders(caremanagers,therapistsetc.),eachwith theirownsetofexpectations.Therefore, itwasessential to incorporatethe feedback of the patients as well as the therapy specialists andpractitioners,especiallyduringtheearlystagesof thedesignprocess,andtoestablish clear communication with the stakeholders. In particular, theformulation of a core game loop model to show how the game mechanicsrelatetotheunderlyingtherapeuticprocesswasusefulinbridgingthegapofunderstanding between professional game designers and clinical staff andfacilitatingafruitfuldiscussion.

5.7.1. MANAGINGSTAKEHOLDEREXPECTATIONSTHROUGH

FRAMING Onekeylearninghighlightedinthisstudyistheimportanceofcarefulframingand communication about the nature of a therapeutic gamification, to helpmanage stakeholder expectations.A particularly interestingobservationwashowthedual(andoftenunclear)natureofamentalhealthcaregamificationasbotha“game”anda“therapeutictool”causedaconflict inexpectationswithpatientsandcarestaff.Theexpectationsheldofa“therapeutictool”,namelytofulfilaseriousroleinsupportingtherapyandimprovingwell‐beinginthereal


135

world seemed inconsistentwith theexpectationsof a game,which ismostlyviewedasanon‐seriouswaytoprovideentertainmentforplayersinacontextdistinct from the real‐world. In addition, when patients viewed thegamificationasa“game”,itlackedtheusualcharacteristicsoftheirexpectationof “entertainment games” (by not having an immersive storyline etc.). Forthosewithhighexpectations,theseinconsistenciesmadethegamificationlessappealing.However, if the real‐world therapeutic aspect is less evident thanthegameworldaspect,patientsandstaffbecomeskepticalaboutthebenefitsandarelessinclinedtoacceptthegameintheirtherapeuticcontext.

One example from the qualitative evaluation with therapists was theirinterpretation of the gamification as a therapy‐supportive application. Mosttherapists viewed and evaluated the application from the perspective of ablendede‐healththerapytoolbutfounditdifficulttointegratetheapplicationintotherapeuticpractice.Forexample,theuseof languageintheapplicationcausedconfusionas itdidnot relate clearly to specific therapyprotocols. Inaddition, the structure of the applicationdid not fit the structure of specifictherapyprotocolsandtherapistswantedtoenhancethemotivationofpatientsfirst,beforeapplyingtheeHealthapplication.

Processessimilartothatofframingcouldbeparticularlybeneficialinhelpingto align user expectations. To enhance the framing of the app as a therapy‐supportive application inyouthmentalhealthcare, one shouldmakeexplicitwhatandhowsuch an app cando to support both therapists andpatients.ThiswillpreventtherapistsfromhavingotherexpectationsorconcernssuchasthateHealthandgamificationwilldistortthetherapeuticalliance.Thevalueofframinginmakingcertainaspectsofinformationmoresalientandthereforemorenoticeableandmeaningful toaudiences[329]couldalsohelphighlightthe perceived therapeutic value of the gamification and avoid raisingunintended expectations of it simply being a “game”. For instance, someresearcherscautionedonusingtheword“game”whendescribingthesystemtoplayersas it couldraiseunintendedexpectationsanddemotivate themtoengage with it [245]. Equally, we cannot frame the app as a “therapeuticapplication”.Inthiscase,therapistsmayhaveexpectationsoftheapplicationbeing amedical application that strictly follows a specific therapy protocol.


136

Whatwasparticularlyinterestingwashowthedesignofthegamificationitselfcould have a similar effect in “framing” the gamification for users. The staffseemed initially to bemore positive towards such “gamifications”when thegameplaymechanics could be easily understood and the underlying gamingactivitycouldberelatedclearly to thedesiredtransfereffect inareal‐world.This seems to allow the perceived benefits to be more easily visualizedcomparedtowhenthedesigniscenteredonagamificationactivitythatistoocomplexandcannoteasilybelinkedtoabeneficialtransfereffect.

5.7.2. INTEGRATINGTHERAPEUTICASPECTS INA GAMEWORLD

EXPERIENCE Anothercomplexity(i.e.wickedproblemaspect)encounteredwhendesigningmental healthcare gamifications is that, depending on the nature of thegamified therapyor activity, the scopeof the real‐world context involved inthe gamification can vary considerably. While gamification of simplercomputerizedtrainingtasks(suchasInterpretationBiasTrainingwhereusersare trained to interpret ambiguous scenarios in a positive manner) is lessinfluencedbythereal‐worldcontext(asthe interactionsrequiredtoachievethe transfer effect are fully confinedwithin the digital “gameworld”), thosewhicharedesignedtosupportmorebehavioralbasedtreatmentapproaches,suchas theReadySetGoalsapplication, canbesignificantly influencedby thereal‐world context. In such treatment, achieving the desired transfer effectrelies mainly on interactions in the real world. Thus, there is often arequirement for constant interaction between the game world and the realworld.

A number of instances were observed where this could have unintendedeffectsonthedesignofthegamification.Wheninteractionsintherealworldare included in the game world, it becomes more difficult for the gamedesigner to accurately structure the player’s experience in advance, as thegame world is no longer a closed environment. As a result, certain gamedesign approaches, such as thosewhich stimulate gameplays of progression(wherethegameplayexperienceisbasedonastructuredorsequentialsetofchallenges [330]), become more difficult to implement as this requires thedesigner to carefully control the scope of the players experience [331]. For


137

example, in the ReadySetGoals gamification, a key element to theimplementation of the risk‐taking mechanism is to understand howchallengingthetasksareforplayers,sothatmorepointscanbeawardedwhenplayers take risks to accomplish taskswithhigherdifficulty levels.However,since the real‐world experience varies greatly for each patient, it is verydifficulttoknowbeforehandhowchallengingthetaskswillbe.Insomecases,the real‐worldexperiencecanevenoverride theexperiencesof the intendedgamerules(forinstance,thereal‐worldtaskcouldbesoeasyastonegatethetime‐pressure challenge from the risk‐taking mechanic). Such problemshighlight the problem of accurately integrating real‐world aspects withinhealthcaregamifications.

Another example which we encountered was related to the motivationalaffordanceofgamification.Therewasaconflictconcerningwhetherincreasingthe motivation of patients should be the role of the gamification or of thetherapist. We thought that gamification would motivate patients forbehavioral change and that therapists would also motivate the patientsindependently. However, there was a problem when combining both.Accordingtothetherapistswhoparticipatedintheinterviews,theyfeltthatitwas necessary to first enhance the motivation of patients for behavioralchangeandtobuildatherapeuticalliance,beforeemployinggamificationandeHealthapplications.Wehadthoughtthatthegamificationcouldenhancethismotivationforbehavioralchange,butthisdidnotsuittheroutinesusedbythetherapists as part of their practice. Therefore, designers should be flexiblewhen applying and studying an eHealth application and gamification in thetherapeuticprocess.Forexample,byprovidingmorescope in thedesign forenhancingatherapeuticallianceduringtheinitialtherapysessions.

5.7.3. THEVALUEOFPERSONALIZATIONINYOUTHMENTALHEALTHCAREGAMIFICATION

Tobetterintegratereal‐worldelementswithinthegameworld,itisessentialfor the gamification to be able to correspond to a broad range of patientcharacteristicsandcontext.Providingareal‐worldexperiencethatuserscanrelatetowouldmakethemmoreawareofthepotentialtherapeuticbenefitsoftheyouthmentalhealthcaregamification,thushelpingtobettermanageuser


138

expectations.Intheinterviewstudy,thetherapiststhoughtthatthegamifiedand non‐gamified therapy goalsmodule of the Luca appwas too structuredand inflexible for a varietyofpatient types. Indesigning the application,weaimed for room for personalization of therapy goals but we also needed astructuretopreserveanoverviewoftherelationshipsbetweenlong‐termandshort‐term therapy goals. To this end, we structured the way long‐termtherapyrelatedgoalscanbesetanddefinedcategoriesforgoalsintheformoflife areas to help patients and therapists to set long‐term therapy‐relatedgoals.Wehaddiscussedproceduraltaskgenerationforgoalsetting,butfoundit too complex to generate goalswhichwouldboth fitwithuser interest (inthiscasethepatients)aswiththeoveralltherapeuticoutcome.However,thelife areas were too diverse and not always suitable for each therapy goal.Lastly,accordingtotherapists,gamificationmaynotbesuitableforeverytypeof patient: gamificationmay bemore suitable for younger adolescents thanolder adolescents, since younger patients may not see the importance andnecessityofactivelyadheringtotherapyandcorrespondingtherapeutictasks.

Oneapproachwhichwebelievecouldbeuseful inhelping tobettermanageuser expectations is the use of personalization, adaptation and tailoringtechniques in the gamification [204].This couldbedone either explicitly bymanuallypersonalizingthegamecontent(suchasthroughashareddecision‐making process with stakeholders) or more implicitly, through automaticadaptation based on a player modelling technique (see [332]). Adaptivetechniqueswhichareabletoautomaticallyadjustthelevelofchallengebasedon performance in the real‐world (such as [333]) could also be particularlyuseful.Suchmechanismsallowforabettercontrolovertheinfluenceofreal‐worldaspectsinthegame.

5.8. CONCLUSIONInthispaper,wediscussedindetail theprocessusedtodesign,develop,andimprove theReadySetGoals application, a gamifiedmobile applicationaimedat supporting therapy within youth addiction care. The overall processinvolved four key stages inwhich various formative researchmethodswereutilized with 53 participants from varying backgrounds (professional game


139

designers,therapistsetc.).Thesestagesincludeidentifyingthetransfereffect,investigating the real‐world context, creating a core game loop and iterativedevelopmentandtesting, firstwithgeneralusersandthenwith thetargetedaudience and care staff. Afterwards, we implemented the improvedReadySetGoalsgamificationasamodulewithinatherapysupportapplicationcalled“Luca”whichwasnotonlyusedinyouthaddictioncarebutalsoinyouthmentalhealthcare.

The overall designprocessprovided several valuable lessons and reflectionsfordesigningmentalhealthcaregamifications.Thisincludesthevalueofuser‐centereddesign approaches to incorporate stakeholder feedback throughoutthe design process and aligning user expectations through framing. TheResearch throughDesignmethodologywasparticularly useful in this regardwiththevariousartefactsandprototypesateachiterativestageservingbothto generate situated design knowledge as well as to align stakeholderexpectations.Inaddition,wealsodiscussedhowtheinteractionsbetweenthegame world and the real world could cause difficulty in gameplay design,especiallywhen real‐world elements are incorporatedwith traditional gamemechanics.

Usage of new eHealth by therapists is essential to successfully conduct anevaluationstudyoneHealthoutcomesandadoptionofeHealthbytherapistsisessential for future successful implementation of eHealth in a healthcaresystem. Even though we trained therapists in the use of the Luca app andgamificationandweremindedthemofthestudy,therearesomelessonstobelearnedfromthetherapiststhatwereinterviewed.Firstly,therapistsneedtobe better informed about the beneficial effect of gamification and thefittingnessofgamificationwith individualusers.Therapists thatparticipatedin the interviews thought that gamification would be more motivating foryounger adolescent patients compared to older ones. However, they do notknowforcertainifthisisactuallythecase.Sincetherapistsarefairlycautiousinapplyingnewtools in therapy,as theydonotalwayssee thebenefits, thegamificationshouldbedesignedinsuchawaythatitisflexibleandallowsforareturntoanon‐gamifiedinterventionifplayersdonotfinditmotivating.Weexpectthatthiswillmaketherapistslesscautiousinapplyinggamificationin


140

therapeuticpracticewithavarietyofpatients,astheycanalwaysgobacktoa“normal” non‐gamified version. In addition, it is important to manage thestakeholder expectations of eHealth by informing therapists in detail aboutthefittingnessofeHealthwithface‐to‐facepractice.Forexample,thefactthateHealth is not a copy‐paste of the therapy protocol but supports generalassignments and tasks of CBT that a patient can execute outside a therapyroom.Lastly, therapistsneedtoknowwheneHealthandgamificationcanbeapplied in the therapeutic process. For example, therapists thought thateHealth and gamification distorted the therapeutic alliance and wanted toenhance themotivationofpatients forbehavioralchange themselves,beforeapplying eHealth and gamification. Therefore, it is important to explicitlyallow therapists the freedom to apply eHealth and gamification when theythink it is most suitable in a design process, so that it does not distort thetherapeuticalliance.Eventhoughthismayconflictwithastudyset‐up,itwillenhance future adoption and the influxofnewparticipants in anevaluationstudy.

One of the limitations of this study is that the used researchmethodsweremainly generative or formative in nature. Although these methods areessential in providing answers to the questions encountered during thedevelopment of the gamification, more research is needed to properlyevaluatetheeffectivenessofthedesiredtransfereffectofReadySetGoals(see[334]). In particular, a key disadvantage of the Research through Designapproachistheblurringoftheoveralltransfereffect.Forexample,wouldthefactorsfoundinthisdesignstagethatmakethegamemoreentertainingalsomake the overall gamification more effective health‐wise? In the future,experimentalstudiesneedtobecarriedouttocomparetheeffectivenessofagamification integrated in a therapeutic approach as opposed to regulartherapy. Another limitation is that the presented case study focuses on aspecifictherapeuticdomain,i.e.youthmentalhealthandaddictiontreatmenthealthcare,andtreatmenttype,i.e.CognitiveBehavioralTherapy.Todevelopageneralizableframework,thesequentialdesignstagesneedtobetestedwithgamificationsdesignedforothertherapeuticdomainsandtreatmenttypes.


141

The key contribution of this paper is to provide more knowledge aboutdesigning gamifications for the mental healthcare domain. A detaileddescription of the structural process used to develop the ReadySetGoalsapplication as well as the reflected learnings of gamification design in thiscontext are provided. In addition, case‐specific examples of the variousmethods used to address the challenges encountered at each stage of thedesign process are also provided to addmore bottom‐up knowledge to thisdomain.Inthefuture,thestructuraldesignprocessesemployedinthisstudycan be used as the basis for the development of a generic framework forgamificationinmentalhealthcare.

142

143

6. GENERALDISCUSSIONAND

IMPLICATIONS

Thischapterdescribestheimplicationsofthepreviousdescribedchaptersandprovidessuggestionsforfutureresearch.

Chapter6|Generaldiscussionandimplications

144

The aim of this dissertation was to study the added value of personalizedgamificationineHealthasadesignfactortoenhanceimplementationpotentialofeHealthinterventionsinyouthmentalhealthcare.EHealthinterventionsinmental healthcare areoften focused on the therapeutic content andprovidelimited interactionmotivation for theusers causingmuchdrop‐out inusage(e.g. [335, 336]).When transforming an existing face‐to‐face therapy into ablended eHealth therapy, users should be motivated to use the eHealthproduct. Thismotivation canbe enhancedbyusing gamification as adesigntechnique. Based on theory and previous studies, Persuasive Game Designseemed successful to enhance thismotivation and consequently the efficacyand implementation of Persuasive Game Design (e.g. [63, 260, 337, 338]).Sincetherewasalackofresearchonsuccessfulimplementationtechniquesofgamified eHealth interventions, personalization was suggested as a designtechnique for a successful implementation. However, clear and sharedconcepts of what personalization entails and the effects on health relatedoutcomeswerelacking.

Sincethenatureandeffectofpersonalizationinpersuasivegamedesignhadnever been systematically studied, and thus making assumptions regardingthebenefitsofpersonalizationungrounded,weconducteda literature studyand developed a model to study the interaction experience, behavior andhealthrelatedtransfereffectsofpersonalizationinpersuasivegamedesignforhealthcare (Chapter 2). We defined ‘personalization’ in this thesis as theinvolvement of stakeholders (e.g. end‐users and domain experts) across thePersonalized Design Process (PDP) phases: Problem Definition, ProductDesignandTailoring.Ourliteraturereviewshowedthatinthecurrentdesignprocesses stakeholdersweremost often involved in theProductDesign andTailoringphase.Becauseamajorityofthestudieswereoflowmethodologicalquality, we could only suggest that it is important to involve stakeholdersacrossthePDP‐phasesinordertoincreasethealignmentoftheproduct,theinteraction time with the product, and consequently to positively influencethehealthrelatedtransfereffect.

Personalization is also often applied in mental healthcare, such as in theapplicationof therapyprotocols inclinicalpractice.Therapistsoftenadapta


145

therapy protocol to the individual situation of a patient, e.g. by using thetherapy protocol as a “toolbox” and choosing the tools they think fit thepatientsbestorcombiningelementsfromtherapyprotocolstotargetmultipleproblems of the patient at the same time. Designers of eHealth for mentalhealthcareoftenrelyontheexistingtherapyprotocols,consistingof therapysessionsinwhichspecificthemesarediscussedthatareintendedtoevokethetherapeuticeffects.However,modificationoftherapyprotocolsbytherapistsandpatients is commonpractice [196]. If eHealthdesigners donot take thetherapeutic practice of therapy protocols into account, it ismore likely thatthefinaleHealthdesigndoesnotsuitthetherapeuticpracticewhichinitsturnnegativelyinfluencesimplementation.

Toproviderecommendations foreHealthdesignersonhowto take this intoaccount,wewanted to generate information about theproportion, type andreasons forpersonalizationofagiven therapyprotocolbyconducting focus‐group discussions with patients and therapists in youth addiction care(Chapter3). Results showed that therapists andpatients bothpersonalizedthe therapy protocol. In addition, both therapists and patients adapted thetherapy and added other non‐protocol therapeutic parts. Based on theseresultsthefollowingrecommendationsforeHealthdesignerswerepresentedto enhance alignment of eHealth to the therapeutic practice andimplementation: a) study and copy at least the actual applied parts of atherapy protocol in eHealth, b) co‐design eHealth in such a way that boththerapists and patients can personalize specific parts of the final eHealthdesign, and c) investigate if parts of the therapy protocol that are notpresently applied by therapists or patients should be part of the eHealthapplication.

However, itwasunknownhowdesignmethodscouldbeappliedforeHealthgamificationwithinayouthmentalhealthcarecontext.Therefore,weusedaspecific design method with cards that represent playful experiences, toexamine whether the input of playful experiences was also experienced byother end‐users from the same context in the actual design itself (Chapter4A).Wefoundthatusingonlyonedesignmethodtoenhancepersonalizationis not sufficient, since the experiences that were used in the design of the


146

prototype did not correspond one‐on‐one with the experiences that werereportedbyotheruserswhoplayed theprototype.Therefore,weconcludedthat stakeholders (e.g. users) should be involved in multiple phases of aPersonalized Design Process and not only at the start, to ensure that theproduct isstillalignedtopreferredexperiencesandlimitpossible individualpreferencesof stakeholder types that cannotbegeneralizable to thespecificstakeholdertype

We also wanted to study the effect of a specific design element in a youthmental healthcare context. Themotivationof patients touse eHealth canbeenhancedbymaking itmoreappealing,byapplyingPersuasiveGameDesign(PGD).Forexample,byusingchallengeandcompletionasgame‐elements toengagepatientswithcysticfibrosistocarryoutflow‐volumetests,wherethepatients assumed the role of a fireman and had to blow on a breathingapparatusinordertoputoutfireinthegame[260].However,itwasunclearwhat game‐elementswould be suitable for gamification in thewhole youthmental healthcare. Therefore, we studied if rewards (the most usedmotivator)wouldbesuitabletoapplywithinasubstancedependencetherapycontext,aspatientswithsubstance‐relateddisordersmaybelesssensitivetonon‐drug‐related rewards than patients without a substance use disorder.Therefore, we studied differences in reward type preferences betweenyoungsters in therapy for substance dependence and youngsters without asubstance use disorder (Chapter4B). Results suggested that, in contrast toour expectations, substancedependentparticipantswerenot lessmotivatedby the typesof rewardscompared tonon‐substancedependentparticipants,andevenmoremotivatedbymonetaryrewards.

ThepreviouschaptersprovidedargumentationforpersonalizationtoenhanceimplementationofgamifiedeHealth.Thiswastakenintoaccountinthedesignprocess of an eHealth application for youthmental healthcare (Chapter5).However, evenwhen eHealth is personalized and gamified, implementationcan still be negatively influenced by negative expectations of stakeholdersabout the effect of a therapeutic gamification, a limited integration of theeHealthproductwithin current therapy, anda lackofpersonalization in thedesign of an application. In addition, the way an evaluation is set‐up can


147

negatively influence implementation of an eHealth application.Personalization to the context of application is thus neededwhendesigningeHealth,gamificationandsetting‐upanexperiment.ThisisusefulinformationforfutureiterationsoftheeHealthapplicationtoenhanceitsimplementationpotential.

In sum, this thesis has provided knowledge on how to improveimplementationpotentialofeHealthwithinayouthmentalhealthcarecontext.Since not all game‐elements are suitable ormotivating for specific users oruser‐groups, the gamified eHealth product should be personalized. Thispersonalization process can be structured by using the PersonalizedDesignProcessmodelwehavedeveloped.

Theresearch that isdescribed in thisdissertationwaspartof theNextLevelproject.ThebroadergoalofNextLevelwastogenerategamedesignprinciplesformentalhealthcarerelatedeHealth,validatetheaddedvalueofgamedesignin eHealth, and to study whether implementation of eHealth could befacilitatedbypersonalizationdesignin(youth)mentalhealthcare.

6.1. IMPLICATIONSThestudiesthataredescribedinthisdissertationhavetwomainimplicationsfor eHealth design in youth mental healthcare. The first focusses onpersonalization ingamificationofeHealthandthat themotivationaleffectofgamificationcanbeenhancedifstakeholdersaremoreactivelyinvolvedinthedesignphasesofagamification.Thesecondimplicationnoticesthattheset‐upof effect‐studies should be adapted to the context of application to limitinvasivenessoftherapeuticpracticeandtoenhancefeasibilityofthestudy.

6.1.1. IMPLICATIONSOFPERSONALIZEDGAMIFICATIONINEHEALTH

The implementation potential of a gamified eHealth can be enhanced ifstakeholders are actively involved in the design process. Gamification canenhance motivation of users to use eHealth, but the motivational effect ofgame‐elements can differ across users and user‐groups. When a gamifiedeHealthproductisnotonlydesignedinaco‐creativewayanditispossibletotailortheend‐producttoindividualusers,itismorelikelythattheproductis


148

betteradoptedbytheend‐users.Inthisway,theynotonlyseethenecessityoftheproblemthat theproduct tries tosupport,butarealsomotivated tousetheproductbythespecificgame‐elements.

Personalizationin(youth)mentalhealthcare

In (youth) mental healthcare, it is currently unclear what exact therapyprotocolelementsarecrucial forenhancingtherapeuticeffects.Forexample,research in therapy forpsychosis showed thatdeliveringbothcognitiveandbehavioral techniques (e.g. enhancing self‐regulatory strategies) wasassociated with better therapy outcomes compared to partial therapy thatonly involved engagement and assessment [339]. However, what specificpartsofthetherapysignificantlyimprovedtherapy‐effectsremainedunclear.This can be studied by eliminating and adding individual therapy protocolelementswhileatthesametimefocusingonthetherapeuticeffect.However,since protocol elements interact, it may never be fully clear what protocolelementsensuretherapeuticeffects.

Therapy protocols are often personalized in face‐to‐face (youth) mentalhealthcare practice by both therapists and patients tomatch their personalpreferences and situation (e.g. [340]). This can facilitate protocolimplementation[204],enhancepatients’engagement[210,216,341,342]andpositively influence the therapeutic alliance [343]. Therefore, informationregardingcrucial therapyprotocolelements is important for information fortherapists, so they know which elements they can and cannot personalizewhile still ensuring therapeutic effects. However, it is also importantinformation for eHealth designers to limit iterations in a design, especiallysince there often seems to be little timeormoney to implement changes inalreadydesignedproducts[202].ManyeHealthinterventionsarebasedonaone‐size‐fits‐all approach, e.g. by copying the therapy protocol, and are notpersonalized to the user and user context. This may enhance the patient’sfeelings that the eHealth product is unresponsive to their individual needs[344]. In addition, not aligning the eHealth product to therapeutic practicemay limit implementation. For example, when eHealth does not suit howtherapistsusethetherapyprotocol(makingtheinterventionlessflexibleandpersonalized,e.g.tomaintainorenhancemotivationinpatientstochange)or


149

if therapists have negative expectations about the benefits of eHealthcompared to face‐to‐face therapy [199‐201]. For successful implementationwithinthiscontext,boththerapistsandpatientasuserareimportanttotakeintoaccount.

DesigningblendedeHealthtosupportyouthmentalhealthcare

Therapists noticed that patients in youth addiction care found it difficult tothink of non‐substance related activities. We therefore decided to design ablendedeHealthtoincreasethereachoftraditionalface‐to‐facetherapy,andto support homework compliance of therapy [345]. More specifically, wefocused on leisure‐related goal‐setting and provided patients with pre‐setleisure goals they could select. However, patients did not find these goalschallenging nor personally relevant and wanted more freedom in settinggoals.Animportantlessonthatcanbedrawnbythis,isthatwhatanend‐userwantsdoesnothavetobeinlinewithwhatothersmaythinkend‐usersneedandthatitisimportanttotakethetimetogettoknowthisuser‐information.We therefore enhanced personalization in goal‐setting by letting therapistsandpatientschoosetheirowntherapygoalsinsteadoflettingthemselectpre‐setgoals.

Wedesigned theReadySetGoals application to support goal‐setting in youthmentalhealthcare,amethodthatisusedintherapyformultipledisordersin(youth)mentalhealthcareandothercontexts.However,itispossiblethatthemotivatingeffectsofthechosengame‐elementsand/ormethodofgoal‐settingaredifferentbetweendifferent typesofpatients inyouthmentalhealthcare.For example, patients with addiction problems generally have significantlyhigher levels of impulsivity and sensation‐seeking personality traits [346]whilepatientswithdepressionproblemsgenerallyhavehigherlevelsintraitssuchasneuroticismandconscientiousness.Itcanbeassumedthatindividualswith different personality traits are more motivated by different game‐elements. For example, individuals with higher sensation seeking andimpulsivitylevelsseemtobemoremotivatedbya“betting”systemcomparedto individualswithhigher levelsof conscientious [347].Oneoption to coverthis,isbytailoringthegame‐elementstothepersonalitytypeofthepatient.


150

TailoringgameelementsineHealth

Whentailoringaproductto individualusers,onecanalsochoosetoconnectgame‐rewards to personality traits. Tailoring games to specific personalitytypeshasalreadybeendoneintheHexadframework[282].Thisframeworkcategorizesusers into six typesofplayerpersonalities loosely related to theBig Five personality traits. In addition, research has shown that thepersonalityofindividualsplaysalargeroleintheperceivedpersuasivenessofvariouspersuasivestrategiesusedinhealthcarecontexts[347].Furthermore,researchshowedthatindividualswithhigherlevelsofsensationseekingandimpulsivityaremoremotivatedbythrillanduseofrewards[298].However,itmaybe thatpersonalityprofiles, asdefined in “real life”, aredifferent in thecontextof (serious) games.Forexample, someonemightbea socializer in a(serious)gamecontextbutanachieverina(reallife)workingcontext.Auserthus canhavemultiple identities, both in a real and gameworld, since theymightwanttoachieveandexperiencedifferentthingsindifferentcontexts.Itmayalsobethatthereisoneconsistentfactorpresentinbothworlds,butthisshouldbeinvestigated.

Goal‐settingindifferentcontexts

Thegeneralstructureofgoal‐setting,firstsettinglong‐termgoalsfollowedbyshort‐term tasks that work towards the long‐term goal(s), is the same indifferent contexts. However, there are also some differences in goal‐settingacross contexts. For example, goal‐setting in a face‐to‐face physicalrehabilitation context is different from goal‐setting in a youth mentalhealthcare context. In physical therapy, a physical therapist/physiotherapistfirst identifiesthephysicalgoal(s)ofthepatient(e.g.beingabletoplayfieldhockey again) and sets‐up a plan of shorter‐term goals that need to beachieved inorder to reach the long‐termgoal(s).Apatienthas tophysicallyperformthetasks,inordertoreachthelonger‐termgoals.In(youth)mentalhealthcare,thegoalsareadifferenttypeofbehavioralgoals(e.g.donotsmokecannabis),butintheendaimtoimprovethepsychologicalaspectofapatient(e.g. improve a relationship with a family member). Therefore, it may beinterestingtodesignagamifiedeHealthapplicationtohelppatientswithgoal‐setting in a different context. Since the ReadySetGoals application already


151

motivatedpatients in youthmental healthcare to set and achieve goals, andtheproblem (enhancingmotivationof users to set and achieve goals) is thesame in different contexts, other users should be involved in the productdesignphasetoimproveandaligntheReadySetGoalstothenewusercontext.These other users from a different context should first evaluate the alreadydesignedReadySetGoalsapplication.Firstly,toseehowsuitablethewaygoal‐setting is designedwould bewithin their context and secondly, to see howmotivatingtheyfindthegame‐elements.

EffectsofpersonalizedgamificationineHealth

TheremaybemultiplereasonswhypersonalizedgamificationineHealthmayhave a positive effect on interaction motivation and corresponding healthrelatedtransfereffects.WithpersonalizedgamificationofeHealth,onealignsthedesigntothepreferredexperience,capacityandcontextofauser(e.g.assuggestedby[202]).Often,eHealthdesignfocussesmoreonthecontent(suchas therapy techniquesandelements) thanon the interactionwith theactualuser [348]. This may result in a mismatch between the design, userexperience,capacityandcontextandasaconsequenceenhancedrop‐outratesand lower implementation potential. For example, individuals with specificpersonality traits aremoremotivatedby someexperiences than others (e.g.[347]). In addition, some contexts of use may bemore suitable for specificdesignsthanothers.Forexample,anapplicationthataimstoenhancephysicalactivity is more suitable for someone that has the space and freedomcomparedtosomeonethathastostayinahospitalbedforalongerperiodoftime.

Often,theproblemthattheto‐bedesignedproductaimstosolveisdefinedbyothers than the user (e.g. designers of therapy programs, managers, etc.).However, it is possible that the problem is not experienced by the usersthemselves. This will lower implementation, because they do not see theimportance of the eHealth product. Therefore, future eHealth designers areadvised to also involve both patients and therapists when choosing theproblemthatagamifiedeHealthproductwillfocuson.Especiallyexperiencedpatientsshouldbeinvolved,sincetheyalreadyhaveexperiencewiththerapyand the difficulties they encountered during their therapy process. For


152

example, these type of patients know what therapy‐items influenced theirtherapyadherence,andcanhaveaclearerideaofwhatisrequiredwithintheeHealth design to support them within this process. Involving patients isalready being done within a youth mental healthcare context, where theproblem that mental health therapy will focus on is defined at the start oftherapybyboththerapistsandpatients.Apatientusuallyfirstpresentswithone ormultiplemental health issues and the therapist and patient togetherdecide what type of therapy will be used to tackle the mental healthproblem(s).Thisisdonebyusingsuggestionsfromthetherapistsinordertoselect the right evidence‐based type of therapy and to tackle the patient’sproblemandthepatientpreferences[349‐351].Inthisway,thetherapysuitsthecontextandcapacityofthepatient.

LimitationsofpersonalizationingamifiedeHealth

WhendesigningthegamifiedeHealthproduct,oneshouldkeep inmindthatthereisalimitationregardingtheamountofpersonalizationduringthedesignprocess. Stakeholders, like therapists and patients, should be involved andprovideinputforthedesigner.However,theactualdesignshouldbedonebythe designers since they are the most experienced in this aspect (e.g. toimplement stakeholder’s insights into thedesign).Therefore, it is importantforadesigner toguideusersandotherstakeholders in thisprocessandasktherightquestionsabouttheirinterests,capacityandcontext.Enhancingtheengagement of specific individuals within a user group can be done bytailoringtheproduct.Thisisimportant,sinceaproductwillneverfullysuitallusers within a user group. By tailoring a product, one can focus more onenhancing theengagementof specificusers.Next to thebeneficialeffectsonmotivation and health related outcomes, the added value of this type ofpersonalization as a design technique is also the easiest type to test. Forexample,bycomparinganeHealthapplicationthatusesonlygeneraltherapypartswithaneHealthapplicationthatalsoconsistsofotherproblem‐specificmodulesthatcanbetailoredtotheindividualpatient.

Most importantly, designers and researchers should make explicit what ispersonalizedinaneHealthdesign.Thiswillbenefitfurtherstructuringofthepersonalizationprocess,whichis important fordesignersandresearchersto


153

clarify how they have applied personalization and what personalizationaspects are effective. Personalization in the first phase of a PDP (problemdefinition)maybemoredifficultcomparedtopersonalizationintheothertwophasesofaPDP.Whendefiningtheproblemthattheproductaimstosolve,itis importantthattheproblemisalsoacknowledgedbytheusersthemselves.Otherwise,itmaybepossiblethataproductisdesignedbutnotexperiencedasneededornecessarybythetargetgroup.

EffectsofpersonalizedgamificationineHealth

However,theeffectofpersonalizationcomparedtonopersonalizationacrossawholePDPhasnotbeenstudiedyet.Theresultsofthestudiesdescribedinthis dissertation only suggest that personalization across all phases canenhancemotivationand/orengagementto interactwithaneHealthproduct.In the literature, personalization has shown beneficial effects in productdesignandtailoring.Forexample,byusinghealthrelatedbehaviorsofuserstomotivatethemforweightloss[352]ortailoringcontenttomotivateuserstoadhere to theirmedicationbyusing theirown input [353].An interestingquestion that arises when focusing on the PDP, is when there is enoughpersonalizationinaproduct.Forexample,ifonehastopersonalizeaproductbasedonage,typeofdisorder,culturalbackground,and/orthegenderoftheuser.Ingeneral,thereshouldbeabalancebetweentheamountofmoneyandtimespentandtheamountofpersonalizationthatisputintoaproduct.Thisisbecauseitisimpossibletoperfectlypersonalizeaproductandthereshouldbeamomentwhenpersonalizationcanbeseenas“enough”.

Concluding, this dissertation showed that the implementation potential of agamifiedeHealthcanbeenhancedifstakeholdersareactivelyinvolvedinthedesign process. Since eHealth is often focused on the therapeutic contentinstead of enhancing interaction motivation, personalized gamification canenhancemotivationofuserstouseaproduct[298,318]andthattheyseethenecessityoftheproblemthattheproducttriestosupport.However,theeffectofpersonalizationcomparedtonopersonalizationacrossawholePDPhasnotbeenstudiedyetandresultsofthestudiesdescribedinthisdissertationonlysuggest that personalization across the whole PDP can enhance motivationand/orengagementtointeractwithaneHealthproduct.


154

6.1.2. IMPLICATIONSFORTHESET‐UPOFFUTUREEHEALTHEVALUATIONMETHODSWITHINTHIS FIELD

Whentryingtostudytheaddedeffectofgamificationandpersonalization,wenoticed that the set‐up of the evaluation study was not matching thetherapeutic practice. This is needed to limit invasiveness in the context ofapplicationandtoenhance implementationpotentialof thestudy in therapypractice.Iftheset‐upofanevaluationmethodisnotalignedtothecontextofapplication, itwill bemoredifficult to successfully run the study. Especiallywhen therapists are crucial for the influx of patients, and the set‐up of theevaluationmethoddoesnotsuitcurrenttherapeuticpractice,theywillbelessmotivated to do extrawork such asmotivating patients for participation oruseanextraproductduringtherapy.

WhentheReadySetGoalswasimplementedintheLucaappasatherapy‐goalsmodule, we wanted to study the beneficial effect of gamification. Being thegolden standard evaluationmethod,we therefore started a non‐randomizedpre‐post study and contrasting the Luca app with a gamified therapy‐goalsmoduletotheLucaappwithanon‐gamifiedtherapy‐goalsmodule.However,theinflowofpatientsintheLucastudywaslow,andatthetimeofwritingitwasuncertainwhether the study couldbe finalized asplannedaccording toourprotocol.Sincepatientsfrombothyouthaddictioncareandyouthmentalhealthcarewouldparticipateinthestudy,wethoughtitwasimportanttokeepsomeelementsintheset‐upofthestudyasconstantaspossible.Forexample,patientscouldonlyparticipateiftheywereatthestartoftheirtherapyandiftheyreceived individualCognitiveBehavioralTherapy.However,wenoticedthatbecausepatientscouldonlyparticipate if theywereat thestartof theirtherapy, therapistsoftendidnot informtheirpatientsabouttheexperiment.Therapists found itmore important to first build a therapeutic relationshipwith their patients, one of the factors related to better therapy retention ofpatients (e.g. [354]),before theystarted therapyandwanted tomention theexistenceof theapplicationandstudy.Whenweset‐upthestudy,weshouldhavetakenthisintoaccountandmakethestartofastudymoreflexible(e.g.start introducing a study after the therapeutic alliance has been build). Inaddition, we should have made clear what we meant with CBT, sincetherapists foundthedefinitionofCBTunclear.Forexample, inyouthmental


155

healthcare different types of therapy protocols are often combined due tocomorbidities, compared to youth addiction care. Therefore, the content ofwhatisconsideredCBToftendiffersbetweenthetwocontexts,whichmakesitimportant for researchers tomake it explicitwhen they consider CBT to beCBT.

Ifastudyset‐upisnotinlinewiththecurrenttherapeuticpractice,therapistsandpatientsmaythusnotbefullywillingtoparticipateinanexperiment.Inaddition,differenttypesofevaluationscouldbemoresuitablewithinspecificcontexts than others. The classical validationmethods that use randomizedcontrolledtrialproceduresdonotseemtobesuitableenoughtomeasurethepotential of gamification and serious games [178]. For example, since thepractice of RCTs is different and much stricter compared to therapeuticpractice. Therefore, one should try to find methods that can study thebeneficialeffectsofpersonalizationwithinthiscontext.Forexample, itcouldbe that multiple N = 1 studies are more suitable within this context. WithmultipleN=1studies,lessparticipantsareneededwhoaremoreintensivelyfollowedcomparedtothoseinRCTs.Thiswouldindeedmeanthatmoretimeandeffortisneededbyparticipatingpatientsandtherapists,butalsothatonlya few participants are needed to test the effect of an intervention. Forexample, if a few therapists participate in a study theywould only need toinvolve one or two patients instead of a large number. However, it is alsopossible thatmultiple N = 1 studies are less suitable for this context, sincepatients oftenhave comorbidities. Thiswouldmake it difficult to generalizetheresultsofafewpatientstoawholepatientpopulation.Toreducetheriskthatpsychologicalimprovementsareduetothetreatment‐as‐usual,itcanalsobesuitabletousemultiplebaselinestudieswithina(youth)mentalhealthcarecontext.Multiplebaseline studies, sameasmultipleN=1 studies, aremuchmoreflexibleinimplementationandevaluationofinterventionsthanthelargestudiessuchasRCT’s[355].Withthesetypeofstudies,themomentapatientstartswithparticipatinginthestudydiffers.Inthisway,thebestmomentforimplementing the product will be known. This knowledge can be used bytherapists. In addition, the continuous assessment allows detailedexamination of patterns of change over time. However, results of multiple‐baselineandN=1studieslackgeneralityofobtainedeffectswhichisespecially


156

a limitation in a contextwhere individuals can havemultiple problems (i.e.comorbiddisorderssuchasinyouthmentalhealthcare)[355].

Another possibility that can enhance the feasibility of a study set‐up is theinvolvement of likewise participants that will participate in co‐creating theset‐up of a study. Especially in studies that are run in mental healthcarecontexts,therapistsaretheoneswhohavetorecruitpatients.Sincetheyareexpertsincurrenttherapeuticpractice,theycanhelpinaligningtheset‐upofan experiment to this context.Whenwe interviewed therapists in the focusgroupdiscussions,wediscoveredthattheirmainreasonfornotexplainingtheexperimenttotheirpatientswasthattheywantedtoenhancethetherapeuticalignmentwithandmotivationoftheirpatientsfirst.Ifweknewthatthestartof the Luca experiment should have been with this offset of delay in thetherapeuticprocess,thisthresholdcouldhavebeenprevented.

6.1.3. IMPLICATIONSANDRECOMMENDATIONSFORFUTURERESEARCHANDEHEALTHDEVELOPMENT

Thestudiesdescribedinthisdissertationarerelevantfordesignresearchersandhealthcareprofessionals.Designresearchersnowhaveanunderstandingofpersonalizeddesigninhealthcare(Chapter2)andknowhowtheycanalignadesigntoyouthmentalhealthcare(Chapter3). Inaddition, theyknowthatdifferent game design methods are needed when focussing on thepersonalisationofexperiencesindesign(Chapter4A)andthatgameelementscan be implemented to motivate youngsters in an addiction care context(Chapter 4B). Lastly, they can use the learnings from our exemplary designimplementation case (Chapter 5). Healthcare professionals now haveadditional information that there is a difference between therapy protocolsand the therapeutic practice of these therapy protocols. They can use thisinformation to improve therapeutic practice, e.g. by updating therapyguidelines, providing training and/ormore supervision to ensure evidence‐based therapeuticpractice (Chapter 3). In addition, healthcareprofessionalsknow that rewards are as motivating for youngsters with substancedependence compared to non‐dependent youngsters and that monetaryrewardsareevenmoremotivatingforyoungsterswithsubstancedependence,even though youngsters with substance dependence have an overall


157

decreasedrewardsensitivity.Thismightbeanextramotivationforhealthcareprofessionals to apply (monetary) rewards to motivate youngsters withsubstancedependencetoengageandremainintherapy(Chapter4B).Lastly,healthcare professionals can use our case‐study learnings for developmentandimplementationofgamifiedeHealth(Chapter5).

Based on the description of the design process of our gamified eHealththerapy,we recommend design researchers to link high quality research tothe development of gamified eHealth. More precise: design researchers arerecommended to study the effects of the whole design and the effects ofspecificdesign‐elements (e.g. the effectsonwillingnessormotivation tousethe design, health‐related effects, etc.). Currently, we cannot draw definiteconclusionsaboutwhetherandwheninvolvingstakeholdersinaPersonalizedDesign Process has added value in terms of effect. Therefore, research isneededtotesttheeffectoftheproductandtoensurethattheaimisachieved.Today,manyeHealthapplicationsareonthemarketofwhichtheefficacyhasnot been established. This is really important, in order to prevent possiblenegativeeffects[356],especiallyifusersonlyrelyoneHealth.Since2018,TheNationaleHealthLivingLab(NeLL)facilitatesthedevelopmentandvalidationofandresearchoneHealthinmultidisciplinaryteams.Forexample,theyhavefound that patients with COPD cannot trust on eHealth apps and websitesbecauseoflimitedornon‐existingevidence[357].Therefore,itisimportanttotesttheeffectofanapplicationbeforebringingittothemarked.ThePDPcanbeusedasastructurefordesigningeHealth.Thiswillenhancethechanceofasuccessfulimplementationandthattheproductdoeswhatonewantsittodo.

The gamification and eHealth in this dissertation focused mainly on thepatientasanend‐userinayouthmentalhealthcarecontext.However,ayouthmental healthcare context consists of more stakeholders than just patients,liketherapistsandmanagers.StakeholdersareimportantfordevelopmentofeHealth(gamifiedornotgamified)whentheycanhelptoaligntheproducttothe c o n t e x t a n d user’s preferences, needs and competences, which inturn can enhance motivation of users to interact with the product andconsequently implementation potential. This is because stakeholders withdifferent expertise (e.g., in design, the health context, or in their own


158

preferences) have different points of view and can provide more completeinsights into what theproduct should consistsof and focuson. In addition,involvingstakeholderscanhelp tomanage theirexpectationswithregard tothe intervention through framing. For example, negative expectations oftherapists, important for implementation in therapeutic practice, about thebenefitsofeHealthcomparedtoface‐to‐facetherapy[199‐201].

DesignresearchersshouldmakeexplicitwhatagamifiedeHealthapplicationcan support both therapists and patients and how this is possible, andconvince themthateHealthandgamificationwillnotdistort the therapeuticalliance. One could do this by making demos with examples of often‐usedtherapeuticsituationsthatonecanrecognize.

Toconclude, it is important to integrate therapeuticaspectswhendesigningpersonalizedandgamifiedeHealthandsetting‐upstudiestotesttheeffectsofsuch eHealth applications. However, there is also a limitation regardingpractical feasibility concerning the amount of stakeholder involvement. It isessential that these types of applications are able to correspond withtherapeutic practice and a broad range of patient characteristics. EHealthdesigners have to know why and how patients and therapists personalizetheir therapy, inorder to identify theparts ineHealthdesign thatshouldbeopentopersonalizationinordertofacilitateengagementandimplementation,while still ensuringpositive therapeutic effects. A designer canwork on theabove aspects by collaborating with all stakeholders in specific contexts,whichcanbestructuredbyusingthePersonalizedDesignProcessmodel.ItisexpectedthattheabovewillenhanceimplementationandimprovetheeffectofgamifiedeHealth.


159

160


161

REFERENCES[1] J.Ormeletal.,"MentalhealthinDutchadolescents:aTRAILSreporton

prevalence,severity,ageofonset,continuityandco‐morbidityofDSMdisorders,"PsychologicalMedicine,vol.45,no.2,pp.345‐360,2015.

[2] R.C.Kessleretal.,"Lifetimeprevalenceandage‐of‐onsetdistributionsofmentaldisordersintheWorldHealthOrganization'sWorldMentalHealthSurveyInitiative,"WorldPsychiatry,vol.6,no.3,pp.168‐176,2007.

[3] E.Hazen,S.Schlozman,andE.Beresin,"Adolescentpsychologicaldevelopment:areview,"(ineng),PediatrRev,vol.29,no.5,pp.161‐7;quiz168,May2008.

[4] L.Steinberg,"Cognitiveandaffectivedevelopmentinadolescence,"(ineng),TrendsCognSci,vol.9,no.2,pp.69‐74,Feb2005.

[5] D.A.SturmanandB.Moghaddam,"Theneurobiologyofadolescence:Changesinbrainarchitecture,functionaldynamics,andbehavioraltendencies,"Neuroscience&BiobehavioralReviews,vol.35,no.8,pp.1704‐1712,2011/08/01/2011.

[6] S.BurnettandS.‐J.Blakemore,"TheDevelopmentofAdolescentSocialCognition,"AnnalsoftheNewYorkAcademyofSciences,vol.1167,no.1,pp.51‐56,06/242009.

[7] R.W.Larson,G.Moneta,M.H.Richards,andS.Wilson,"Continuity,stability,andchangeindailyemotionalexperienceacrossadolescence,"Childdevelopment,vol.73,no.4,pp.1151‐1165,2002.

[8] R.C.Kessler,C.L.Foster,W.B.Saunders,andP.E.Stang,"Socialconsequencesofpsychiatricdisorders,I:Educationalattainment,"TheAmericanjournalofpsychiatry,vol.152,no.7,p.1026,1995.

[9] D.M.FergussonandL.J.Woodward,"Mentalhealth,educational,andsocialroleoutcomesofadolescentswithdepression,"ArchivesofGeneralPsychiatry,vol.59,no.3,pp.225‐231,2002.

[10] D.M.Fergusson,I.J.Horwood,andE.M.Ridder,"Showmethechildatseven:theconsequencesofconductproblemsinchildhoodforpsychosocialfunctioninginadulthood,"JournalofChildPsychologyandPsychiatry,vol.46,no.8,pp.837‐849,2005.

[11] S.L.Andersen,A.Tomada,E.S.Vincow,E.Valente,A.Polcari,andM.H.Teicher,"Preliminaryevidenceforsensitiveperiodsintheeffectofchildhoodsexualabuseonregionalbraindevelopment,"TheJournalofneuropsychiatryandclinicalneurosciences,vol.20,no.3,pp.292‐301,2008.

[12] L.M.Bylsma,B.H.Morris,andJ.Rottenberg,"Ameta‐analysisofemotionalreactivityinmajordepressivedisorder,"ClinicalPsychologyReview,vol.28,no.4,pp.676‐691,2008/04/01/2008.

[13] D.A.BrentandD.J.Kolko,"Psychotherapy:definitions,mechanismsofaction,andrelationshiptoetiologicalmodels,"(ineng),JAbnormChildPsychol,vol.26,no.1,pp.17‐25,Feb1998.

[14] D.CottrellandP.Boston,"Practitionerreview:Theeffectivenessofsystemicfamilytherapyforchildrenandadolescents,"JournalofChildPsychologyandPsychiatry,vol.43,no.5,pp.573‐586,2002.

References

162

[15] M.A.Reinecke,N.E.Ryan,andD.L.DuBois,"Cognitive‐behavioraltherapyofdepressionanddepressivesymptomsduringadolescence:areviewandmeta‐analysis,"JournaloftheAmericanAcademyofChildandAdolescentPsychiatry,vol.37,no.1,pp.26‐34,1998.

[16] S.Cartwright‐Hatton,C.Roberts,P.Chitsabesan,C.Fothergill,andR.Harrington,"Systematicreviewoftheefficacyofcognitivebehaviourtherapiesforchildhoodandadolescentanxietydisorders,"Britishjournalofclinicalpsychology,vol.43,no.4,pp.421‐436,2004.

[17] J.R.Weiszetal.,"Whatfivedecadesofresearchtellsusabouttheeffectsofyouthpsychologicaltherapy:Amultilevelmeta‐analysisandimplicationsforscienceandpractice,"(ineng),AmPsychol,vol.72,no.2,pp.79‐117,Feb‐Mar2017.

[18] S.T.Gaynor,P.S.Lawrence,andR.O.Nelson‐Gray,"MeasuringHomeworkComplianceinCognitive‐BehavioralTherapyforAdolescentDepressionReview,PreliminaryFindings,andImplicationsforTheoryandPractice,"Behaviormodification,vol.30,no.5,pp.647‐672,2006.

[19] B.T.Mausbach,R.Moore,S.Roesch,V.Cardenas,andT.L.Patterson,"Therelationshipbetweenhomeworkcomplianceandtherapyoutcomes:Anupdatedmeta‐analysis,"CognitiveTherapyandResearch,vol.34,no.5,pp.429‐438,2010.

[20] E.HansandW.Hiller,"Effectivenessofanddropoutfromoutpatientcognitivebehavioraltherapyforadultunipolardepression:Ameta‐analysisofnonrandomizedeffectivenessstudies,"JournalofConsultingandClinicalPsychology,vol.81,no.1,p.75,2013.

[21] H.Xiao,J.A.Hayes,L.G.Castonguay,A.A.McAleavey,andB.D.Locke,"Therapisteffectsandtheimpactsoftherapynonattendance,"(ineng),Psychotherapy(Chic),vol.54,no.1,pp.58‐65,Mar2017.

[22] V.Daleetal.,"Exploringtreatmentattendanceanditsrelationshiptooutcomeinarandomizedcontrolledtrialoftreatmentforalcoholproblems:secondaryanalysisoftheUKAlcoholTreatmentTrial(UKATT),"(ineng),AlcoholAlcohol,vol.46,no.5,pp.592‐9,Sep‐Oct2011.

[23] A.M.deHaan,A.E.Boon,J.T.deJong,M.Hoeve,andR.R.Vermeiren,"Ameta‐analyticreviewontreatmentdropoutinchildandadolescentoutpatientmentalhealthcare,"ClinicalPsychologyReview,vol.33,no.5,pp.698‐711,2013.

[24] G.Eysenbach,"Whatise‐health?,"JournalofMedicalInternetResearch,vol.3,no.2,p.e20,2001.

[25] C.Pagliarietal.,"WhatIseHealth(4):AScopingExercisetoMaptheField,"JournalofMedicalInternetResearchvol.7,no.1,p.e9,3/312005.

[26] H.Oh,C.Rizo,M.Enkin,andA.Jadad,"WhatiseHealth(3):asystematicreviewofpublisheddefinitions,"JournalofMedicalInternetResearchvol.7,no.1,p.e1,2005.

[27] G.Andersson,P.Cuijpers,P.Carlbring,H.Riper,andE.Hedman,"GuidedInternet‐basedvs.face‐to‐facecognitivebehaviortherapyforpsychiatricandsomaticdisorders:asystematicreviewandmeta‐analysis,"(ineng),WorldPsychiatry,vol.13,no.3,pp.288‐95,Oct2014.


163

[28] I.Titzler,K.Saruhanjan,M.Berking,H.Riper,andD.D.Ebert,"Barriersandfacilitatorsfortheimplementationofblendedpsychotherapyfordepression:Aqualitativepilotstudyoftherapists'perspective,"InternetInterventions,vol.12,pp.150‐164,2018/06/01/2018.

[29] (2018,December28).Internet;toegang,gebruikenfaciliteiten.[30] S.HelbigandL.Fehm,"ProblemswithhomeworkinCBT:Rareexceptionor

ratherfrequent?,"Behaviouralandcognitivepsychotherapy,vol.32,no.3,pp.291‐301,2004.

[31] N.Kazantzis,F.P.Deane,andK.R.Ronan,"HomeworkAssignmentsinCognitiveandBehavioralTherapy:AMeta‐Analysis,"ClinicalPsychology:ScienceandPractice,vol.7,no.2,pp.189‐202,2000.

[32] D.Lupton,"Thedigitallyengagedpatient:Self‐monitoringandself‐careinthedigitalhealthera,"SocialTheory&Health,vol.11,no.3,pp.256‐270,2013.

[33] J.Wentzel,R.vanderVaart,E.T.Bohlmeijer,andJ.E.W.C.vanGemert‐Pijnen,"MixingOnlineandFace‐to‐FaceTherapy:HowtoBenefitFromBlendedCareinMentalHealthCare,"JMIRMentalHealth,vol.3,no.1,p.e9,2016.

[34] R.vanderVaart,M.Witting,H.Riper,L.Kooistra,E.T.Bohlmeijer,andJ.E.W.C.vanGemert‐Pijnen,"Blendingonlinetherapyintoregularface‐to‐facetherapyfordepression:content,ratioandpreconditionsaccordingtopatientsandtherapistsusingaDelphistudy,"BMCPsychiatry,journalarticlevol.14,no.1,p.355,December142014.

[35] A.O.HorvathandB.D.Symonds,"Relationbetweenworkingallianceandoutcomeinpsychotherapy:Ameta‐analysis,"JournalofCounselingPsychology,vol.38,no.2,pp.139‐149,1991.

[36] S.R.Shirk,M.S.Karver,andR.Brown,"Theallianceinchildandadolescentpsychotherapy,"Psychotherapy,vol.48,no.1,p.17,2011.

[37] C.Sundström,M.Blankers,andZ.Khadjesari,"Computer‐BasedInterventionsforProblematicAlcoholUse:aReviewofSystematicReviews,"InternationalJournalofBehavioralMedicine,vol.24,no.5,pp.646‐658,2017/10/012017.

[38] N.Boumparis,E.Karyotaki,M.P.Schaub,P.Cuijpers,andH.Riper,"Internetinterventionsforadultillicitsubstanceusers:ameta‐analysis,"Addiction(Abingdon,England),vol.112,no.9,pp.1521‐1532,2017.

[39] G.Andrewsetal.,"Computertherapyfortheanxietyanddepressiondisordersiseffective,acceptableandpracticalhealthcare:Anupdatedmeta‐analysis,"JournalofAnxietyDisorders,vol.55,pp.70‐78,2018.

[40] X.Yeetal.,"Effectivenessofinternet‐basedinterventionsforchildren,youth,andyoungadultswithanxietyand/ordepression:asystematicreviewandmeta‐analysis,"BMChealthservicesresearch,vol.14,no.1,p.313,2014.

[41] L.Farreretal.,"Technology‐BasedInterventionsforMentalHealthinTertiaryStudents:SystematicReview,"JournalofMedicalInternetResearch,vol.15,no.5,p.e101,2013.

[42] G.AnderssonandP.Cuijpers,"Internet‐basedandothercomputerizedpsychologicaltreatmentsforadultdepression:ameta‐analysis,"CognBehavTher,vol.38,2009.

References

164

[43] V.Spek,P.Cuijpers,I.Nyklícek,H.Riper,J.Keyzer,andV.Pop,"Internet‐basedcognitivebehaviourtherapyforsymptomsofdepressionandanxiety:ameta‐analysis,"PsychologicalMedicine,vol.37,no.3,pp.319‐328,2007.

[44] R.M.F.Kenteretal.,"CostsandeffectsofInternetcognitivebehavioraltreatmentblendedwithface‐to‐facetreatment:Resultsfromanaturalisticstudy,"InternetInterventions,vol.2,no.1,pp.77‐83,2015/03/01/2015.

[45] H.Christensen,K.M.Griffiths,andL.Farrer,"Adherenceininternetinterventionsforanxietyanddepression,"(ineng),JMedInternetRes,vol.11,no.2,p.e13,Apr242009.

[46] T.L.Dunn,L.M.Casey,J.Sheffield,P.Newcombe,andA.B.Chang,"Dropoutfromcomputer‐basedinterventionsforchildrenandadolescentswithchronichealthconditions,"Journalofhealthpsychology,vol.17,no.3,pp.429‐442,2012.

[47] W.vanBallegooijenetal.,"AdherencetoInternet‐basedandface‐to‐facecognitivebehaviouraltherapyfordepression:ameta‐analysis,"PLoSOne,vol.9,no.7,p.e100674,2014.

[48] M.Wilhelmsenetal.,"Motivationtopersistwithinternet‐basedcognitivebehaviouraltreatmentusingblendedcare:aqualitativestudy,"BMCPsychiatry,vol.13,2013.

[49] S.K.Orlowskietal.,"Participatoryresearchasonepieceofthepuzzle:asystematicreviewofconsumerinvolvementindesignoftechnology‐basedyouthmentalhealthandwell‐beinginterventions,"JMIRhumanfactors,vol.2,no.2,2015.

[50] T.Mattelmäki,&SleeswijkVisser,F.,"LostinCo‐X‐InterpretationsofCo‐designandCo‐creation,"inProceedingsofthe4thWorldConferenceonDesignResearch(IASDR2011):DiversityandUnity,Delft,TheNetherlands,2011,vol.31.

[51] R.M.RyanandE.L.Deci,"Self‐determinationtheoryandthefacilitationofintrinsicmotivation,socialdevelopment,andwell‐being,"Americanpsychologist,vol.55,no.1,p.68,2000.

[52] A.DeSmet,D.Thompson,T.Baranowski,A.Palmeira,M.Verloigne,andI.DeBourdeaudhuij,"IsParticipatoryDesignAssociatedwiththeEffectivenessofSeriousDigitalGamesforHealthyLifestylePromotion?AMeta‐Analysis,"(inEnglish),JournalofMedicalInternetResearch,OriginalPapervol.18,no.4,p.e94,2016.

[53] E.B.‐N.SandersandP.J.Stappers,"Co‐creationandthenewlandscapesofdesign,"CoDesign:InternationalJournalofCoCreationinDesignandtheArts,vol.4,no.1,pp.5‐18,2008.

[54] E.Frias‐Martinez,S.Y.Chen,andX.Liu,"Evaluationofapersonalizeddigitallibrarybasedoncognitivestyles:Adaptivityvs.adaptability,"InternationalJournalofInformationManagement,vol.29,no.1,pp.48‐56,2//2009.

[55] R.K.McHughandD.H.Barlow,"Thedisseminationandimplementationofevidence‐basedpsychologicaltreatments:areviewofcurrentefforts,"AmericanPsychologist,vol.65,no.2,pp.73‐84,2010.


165

[56] A.Hogue,S.Dauber,J.Samuolis,andH.A.Liddle,"TreatmentTechniquesandOutcomesinMultidimensionalFamilyTherapyforAdolescentBehaviorProblems,"JournalofFamilyPsychologyvol.20,no.4,pp.535‐543,2006.

[57] S.W.Stirmanetal.,"ModificationstoCognitiveTherapybyCommunityMentalHealthProviders:ImplicationsforEffectivenessandSustainability,"PsychiatricServices,vol.64,no.10,pp.1056‐1059,2013.

[58] K.M.CarrollandB.J.Rounsaville,"Avisionofthenextgenerationofbehavioraltherapiesresearchintheaddictions,"Addiction,vol.102,no.6,pp.850–869,2007.

[59] C.B.TaylorandV.Y.Chang,"Issuesinthedisseminationofcognitive–behaviortherapy,"NordicJournalofPsychiatry,vol.62,no.47,pp.37‐44,2008.

[60] A.O.Horvath,A.C.DelRe,C.Flückiger,andD.Symonds,"Allianceinindividualpsychotherapy,"Psychotherapy,vol.48,no.1,pp.9‐16,2011.

[61] S.R.Shirk,G.Gudmundsen,H.C.Kaplinski,andD.L.McMakin,"Allianceandoutcomeincognitive‐behavioraltherapyforadolescentdepression,"JournalofClinicalChild&AdolescentPsychology,vol.37,no.3,pp.631‐639,2008.

[62] M.Blankers,T.Donker,andH.Riper,"Wetenschappelijkeevidentieenuitdagingenvoordepraktijk.E‐MENTALHEALTHINNEDERLAND,"Depsycholoog,vol.2013,pp.12‐23,2013.

[63] V.T.Visch,N.J.H.Vegt,H.Anderiesen,andK.vanderKooij,"PersuasiveGameDesign:Amodelanditsdefinitions,"presentedattheCHIWorkshopDesigningGamification:CreatingGamefulandPlayfulExperiences,Paris,France,2013.

[64] T.Holocher‐Ertl,B.Kieslinger,andC.M.Fabian,"Designingfortheusersorwiththeusers?Aparticipatorydesignapproachforscienceteachinginschools,"inProceedingsofthe2012eChallengesAnnualConference,Lisbon,Portugal,2012.

[65] C.Spinuzzi,"TheMethodologyofParticipatoryDesign,"TechnicalCommunication,vol.52,no.2,pp.163‐174,//2005.

[66] M.Steen,M.Manschot,andN.DeKoning,"BenefitsofCo‐designinServiceDesignProjects,"InternationalJournalofDesign,Benefits,Co‐Design,ServiceDesign,Casesvol.5,no.2,pp.53‐60,2011.

[67] C.K.PrahaladandV.Ramaswamy,"Co‐creationexperiences:Thenextpracticeinvaluecreation,"JournalofInteractiveMarketing,vol.18,no.3,pp.5‐14,2004.

[68] C.‐I.Teng,"Customization,immersionsatisfaction,andonlinegamerloyalty,"ComputersinHumanBehavior,vol.26,no.6,pp.1547‐1554,11//2010.

[69] C.P.Lin,H.N.Huang,S.W.Joe,andH.C.Ma,"Learningthedeterminantsofsatisfactionandusageintentionofinstantmessaging,"Cyberpsychology&Behavior,vol.11,no.3,pp.262‐267,2008.

[70] A.Dijkstra,"Thepersuasiveeffectsofpersonalizationthrough:Namementioninginasmokingcessationmessage,"UserModelingandUser‐AdaptedInteraction,vol.24,no.5,pp.393‐411,Dec2014.

[71] D.Johnson,S.Deterding,K.‐A.Kuhn,A.Staneva,S.Stoyanov,andL.Hides,"Gamificationforhealthandwellbeing:Asystematicreviewoftheliterature,"InternetInterventions,vol.6,no.SupplementC,pp.89‐106,2016.

[72] H.Korhonen,M.Montola,andJ.Arrasvuori,"Understandingplayfuluserexperiencethroughdigitalgames,"inInternationalConferenceonDesigning

References

166

PleasurableProductsandInterfaces,DPPI09,France,2009:CompiegneUniversityofTechnology.

[73] L.SherryJohn,"FlowandMediaEnjoyment,"CommunicationTheory,vol.14,no.4,pp.328‐347,2006.

[74] A.K.Przybylski,C.S.Rigby,andR.M.Ryan,"Amotivationalmodelofvideogameengagement,"ReviewofGeneralPsychology,vol.14,no.2,p.154,2010.

[75] A.UysalandI.G.Yildirim,"Self‐DeterminationTheoryinDigitalGames,"inGamerPsychologyandBehavior,B.Bostan,Ed.Cham:SpringerInternationalPublishing,2016,pp.123‐135.

[76] S.Bakkes,C.T.Tan,andY.Pisan,"Personalisedgaming:amotivationandoverviewofliterature,"inProceedingsofthe8thAustralasianConferenceonInteractiveEntertainment:PlayingtheSystem,2012,p.4:ACM.

[77] L.Sardi,A.Idri,andJ.L.Fernández‐Alemán,"ASystematicReviewofGamificationine‐Health,"JournalofBiomedicalInformatics,2017.

[78] T.M.Flemingetal.,"Seriousgamesandgamificationformentalhealth:currentstatusandpromisingdirections,"Frontiersinpsychiatry,vol.7,p.215,2017.

[79] T.A.DennisandL.O'Toole,"MentalHealthontheGo:EffectsofaGamifiedAttentionBiasModificationMobileApplicationinTraitAnxiousAdults,"(ineng),ClinPsycholSci,vol.2,no.5,pp.576‐590,Sep12014.

[80] T.Flemingetal.,"Seriousgamesforthetreatmentorpreventionofdepression:asystematicreview,"RevistadePsicopatologíayPsicologíaClínica,vol.19,no.3,pp.227‐242,2014.

[81] L.Gamberini,G.Barresi,A.Majer,andF.Scarpetta,"Agameadaykeepsthedoctoraway:ashortreviewofcomputergamesinmentalhealthcare,"JournalofCyberTherapy&Rehabilitation,vol.1,no.2,pp.127‐149,2008.

[82] J.J.Santamariaetal.,"Seriousgamesasadditionalpsychologicalsupport:areviewoftheliterature,"JournalofCyberTherapyandRehabilitation,vol.4,no.4,pp.469‐477,2011.

[83] A.DeSmetetal.,"Ameta‐analysisofseriousdigitalgamesforhealthylifestylepromotion,"Preventivemedicinevol.69,pp.95‐107,2014.

[84] P.M.Kato,"Videogamesinhealthcare:Closingthegap,"ReviewofGeneralPsychology,vol.14,no.2,p.113,2010.

[85] H.L.Horne‐Moyer,B.H.Moyer,D.C.Messer,andE.S.Messer,"TheUseofElectronicGamesinTherapy:aReviewwithClinicalImplications,"CurrentPsychiatryReports,journalarticlevol.16,no.12,p.520,2014.

[86] J.Li,Y.‐L.Theng,andS.Foo,"Game‐BasedDigitalInterventionsforDepressionTherapy:ASystematicReviewandMeta‐Analysis,"Cyberpsychology,Behavior,andSocialNetworking,vol.17,no.8,pp.519‐527,2014/08/012014.

[87] V.Brezinka,"Computergamessupportingcognitivebehaviourtherapyinchildren,"Clinicalchildpsychologyandpsychiatry,vol.19,no.1,pp.100‐110,2014.

[88] H.M.Lau,J.H.Smit,T.M.Fleming,andH.Riper,"SeriousGamesforMentalHealth:AreTheyAccessible,Feasible,andEffective?ASystematicReviewandMeta‐analysis,"FrontiersinPsychiatry,vol.7,p.209,2016.


167

[89] S.M.Kelders,M.Sommers‐Spijkerman,andJ.Goldberg,"InvestigatingtheDirectImpactofaGamifiedVersusNongamifiedWell‐BeingIntervention:AnExploratoryExperiment,"(inEnglish),JMedInternetRes,OriginalPapervol.20,no.7,p.e247,2018.

[90] F.M.Brooks,K.L.Chester,N.C.Smeeton,andN.H.Spencer,"Videogaminginadolescence:factorsassociatedwithleisuretimeuse,"JournalofYouthStudies,vol.19,no.1,pp.36‐54,2016.

[91] A.DeaconandK.O’Farrell,"Theuseofseriousgamesandgamifieddesigntoimprovehealthoutcomesinadolescentswithchronicdisease:Areviewofrecentliterature,"2016.

[92] M.Brown,N.O'Neill,H.vanWoerden,P.Eslambolchilar,M.Jones,andA.John,"GamificationandAdherencetoWeb‐BasedMentalHealthInterventions:ASystematicReview,"JMIRMentalHealth,vol.3,no.3,p.e39,08/242016.

[93] G.MireyaSilva,A.Raposo,andM.Suplino,"ExploringCollaborationPatternsinaMultitouchGametoEncourageSocialInteractionandCollaborationAmongUserswithAutismSpectrumDisorder,"(inEnglish),ComputerSupportedCooperativeWork(CSCW),pp.1‐27,2014/12/102014.

[94] S.RaghupathyandA.L.Forth,"TheHAWK2program:acomputer‐baseddrugpreventioninterventionforNativeAmericanyouth,"(ineng),TheAmericanjournalofdrugandalcoholabuse,vol.38,no.5,pp.461‐7,Sep2012.

[95] C.Holmgård,G.N.Yannakakis,K.I.Karstoft,andH.S.Andersen,"StressDetectionforPTSDviatheStartleMartGame,"InproceedingsoftheHumaineAssociationConferenceonAffectiveComputingandIntelligentInteraction(ACII),pp.523‐528,2013.

[96] M.Glasemann,A.M.Kanstrup,andT.Ryberg,"Designandexplorationofamobilegamescenarioinadiabeticyouthcamp,"IADISInternationalconferenceMobileLearning,pp.132‐140,2010.

[97] J.HollingdaleandT.Greitemeyer,"Thechangingfaceofaggression:Theeffectofpersonalizedavatarsinaviolentvideogameonlevelsofaggressivebehavior,"JournalofAppliedSocialPsychology,vol.43,no.9,pp.1862‐1868,Sep2013.

[98] E.B.‐N.Sanders,"Fromuser‐centeredtoparticipatorydesignapproaches,"Designandthesocialsciences:Makingconnections,pp.1‐8,2002.

[99] R.ProfittandB.Lange,"Usercentereddesignanddevelopmentofagameforexerciseinolderadults,"InternationalJournalofTechnology,Knowledge&Society,vol.8,no.5,pp.95‐112,2012.

[100] R.Mugge,J.P.L.Schoormans,andH.N.J.Schifferstein,"Incorporatingconsumersinthedesignoftheirownproducts.Thedimensionsofproductpersonalisation,"CoDesign,vol.5,no.2,pp.79‐97,2009/06/012009.

[101] S.S.SundarandS.S.Marathe,"PersonalizationversusCustomization:TheImportanceofAgency,Privacy,andPowerUsage,"HumanCommunicationResearch,vol.36,no.3,pp.298‐322,2010.

[102] A.Dijkstra,"ThePsychologyofTailoring‐IngredientsinComputer‐TailoredPersuasion,"SocialandPersonalityPsychologyCompass,vol.2,no.2,pp.765‐784,2008.

References

168

[103] S.M.Noar,N.G.Harrington,andR.S.Aldrich,"TheRoleofMessageTailoringintheDevelopmentofPersuasiveHealthCommunicationMessages,"AnnalsoftheInternationalCommunicationAssociation,vol.33,no.1,pp.73‐133,2009/01/012009.

[104] B.S.Fjeldsoe,A.L.Marshall,andY.D.Miller,"BehaviorChangeInterventionsDeliveredbyMobileTelephoneShort‐MessageService,"AmericanJournalofPreventiveMedicine,vol.36,no.2,pp.165‐173,2//2009.

[105] M.L.A.Lustria,J.Cortese,S.M.Noar,andR.L.Glueckauf,"Computer‐tailoredhealthinterventionsdeliveredovertheWeb:reviewandanalysisofkeycomponents,"Patienteducationandcounseling,vol.74,no.2,pp.156‐173,2009.

[106] S.Easterbrook,"ResolvingRequirementsConflictswithComputer‐SupportedNegotiation,"inInRequirementsEngineering:SocialandTechnicalIssues,M.Jirotka,andGoguen,J.,Ed.:USA:AcademicPress,1994,pp.41‐65.

[107] A.vanBoeijen,J.Daalhuizen,J.Zijlstra,andR.vanderSchoor,DelftDesignGuide:DesignMethods.Amsterdam:BISPublishers,2013.

[108] D.SzebekoandL.Tan,"Co‐designingforSociety,"AustralasianMedicalJournal,vol.3,no.9,pp.580‐590,2010.

[109] V.Baert,E.Gorus,T.Mets,C.Geerts,andI.Bautmans,"Motivatorsandbarriersforphysicalactivityintheoldestold:asystematicreview,"(ineng),Ageingresearchreviews,vol.10,no.4,pp.464‐74,Sep2011.

[110] P.Krebs,J.O.Prochaska,andJ.S.Rossi,"Ameta‐analysisofcomputer‐tailoredinterventionsforhealthbehaviorchange,"(ineng),Preventivemedicine,vol.51,no.3‐4,pp.214‐21,Sep‐Oct2010.

[111] D.B.Portnoy,L.A.Scott‐Sheldon,B.T.Johnson,andM.P.Carey,"Computer‐deliveredinterventionsforhealthpromotionandbehavioralriskreduction:ameta‐analysisof75randomizedcontrolledtrials,1988‐2007,"(ineng),Preventivemedicine,vol.47,no.1,pp.3‐16,Jul2008.

[112] D.A.Lieberman,C.H.Bates,andS.Jiyeon,"YoungChildren'sLearningWithDigitalMedia,"ComputersintheSchools,vol.26,no.4,pp.271‐283,2009.

[113] S.N.Merry,K.Stasiak,M.Shepherd,C.Frampton,T.Fleming,andM.F.G.Lucassen,TheeffectivenessofSPARX,acomputerisedselfhelpinterventionforadolescentsseekinghelpfordepression:randomisedcontrollednon‐inferioritytrial.2012.

[114] R.Moss‐Morris,P.McCrone,L.Yardley,K.vanKessel,G.Wills,andL.Dennison,"ApilotrandomisedcontrolledtrialofanInternet‐basedcognitivebehaviouraltherapyself‐managementprogramme(MSInvigor8)formultiplesclerosisfatigue,"Behaviourresearchandtherapy,vol.50,no.6,pp.415‐421,2012.

[115] S.Crovato,G.Mascarello,A.Pinto,andL.Ravarotto,"Testingaseriousgameforcommunicatingfoodrisktoyoungconsumers:“Amysteriouspoisoning","Proceedingsofthe5thInternationalConferenceonEducationandNewLearningTechnologies:Edulearn13,pp.283‐289,2013.

[116] K.C.Organista,N.J.Alvarado,A.Balblutin‐Burnham,P.Worby,andS.R.Martinez,"AnExploratoryStudyofHIVPreventionwithMexican/Latino


169

MigrantDayLaborers,"JournalofHIV/AIDS&SocialServices,vol.5,no.2,pp.89‐114,2006.

[117] A.L.McGaffey,D.J.Abatemarco,I.K.Jewell,S.K.Fidler,andK.Hughes,"FitwitsMD(TM):AnOffice‐BasedToolandGamesforConversationsaboutObesitywith9‐to12‐Year‐OldChildren,"JournaloftheAmericanBoardofFamilyMedicine,vol.24,no.6,pp.768‐771,Nov‐Dec2011.

[118] P.SinghandI.Niranjan,"Usinggroupworkinthedesigningandplayingofboardgamesinoccupationalhealthandsafety,"Groupwork,vol.22,no.3,pp.65‐82,2012.

[119] J.M.Cahill,"HealthWorks:InteractiveAIDSeducationvideogames,"ComputersinHumanServices,vol.11,no.1‐2,pp.159‐176,1994.

[120] S.Costaetal.,"Anapproachtopromotesocialandcommunicationbehaviorsinchildrenwithautismspectrumdisorders:Robotbasedintervention,"Proceedingsofthe20thIEEEinternationalsymposiumonrobotandhumaninteractivecommunication:RO‐MAN,pp.101‐1062011.

[121] S.Cobbetal.,"AppliedvirtualenvironmentstosupportlearningofsocialinteractionskillsinuserswithAsperger'sSyndrome,"DigitalCreativity,vol.13,no.1,pp.11‐22,2002.

[122] B.Schmitz,R.Klemke,J.Walhout,andM.Specht,"Attuningamobilesimulationgameforschoolchildrenusingadesign‐basedresearchapproach,"Computers&Education,vol.81,pp.35‐48,Feb2015.

[123] A.S.AlklindTaylor,P.Backlund,H.Engstrom,M.Johannesson,andM.Lebram,"ThebirthofElinor:acollaborativedevelopmentofagamebasedsystemforstrokerehabilitation,"inProceedingsofthe2ndInternationalConferenceinVisualisation,2009,pp.52‐60.

[124] M.Gotsisetal.,"Skyfarer:DesignCaseStudyofaMixedRealityRehabilitationVideoGame,"presentedattheInDesign,UserExperience,andUsability.UserExperienceDesignforDiverseInteractionPlatformsandEnvironments,thirdInternationalConference,DUXU:Design,UserExperience,andUsability,2014.

[125] N.v.H.Nijhof,J.vanRijn,H.vanGemert‐Pijnen,J.E.W.C.,"Thebehavioraloutcomesofatechnology‐supportedleisureactivityinpeoplewithdementia,"TechnologyandDisability,vol.24,no.4,pp.263‐273,2013

[126] A.McGaffey,K.Hughes,S.Fidler,F.D'Amico,andM.Stalter,"CanElvisPretzleyandtheFitwitsimproveknowledgeofobesity,nutrition,exercise,andportionsinfifthgraders?,"InternationalJournalofObesity,vol.34,no.7,pp.1134‐1142,Jul2010.

[127] R.Bailey,K.Wise,andP.Bolls,"Howavatarcustomizabilityaffectschildren'sarousalandsubjectivepresenceduringjunkfood‐sponsoredonlinevideogames,"CyberPsychology&Behavior,vol.12,no.3,pp.277‐283,Jun2009.

[128] A.Ben‐Sasson,L.Lamash,andE.Gal,"Toenforceornottoenforce?Theuseofcollaborativeinterfacestopromotesocialskillsinchildrenwithhighfunctioningautismspectrumdisorder,"Autism,vol.17,no.5,pp.608‐622,Sep2013.

References

170

[129] C.S.González,C.Collazos,J.L.González,P.Toledo,andF.Blanco,"Theimportanceofhumanfactorstoenhancetheuserexperienceinvideogames,"InternationalSymposiumonComputersinEducation,pp.1‐4,2012.

[130] G.D.Cohen,"TwonewintergenerationalinterventionsforAlzheimer'sdiseasepatientsandfamilies,"AmericanJournalofAlzheimer'sDiseaseandOtherDementias,vol.15,no.3,pp.137‐142,May1,20002000.

[131] A.M.HussaanandK.Sehaba,"AdaptiveSeriousGameforRehabilitationofPersonswithCognitiveDisabilities,"Proceedingsofthe13thInternationalConferenceonAdvancedLearningTechnologies,pp.65‐69,2013.

[132] Y.Kim,"Canyouravatarimproveyourhealth?theimpactofavatarcustomization(Doctoraldissertation),"PennsylvaniaStateUniversity,2010.

[133] H.H.LundandC.B.Nielsen,"Modularityformodulatingexercisesandlevels‐observationsfromcardiac,stroke,andCOLDpatientstherapy,"presentedattheProceedingsofthe8thInternationalConferenceonUbiquitousRobotsandAmbientIntelligence,2011.

[134] M.ChilcottandA.Smith,"AgeingWellandLearningthroughOnlineImmersiveParticipationUsingaMulti‐userWeb3DEnvironment,"ProceedingsoftheThirdInternationalConferenceonGamesandVirtualWorldsforSeriousApplications,pp.70‐75,2011.

[135] S.Albiol‐Pérezetal.,"ANovelVirtualMotorRehabilitationSystemforGuillain‐BarreSyndrome.TwoSingleCaseStudies,"(inEng),MethodsofInformationinMedicine,vol.54,no.2,pp.127‐134,Jan222015.

[136] C.Boulangeretal.,"Strokerehabilitationwithasensingsurface,"inProceedingsoftheSIGCHIConferenceonHumanFactorsinComputingSystems,2013,pp.1243‐1246.

[137] M.S.Cameirao,S.B.Badia,E.D.Oller,andP.F.Verschure,"NeurorehabilitationusingthevirtualrealitybasedRehabilitationGamingSystem:methodology,design,psychometrics,usabilityandvalidation,"(ineng),JournalofNeuroengineeringandRehabilitation,vol.7,p.48,2010.

[138] S.S.Kuys,K.Hall,M.Peasey,M.Wood,R.Cobb,andS.C.Bell,"Gamingconsoleexerciseandcycleortreadmillexerciseprovidesimilarcardiovasculardemandinadultswithcysticfibrosis:arandomisedcross‐overtrial,"(ineng),JournalofPhysiotherapy,vol.57,no.1,pp.35‐40,2011.

[139] J.C.Metzgeretal.,"Assessment‐drivenselectionandadaptationofexercisedifficultyinrobot‐assistedtherapy:apilotstudywithahandrehabilitationrobot,"JournalofNeuroengineeringandRehabilitation,vol.11,Nov2014,Art.no.154.

[140] C.R.Lylesetal.,"QualitativeEvaluationofaMobilePhoneandWeb‐BasedCollaborativeCareInterventionforPatientswithType2Diabetes,"DiabetesTechnology&Therapeutics,vol.13,no.5,pp.563‐569,May2011.

[141] J.Wainer,B.Robins,F.Amirabdollahian,andK.Dautenhahn,"UsingthehumanoidrobotKASPARtoautonomouslyplaytriadicgamesandfacilitatecollaborativeplayamongchildrenwithautism,"IEEETransactionsonAutonomousMentalDevelopment,vol.6,no.3,pp.183‐199,2014.


171

[142] H.K.Yuen,H.L.Breland,L.K.Vogtle,K.Holthaus,D.L.Kamen,andD.Sword,"Theprocessassociatedwithmotivationofahome‐basedWiiFitexerciseprogramamongsedentaryAfricanAmericanwomenwithsystemiclupuserythematosus,"DisabilityandHealthJournal,vol.6,no.1,pp.63‐68,Jan2013.

[143] M.Gonzalez‐Franco,S.Gilroy,andJ.O.Moore,"Empoweringpatientstoperformphysicaltherapyathome,"Proceedingsofthe36thAnnualInternationalConferenceoftheIEEEEngineeringinMedicineandBiologySocietypp.6308‐6311,2014.

[144] I.Sparrer,T.A.DuongDinh,J.Ilgner,andM.Westhofen,"VestibularrehabilitationusingtheNintendo(R)WiiBalanceBoard‐‐auser‐friendlyalternativeforcentralnervouscompensation,"(ineng),ActaOtolaryngol,vol.133,no.3,pp.239‐245,Mar2013.

[145] L.C.Abromsetal.,"Text2Quit:Resultsfromapilottestofapersonalized,interactivemobilehealthsmokingcessationprogram,"JournalofHealthCommunication,vol.17,no.Suppl1,pp.44‐53,May2012.

[146] J.ChoiandA.Medalia,"Intrinsicmotivationandlearninginaschizophreniaspectrumsample,"SchizophreniaResearch,vol.118,no.1‐3,pp.12‐19,May2010.

[147] A.Gouaïch,N.Hocine,L.VanDokkum,andD.Mottet,"Digital‐pheromonebaseddifficultyadaptationinpost‐stroketherapeuticgames,"inProceedingsofthe2ndACMSIGHITInternationalHealthInformaticsSymposium,2012,pp.5‐12.

[148] C.Peretz,A.D.Korczyn,E.Shatil,V.Aharonson,S.Birnboim,andN.Giladi,"Computer‐based,personalizedcognitivetrainingversusclassicalcomputergames:Arandomizeddouble‐blindprospectivetrialofcognitivestimulation,"Neuroepidemiology,vol.36,no.2,pp.91‐99,Apr2011.

[149] M.PlowandM.Finlayson,"AqualitativestudyexploringtheusabilityofNintendoWiiFitamongpersonswithmultiplesclerosis,"OccupationalTherapyInternational,vol.21,no.1,pp.21‐32,Mar2014.

[150] M.Frutos‐Pascual,B.G.Zapirain,andA.M.Zorrilla,"AdaptiveTele‐TherapiesBasedonSeriousGamesforHealthforPeoplewithTime‐ManagementandOrganisationalProblems:PreliminaryResults,"InternationalJournalofEnvironmentalResearchandPublicHealth,vol.11,no.1,pp.749‐772,Jan2014.

[151] W.Lee,Y.M.Chae,S.Kim,S.H.Ho,andI.Choi,"Evaluationofamobilephone‐baseddietgameforweightcontrol,"JournalofTelemedicineandTelecare,vol.16,no.5,pp.270‐275,2010.

[152] M.E.Hoque,J.K.Lane,R.ElKaliouby,M.Goodwin,andR.W.Picard,"Exploringspeechtherapygameswithchildrenontheautismspectrum,"inInINTERSPEECH,10thAnnualConferenceoftheInternationalSpeechCommunicationAssociation,2009,pp.1455‐1458.

[153] J.R.OctaviaandK.Coninx,"AdaptivePersonalizedTrainingGamesforIndividualandCollaborativeRehabilitationofPeoplewithMultipleSclerosis,"BiomedResearchInternational,2014,Art.no.345728.

[154] K.Ustinova,J.Perkins,W.Leonard,andC.Hausbeck,"Virtualrealitygame‐basedtherapyfortreatmentofposturalandco‐ordinationabnormalities

References

172

secondarytoTBI:Apilotstudy,"BrainInjury,vol.28,no.4,pp.486‐495,Apr2014.

[155] M.A.Padilla‐Castañedaetal.,"Avirtualrealitysystemforrobotic‐assistedorthopedicrehabilitationofforearmandelbowfractures,"inIEEE/RSJInternationalConferenceonIntelligentRobotsandSystems,2013,vol.11,pp.1506‐1511.

[156] B.BossavitandA.Pina,"DesigningEducationalTools,BasedonBodyInteraction,forChildrenwithSpecialNeedsWhoPresentDifferentMotorSkills,"inProceedingsofthe2014InternationalConferenceonInteractiveTechnologiesandGames2014,pp.63‐70.

[157] J.H.Shin,H.Ryu,andS.H.Jang,"Atask‐specificinteractivegame‐basedvirtualrealityrehabilitationsystemforpatientswithstroke:ausabilitytestandtwoclinicalexperiments,"JournalofNeuroengineeringandRehabilitation,vol.11,Mar2014,Art.no.32.

[158] M.Shepherd,T.Fleming,M.Lucassen,K.Stasiak,I.Lambie,andS.N.Merry,"Thedesignandrelevanceofacomputerizedgamifieddepressiontherapyprogramforindigenousmaoriadolescents,"(ineng),JMIRSeriousGames,vol.3,no.1,p.e1,2015.

[159] R.Cornejo,D.Hernandez,M.Tentori,andJ.Favela,"Anambientcasualgametopromotesocializationandactiveageing,"inInCollaborationandTechnology,20thInternationalConference,2014,pp.75‐88.

[160] F.Lu,K.Turner,andB.Murphy,"Reducingadolescentobesitywithamobilefitnessapplication:studyresultsofyouthage15to17,"IEEEHealthcom,15thInternationalConferenceonE‐Health,Networking,Application&Service,pp.554‐558,2013.

[161] H.M.Edwards,S.McDonald,T.Zhao,andL.Humphries,"Designrequirementsforpersuasivetechnologiestomotivatephysicalactivityinadolescents:Afieldstudy,"Behaviour&InformationTechnology,vol.33,no.9,pp.968‐986,Sep2014.

[162] L.Bartoli,F.Garzotto,M.Gelsomini,L.Oliveto,andM.Valoriani,"DesigningandevaluatingtouchlessplayfulinteractionforASDchildren,"inProceedingsofthe2014conferenceonInteractionDesignandChildren,2014,pp.17‐26.

[163] D.CoyleandG.Doherty,"ClinicalEvaluationsandCollaborativeDesign:Developingnewtechnologiesformentalhealthcareinterventions,"inProceedingsoftheSIGCHIConferenceonHumanFactorsinComputingSystems2009,pp.2051‐2060.

[164] J.D.Johnston,A.P.Massey,andR.L.Marker‐Hoffman,"Usinganalternaterealitygametoincreasephysicalactivityanddecreaseobesityriskofcollegestudents,"JournalofDiabetesScienceandTechnology,vol.6,no.4,pp.828‐838,2012.

[165] M.‐G.Kimetal.,"Designingrobot‐assistedPivotalResponseTrainingingameactivityforchildrenwithautism,"inProceedingsofIEEEInternationalConferenceonSystems,Man,andCybernetics,2014,pp.1101‐1106.


173

[166] E.I.Barakova,P.Bajracharya,M.Willemsen,T.Lourens,andB.Huskens,"Long‐termLEGOtherapywithhumanoidrobotforchildrenwithASD,"ExpertSystems,vol.32,no.6,pp.698‐709,2015.

[167] H.A.Hernandezetal.,"Designofanexergamingstationforchildrenwithcerebralpalsy,"inProceedingsoftheSIGCHIConferenceonHumanFactorsinComputingSystems,2012,pp.2619‐2628.

[168] N.Hocine,A.Gouaïch,S.A.Cerri,D.Mottet,J.Froger,andI.Laffont,"Adaptationinseriousgamesforupper‐limbrehabilitation:anapproachtoimprovetrainingoutcomes,"UserModelingandUser‐AdaptedInteraction,vol.25,no.1,pp.65‐98,2015.

[169] R.deOliveira,M.Cherubini,N.Oliver,andAcm,"MoviPill:ImprovingMedicationComplianceforEldersUsingaMobilePersuasiveSocialGame,"inProceedingsofthe12thACMinternationalconferenceonUbiquitouscomputing,2010,pp.251‐260.

[170] E.Tsekleves,I.T.Paraskevopoulos,A.Warland,andC.Kilbride,"DevelopmentandpreliminaryevaluationofanovellowcostVR‐basedupperlimbstrokerehabilitationplatformusingWiitechnology,"(inEng),DisabilityandRehabilitation:AssistiveTechnology,vol.11,no.5,pp.413‐422,Nov132014.

[171] A.A.Rizzoetal.,"VirtualIraq:InitialcasereportsfromaVRexposuretherapyapplicationforcombat‐relatedposttraumaticstressdisorder,"inVirtualRehabilitation,2007,pp.124‐130.

[172] P.HekkertandH.N.Schifferstein,"Introducingproductexperience,"Productexperience,pp.1‐8,2008.

[173] P.DesmetandP.Hekkert,"Frameworkofproductexperience,"InternationalJournalofDesign,vol.1,no.1,pp.57‐66,2007.

[174] H.Oinas‐Kukkonen,"Afoundationforthestudyofbehaviorchangesupportsystems,"(inEnglish),PersonalandUbiquitousComputing,vol.17,no.6,pp.1223‐1235,2013/08/012013.

[175] S.J.Brown,D.A.Lieberman,B.A.Gemeny,Y.C.Fan,D.M.Wilson,andD.J.Pasta,"Educationalvideogameforjuvenilediabetes:resultsofacontrolledtrial,"InformaticsforHealthandSocialCare,vol.22,no.1,pp.77‐89,1997.

[176] P.M.Kato,S.W.Cole,A.S.Bradly,andB.H.Pollock,"Avideogameimprovesbehavioraloutcomesinadolescentsandyoungadultswithcancer:arandomizedtrial,"Pediatrics,vol.122,p.e305,2007.

[177] A.C.McPherson,C.Glazebrook,D.Forster,C.James,andA.Smyth,"Arandomized,controlledtrialofaninteractiveeducationalcomputerpackageforchildrenwithasthma,"Pediatrics,vol.117,pp.1046‐1054,2006.

[178] K.vanderKooij,E.Hoogendoorn,R.Spijkerman,andV.Visch,"ValidationofGamesforBehavioralChange:ConnectingthePlayfulandSerious,,"TheInternationalJournalofSeriousGames,vol.2,no.3,pp.63‐75,2015.

[179] B.Laugwitz,T.Held,andM.Schrepp,Constructionandevaluationofauserexperiencequestionnaire.Springer,2008.

[180] J.Brooke,"SUS‐Aquickanddirtyusabilityscale,"Usabilityevaluationinindustry,vol.189,no.194,pp.4‐7,1996.

References

174

[181] T.L.Mayes,I.H.Bernstein,C.L.Haley,B.D.Kennard,andG.J.Emslie,"PsychometricpropertiesoftheChildren'sDepressionRatingScale–Revisedinadolescents,"Journalofchildandadolescentpsychopharmacology,vol.20,no.6,pp.513‐516,2010.

[182] S.Deterding,D.Dixon,R.Khaled,andL.Nacke,"Fromgamedesignelementstogamefulness:defining"gamification","presentedattheProceedingsofthe15thInternationalAcademicMindTrekConference:EnvisioningFutureMediaEnvironments,Tampere,Finland,2011.

[183] R.Hunicke,LeBlanc,M.,Zubek,R.,"MDA:AFormalApproachtoGameDesignandGameResearch,"ProceedingsoftheChallengesinGameAIWorkshop,NineteenthNationalConferenceonArtificialIntelligence.,2004.

[184] S.J.Hu,"EvolvingParadigmsofManufacturing:FromMassProductiontoMassCustomizationandPersonalization,"ProcediaCIRP,vol.7,no.0,pp.3‐8,//2013.

[185] J.WattsandM.Hirst,"Userparticipationintheearlystagesofbuildingdesign,"DesignStudies,vol.3,no.1,pp.11‐18,1982/01/011982.

[186] W.D.Hoyer,Chandy,R.,Dorotic,M.,Krafft,M.,&Singh,S.S.,"Consumercocreationinnewproductdevelopment,"JournalofServiceResearch,vol.13,no.3,pp.283‐296,2010.

[187] I.KatzandA.Assor,"Whenchoicemotivatesandwhenitdoesnot,"EducationalPsychologyReview,vol.19,no.4,pp.429‐442,2007.

[188] G.N.Miceli,F.Ricotta,andM.Costabile,"Customizingcustomization:Aconceptualframeworkforinteractivepersonalization,"JournalofInteractiveMarketing,vol.21,no.2,pp.6‐25,2007.

[189] K.Miller,S.Rodger,B.Kipping,andR.M.Kimble,"Anoveltechnologyapproachtopainmanagementinchildrenwithburns:Aprospectiverandomizedcontrolledtrial,"(ineng),Burns,vol.37,no.3,pp.395‐405,May2011.

[190] E.Malbos,R.M.Rapee,andM.Kavakli,"Creationofinteractivevirtualenvironmentsforexposuretherapythroughgame‐leveleditors:Comparisonandtestsonpresenceandanxiety,"InternationalJournalofHuman‐ComputerInteraction,vol.29,no.12,pp.827‐837,Dec2013.

[191] R.Orji,J.Vassileva,andR.L.Mandryk,"Modelingtheefficacyofpersuasivestrategiesfordifferentgamertypesinseriousgamesforhealth,"UserModelingandUser‐AdaptedInteraction,vol.24,no.5,pp.453‐498,2014.

[192] S.Deterding,M.Sicart,L.Nacke,K.O'Hara,andD.Dixon,"Gamification.usinggame‐designelementsinnon‐gamingcontexts,"presentedattheCHI'11ExtendedAbstractsonHumanFactorsinComputingSystems,Vancouver,BC,Canada,2011.

[193] T.Baranowskietal.,"GamesforHealthforChildren‐CurrentStatusandNeededResearch,"(ineng),Gamesforhealthjournal,vol.5,no.1,pp.1‐12,Feb2016.

[194] M.LynskeyandW.Hall,"Theeffectsofadolescentcannabisuseoneducationalattainment:areview,"Addiction,vol.95,no.11,pp.1621‐1630,2000.

[195] J.R.Weisz,B.Weiss,S.S.Han,D.A.Granger,andT.Morton,"Effectsofpsychotherapywithchildrenandadolescentsrevisited:ameta‐analysisof


175

treatmentoutcomestudies,"(ineng),PsychologicalBulletin,vol.117,no.3,pp.450‐468,May1995.

[196] S.J.E.Bruijniks,G.Franx,andM.J.H.Huibers,"Theimplementationandadherencetoevidence‐basedprotocolsforpsychotherapyfordepression:theperspectiveoftherapistsinDutchspecializedmentalhealthcare,"(ineng),BMCPsychiatry,vol.18,no.1,p.190,Jun142018.

[197] S.Gibson,S.L.Brand,S.Burt,Z.V.R.Boden,andO.Benson,"Understandingtreatmentnon‐adherenceinschizophreniaandbipolardisorder:asurveyofwhatserviceusersdoandwhy,"BMCPsychiatry,journalarticlevol.13,no.1,p.153,May292013.

[198] M.E.AddisandA.D.Krasnow,"Anationalsurveyofpracticingpsychologists'attitudestowardpsychotherapytreatmentmanuals,"JournalofConsultingandClinicalPsychology,vol.68,no.2,pp.331‐339,2000.

[199] M.Wilhelmsenetal.,"NorwegianGeneralPractitioners’PerspectivesonImplementationofaGuidedWeb‐BasedCognitiveBehavioralTherapyforDepression:AQualitativeStudy,"JournalofMedicalInternetResearch,vol.16,no.9,p.e208,2014.

[200] M.Kivi,M.C.M.Eriksson,D.Hange,E.‐L.Petersson,C.Björkelund,andB.Johansson,"Experiencesandattitudesofprimarycaretherapistsintheimplementationanduseofinternet‐basedtreatmentinSwedishprimarycaresettings,"InternetInterventions,vol.2,no.3,pp.248‐256,2015/09/01/2015.

[201] J.Bengtsson,S.Nordin,andP.Carlbring,"Therapists'ExperiencesofConductingCognitiveBehaviouralTherapyOnlinevis‐à‐visFace‐to‐Face,"CognitiveBehaviourTherapy,vol.44,no.6,pp.470‐479,2015/11/022015.

[202] H.ScholtenandI.Granic,"UseofthePrinciplesofDesignThinkingtoAddressLimitationsofDigitalMentalHealthInterventionsforYouth:Viewpoint,"(inEnglish),JMedInternetRes,Viewpointvol.21,no.1,p.e11528,2019.

[203] G.M.Schippers,M.Smeerdijk,andM.J.M.Merkx,HandboekCGTbijmiddelengebruikengokkenUtrecht:PerspectiefUitgevers,2014.

[204] M.M.M.VanDooren,V.T.Visch,R.Spijkerman,R.H.M.Goossens,andV.M.Hendriks,"PersonalizationinGameDesignforHealthcare:aLiteratureReviewonitsDefinitionsandEffects,"InternationalJournalofSeriousGames,vol.3,no.4,pp.3‐28,2016.

[205] B.C.O'Brien,I.B.Harris,T.J.Beckman,D.A.Reed,andD.A.Cook,"Standardsforreportingqualitativeresearch:asynthesisofrecommendations,"(ineng),AcadMed,vol.89,no.9,pp.1245‐51,Sep2014.

[206] K.CharmazandL.L.Belgrave,"Groundedtheory,"TheBlackwellencyclopediaofsociology,2007.

[207] V.BraunandV.Clarke,"Usingthematicanalysisinpsychology,"QualResPsychol,vol.3,2006.

[208] J.Aronson,"Apragmaticviewofthematicanalysis,"Thequalitativereport,vol.2,no.1,pp.1‐3,1994.

[209] A.Tong,P.Sainsbury,andJ.Craig,"Consolidatedcriteriaforreportingqualitativeresearch(COREQ):a32‐itemchecklistforinterviewsandfocus

References

176

groups,"InternationalJournalforQualityinHealthCare,vol.19,no.6,pp.349‐357,2007.

[210] B.C.ChuandP.C.Kendall,"TherapistResponsivenesstoChildEngagement:FlexibilityWithinManual‐BasedCBTforAnxiousYouth,"JournalofClinicalPsychology,vol.65,no.7,pp.736‐7542009.

[211] D.Schulte,R.Künzel,G.Pepping,andT.Schulte‐Bahrenberg,"Tailor‐madeversusstandardizedtherapyofphobicpatients,"AdvancesinBehaviourResearchandTherapy,vol.14,no.2,pp.67‐92,1992.

[212] J.Morgenstern,K.A.Blanchard,T.J.Morgan,E.Labouvie,andJ.Hayaki,"Testingtheeffectivenessofcognitive‐behavioraltreatmentforsubstanceabuseinacommunitysetting:withintreatmentandposttreatmentfindings,"JournalofConsultingandClinicalPsychology,vol.69,no.6,p.1007,2001.

[213] J.T.Levitt,L.S.Malta,A.Martin,L.Davis,andM.Cloitre,"TheflexibleapplicationofamanualizedtreatmentforPTSDsymptomsandfunctionalimpairmentrelatedtothe9/11WorldTradeCenterattack,"BehaviourResearchandTherapy,vol.45,no.7,pp.1419‐1433,2007.

[214] D.A.Chambers,R.E.Glasgow,andK.C.Stange,"Thedynamicsustainabilityframework:addressingtheparadoxofsustainmentamidongoingchange,"ImplementationScience,journalarticlevol.8,no.1,p.117,2013.

[215] D.L.Tobin,J.D.Banker,L.Weisberg,andW.Bowers,"Iknowwhatyoudidlastsummer(anditwasnotCBT):afactoranalyticmodelofinternationalpsychotherapeuticpracticeintheeatingdisorders,"(ineng),TheInternationalJournalofEatingDisorders,vol.40,no.8,pp.754‐157,Dec2007.

[216] P.C.Kendall,B.C.Chu,A.Gifford,C.Hayes,andM.Nauta,"Breathinglifeintoamanual:Flexibilityandcreativitywithmanual‐basedtreatments,"CognitiveandBehavioralPractice,vol.5,no.2,pp.177–198,1998.

[217] H.F.Zickgrafetal.,"InterpersonalFactorsAreAssociatedwithLowerTherapistAdherenceinCognitive‐BehaviouralTherapyforPanicDisorder,"ClinicalPsychology&Psychotherapy,vol.23,no.3,pp.272‐284,2016.

[218] A.Hogue,S.Dauber,E.Lichvar,M.Bobek,andC.E.Henderson,"ValidityofTherapistSelf‐ReportRatingsofFidelitytoEvidence‐BasedPracticesforAdolescentBehaviorProblems:CorrespondencebetweenTherapistsandObservers,"AdministrationandPolicyinMentalHealthandMentalHealthServicesResearch,vol.42,no.2,pp.229‐243,2015/03/012015.

[219] K.M.Peavyetal.,"Treatmentadherenceandcompetencyratingsamongtherapists,supervisors,study‐relatedratersandexternalratersinaclinicaltrialofa12‐stepfacilitationforstimulantusers,"(ineng),Journalofsubstanceabusetreatment,vol.47,no.3,pp.222‐228,2014.

[220] V.Tschuschkeetal.,"Theroleoftherapists'treatmentadherence,professionalexperience,therapeuticalliance,andclients'severityofpsychologicalproblems:Predictionoftreatmentoutcomeineightdifferentpsychotherapyapproaches.Preliminaryresultsofanaturalisticstudy,"PsychotherapyResearch,vol.25,no.4,pp.420‐434,2015.

[221] A.Hackshaw,"Smallstudies:strengthsandlimitations,"EuropeanRespiratoryJournal,vol.32,no.5,pp.1141‐1143,2008.


177

[222] J.Arrasvuorietal.,"ApplyingthePLEXframeworkindesigningforplayfulness,"presentedattheProceedingsofthe2011ConferenceonDesigningPleasurableProductsandInterfaces,Milano,Italy,2011.

[223] Andretal.,"Theplayfulexperiences(PLEX)frameworkasaguideforexpertevaluation,"presentedattheProceedingsofthe6thInternationalConferenceonDesigningPleasurableProductsandInterfaces,NewcastleuponTyne,UnitedKingdom,2013.

[224] D.R.Davis,A.N.Kurti,J.M.Skelly,R.Redner,T.J.White,andS.T.Higgins,"Areviewoftheliteratureoncontingencymanagementinthetreatmentofsubstanceusedisorders,2009–2014,"PreventiveMedicine,vol.92,pp.36‐46,2016.

[225] J.Q.Purnell,R.Gernes,R.Stein,M.S.Sherraden,andA.Knoblock‐Hahn,"Asystematicreviewoffinancialincentivesfordietarybehaviorchange,"JournaloftheAcademyofNutritionandDietetics,vol.114,no.7,pp.1023‐1035,2014.

[226] E.L.Giles,S.Robalino,E.McColl,F.F.Sniehotta,andJ.Adams,"Theeffectivenessoffinancialincentivesforhealthbehaviourchange:systematicreviewandmeta‐analysis,"PLOSONE,vol.9,no.3,p.e90347,2014.

[227] E.Mantzari,F.Vogt,I.Shemilt,Y.Wei,J.P.Higgins,andT.M.Marteau,"Personalfinancialincentivesforchanginghabitualhealth‐relatedbehaviors:asystematicreviewandmeta‐analysis,"PreventiveMedicine,vol.75,pp.75‐85,2015.

[228] F.LuthansandA.D.Stajkovic,"Reinforceforperformance:theneedtogobeyondpayandevenrewards,"TheAcademyofManagementExecutive(1993‐2005),vol.13,no.2,pp.49‐57,1999.

[229] P.SweetserandP.Wyeth,"GameFlow:amodelforevaluatingplayerenjoymentingames,"ComputersinEntertainment(CIE),vol.3,no.3,pp.3‐3,2005.

[230] R.M.Ryan,C.S.Rigby,andA.Przybylski,"Themotivationalpullofvideogames:Aself‐determinationtheoryapproach,"Motivationandemotion,vol.30,no.4,pp.344‐360,2006.

[231] J.O.Prochaska,J.C.Norcross,andC.C.DiClemente,"Applyingthestagesofchange,"PsychotherapyinAustralia,vol.19,no.2,pp.10‐15,2013.

[232] J.Cameron,K.M.Banko,andW.D.Pierce,"Pervasivenegativeeffectsofrewardsonintrinsicmotivation:themythcontinues,"TheBehaviorAnalyst,vol.24,no.1,pp.1‐44,2001.

[233] C.I.Muntean,"Raisingengagementine‐learningthroughgamification,"inProc.ofthe6thInternationalConferenceonVirtualLearning,ICVL2011,Cluj‐Napoca,Romania,2011,pp.323‐329.

[234] J.Hamari,J.Koivisto,andH.Sarsa,"Doesgamificationwork?‐‐aliteraturereviewofempiricalstudiesongamification,"inProc.ofthe47thHawaiiInternationalConferenceonSystemSciences,HICSS2014,Waikoloa,Hawaii,2014,pp.3025‐3034:IEEE.

[235] R.Garris,R.Ahlers,andJ.E.Driskell,"Games,motivationandlearning:aresearchandpracticemodel,"SimulationGaming,vol.33,pp.441‐467,2002.

[236] C.P.Cerasoli,J.M.Nicklin,andM.T.Ford,"Intrinsicmotivationandextrinsicincentivesjointlypredictperformance:a40‐yearmeta‐analysis,"(ineng),PsychologicalBulletin,vol.140,no.4,pp.980‐1008,Jul2014.

References

178

[237] M.Papastergiou,"Exploringthepotentialofcomputerandvideogamesforhealthandphysicaleducation:Aliteraturereview,"Computers&Education,vol.53,no.3,pp.603‐622,2009.

[238] A.Pateletal.,"Distractionwithahand‐heldvideogamereducespediatricpreoperativeanxiety,"PediatricAnesthesia,vol.16,no.10,pp.1019‐1027,2006.

[239] T.J.O'Connoretal.,"Evaluationofamanualwheelchairinterfacetocomputergames,"NeurorehabilitationandNeuralRepair,vol.14,no.1,pp.21‐31,2000.

[240] S.G.Fitzgerald,R.A.Cooper,T.Thorman,R.Cooper,S.Guo,andM.L.Boninger,"TheGAME(Cycle)exercisesystem:comparisonwithstandardergometry,"TheJournalofSpinalCordMedicine,vol.27,no.5,pp.453‐9,2004.

[241] S.R.Sharar,G.J.Carrougher,D.Nakamura,H.G.Hoffman,D.K.Blough,andD.R.Patterson,"Factorsinfluencingtheefficacyofvirtualrealitydistractionanalgesiaduringpostburnphysicaltherapy:preliminaryresultsfrom3ongoingstudies,"ArchivesofPhysicalMedicineandRehabilitation,vol.88,no.12,pp.S43‐S49,2007.

[242] S.J.Brown,D.A.Lieberman,B.Gemeny,Y.C.Fan,D.Wilson,andD.Pasta,"Educationalvideogameforjuvenilediabetes:resultsofacontrolledtrial,"MedicalInformatics,vol.22,no.1,pp.77‐89,1997.

[243] D.A.Lieberman,"Managementofchronicpediatricdiseaseswithinteractivehealthgames:Theoryandresearchfindings,"TheJournalofAmbulatoryCareManagement,vol.24,no.1,pp.26‐38,2001.

[244] M.D.HanusandJ.Fox,"Assessingtheeffectsofgamificationintheclassroom:Alongitudinalstudyonintrinsicmotivation,socialcomparison,satisfaction,effort,andacademicperformance,"Computers&Education,vol.80,pp.152‐161,2015.

[245] W.J.Boendermaker,P.J.M.Prins,andR.W.Wiers,"CognitiveBiasModificationforadolescentswithsubstanceuseproblems–Canseriousgameshelp?,"JournalofBehaviorTherapyandExperimentalPsychiatry,vol.49,pp.13‐20,2015/12/01/2015.

[246] K.C.BerridgeandM.L.Kringelbach,"Affectiveneuroscienceofpleasure:rewardinhumansandanimals,"(ineng),Psychopharmacology,vol.199,no.3,pp.457‐480,2008.

[247] P.DayanandB.W.Balleine,"Reward,motivation,andreinforcementlearning,"Neuron,vol.36,no.2,pp.285‐98,2002.

[248] R.M.RyanandE.L.Deci,"IntrinsicandExtrinsicMotivations:ClassicDefinitionsandNewDirections,"ContemporaryEducationalPsychology,vol.25,no.1,pp.54‐67,2000.

[249] C.C.RuffandE.Fehr,"Theneurobiologyofrewardsandvaluesinsocialdecisionmaking,"(ineng),Naturereviews.Neuroscience,vol.15,no.8,pp.549‐562,2014.

[250] W.Schultz,"Behavioraltheoriesandtheneurophysiologyofreward,"Annualreviewofpsychology,vol.57,pp.87‐115,2006.


179

[251] H.WangandC.‐T.Sun,"Gamerewardsystems:Gamingexperiencesandsocialmeanings,"inProc.ofthe2011DiGRAInternationalConference:ThinkDesignPlay,DiGRA'11,Utrecht,Netherlands,2011.

[252] J.HamariandV.Eranti,"FrameworkforDesigningandEvaluatingGameAchievements,"inProc.ofthe2011DiGRAInternationalConference:ThinkDesignPlay,DiGRA'11,Utrecht,Netherlands,2011.

[253] A.K.Przybylski,C.S.Rigby,andR.M.Ryan,"Amotivationalmodelofvideogameengagement,"ReviewofGeneralPsychology,vol.14,no.2,pp.154‐166,2010.

[254] H.Braun,I.Kirsch,andK.Yamamoto,"AnExperimentalStudyoftheEffectsofMonetaryIncentivesonPerformanceonthe12th‐GradeNAEPReadingAssessment,"TeachersCollegeRecord,vol.113,no.11,pp.2309‐2344,2011.

[255] G.Kohls,J.Peltzer,B.Herpertz‐Dahlmann,andK.Konrad,"Differentialeffectsofsocialandnon‐socialrewardonresponseinhibitioninchildrenandadolescents,"DevelopmentalScience,vol.12,no.4,pp.614‐625,2009.

[256] A.D.StajkovicandF.Luthans,"DifferentialEffectsofIncentiveMotivatorsonWorkPerformance,"TheAcademyofManagementJournal,vol.44,no.3,pp.580‐590,2001.

[257] W.‐H.Choietal.,"AnfMRIStudyInvestigatingAdolescentBrainActivationbyRewardsandFeedback,"PsychiatryInvestigation,vol.10,no.1,pp.47‐55,2013.

[258] M.J.Koeppetal.,"Evidenceforstriataldopaminereleaseduringavideogame,"(ineng),Nature,vol.393,no.6682,pp.266‐8,May211998.

[259] N.D.Volkow,J.S.Fowler,G.J.Wang,andJ.M.Swanson,"DopamineinDrugAbuseandAddiction:ResultsFromImagingStudiesandTreatmentImplications,"MolecularPsychiatry,vol.9,pp.557‐569,2004.

[260] P.Siriaraya,V.Visch,A.Vermeeren,andM.Bas,"AcookbookmethodforPersuasiveGameDesign,"InternationalJournalofSeriousGames,vol.5,no.1,pp.37‐71,2018.

[261] G.F.KoobandM.LeMoal,"Drugaddiction,dysregulationofreward,andallostasis,"Neuropsychopharmacology,vol.24,no.2,pp.97‐129,2001.

[262] J.Reuter,T.Raedler,M.Rose,I.Hand,J.Glascher,andC.Buchel,"Pathologicalgamblingislinkedtoreducedactivationofthemesolimbicrewardsystem,"NatureNeuroscience,vol.8,no.2,pp.147‐148,2005.

[263] N.D.Volkow,G.‐J.Wang,J.S.Fowler,D.Tomasi,F.Telang,andR.Baler,"Addiction:Decreasedrewardsensitivityandincreasedexpectationsensitivityconspiretooverwhelmthebrain'scontrolcircuit,"BioEssays:newsandreviewsinmolecular,cellularanddevelopmentalbiology,vol.32,no.9,pp.748‐755,2010.

[264] G.F.KoobandM.LeMoal,"Plasticityofrewardneurocircuitryandthe'darkside'ofdrugaddiction,"NatureNeuroscience,vol.8,p.1442,2005.

[265] J.‐E.Kimetal.,"NeuralresponsestovariousrewardsandfeedbackinthebrainsofadolescentInternetaddictsdetectedbyfunctionalmagneticresonanceimaging,"PsychiatryandClinicalNeurosciences,vol.68,no.6,pp.463‐470,2014.

References

180

[266] N.D.Volkow,G.J.Wang,J.S.Fowler,D.Tomasi,andF.Telang,"Addiction:beyonddopaminerewardcircuitry,"ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica,vol.108,no.37,pp.15037‐15042,2011.

[267] M.Prendergast,D.Podus,J.Finney,L.Greenwell,andJ.Roll,"Contingencymanagementfortreatmentofsubstanceusedisorders:ameta‐analysis,"Addiction,vol.101,no.11,pp.1546‐1560,2006.

[268] S.Shoptawetal.,"BehavioraltreatmentapproachesformethamphetaminedependenceandHIV‐relatedsexualriskbehaviorsamongurbangayandbisexualmen,"DrugandAlcoholDependence,vol.78,no.2,pp.125‐134,2005.

[269] S.Minozzi,R.Saulle,F.DeCrescenzo,andL.Amato,"Psychosocialinterventionsforpsychostimulantmisuse,"CochraneDatabaseofSystematicReviews,vol.9,p.Cd011866,2016.

[270] J.P.Lussier,S.H.Heil,J.A.Mongeon,G.J.Badger,andS.T.Higgins,"Ameta‐analysisofvoucher‐basedreinforcementtherapyforsubstanceusedisorders,"Addiction,vol.101,no.2,pp.192‐203,2006.

[271] W.M.Compton,3rd,L.B.Cottler,A.BenAbdallah,D.L.Phelps,E.L.Spitznagel,andJ.C.Horton,"Substancedependenceandotherpsychiatricdisordersamongdrugdependentsubjects:raceandgendercorrelates,"TheAmericanJournalonAddictions,vol.9,no.2,pp.113‐125,2000.

[272] S.E.Young,R.P.Corley,M.C.Stallings,S.H.Rhee,T.J.Crowley,andJ.K.Hewitt,"Substanceuse,abuseanddependenceinadolescence:prevalence,symptomprofilesandcorrelates,"DrugandAlcoholDependence,vol.68,no.3,pp.309‐322,2002.

[273] S.IVZ.(2016).KerncijfersVerslavingszorg2015.LADIS,LandelijkAlcoholenDrugsInformatieSysteem.Achieved02‐02‐2019fromhttps://www.ggznederland.nl/uploads/assets/Werkdocument%20KC%202015%20.pdf.

[274] R.RigterandP.Blanken,"KerncijfersBrijderJeugd2016.,"DenHaag:ParnassiaAddictionResearchCentre(PARC),BrijderVerslavingszorg,ParnassiaGroep,2017.

[275] JannetdeJonge,RenskeSpijkerman,andM.Müller,HandleidingCGT‐J.Cognitievegedragstherapeutischebehandelingvanproblematischmiddelengebruik,gokkenengamenbijjongerenResultatenScoren&PerspectiefUitgevers2018.

[276] M.Ventura,V.Shute,andW.Zhao,"Therelationshipbetweenvideogameuseandaperformance‐basedmeasureofpersistence,"Computers&Education,vol.60,no.1,pp.52‐58,2013.

[277] L.VanLeijenhorst,B.GuntherMoor,Z.A.OpdeMacks,S.A.Rombouts,P.M.Westenberg,andE.A.Crone,"Adolescentriskydecision‐making:neurocognitivedevelopmentofrewardandcontrolregions,"Neuroimage,vol.51,no.1,pp.345‐355,2010.

[278] B.R.Braams,A.C.vanDuijvenvoorde,J.S.Peper,andE.A.Crone,"Longitudinalchangesinadolescentrisk‐taking:acomprehensivestudyofneuralresponsestorewards,pubertaldevelopment,andrisk‐takingbehavior,"TheJournalof


181

neuroscience:theofficialjournaloftheSocietyforNeuroscience,vol.35,no.18,pp.7226‐7238,May62015.

[279] A.Galvanetal.,"Earlierdevelopmentoftheaccumbensrelativetoorbitofrontalcortexmightunderlierisk‐takingbehaviorinadolescents,"TheJournalofneuroscience:theofficialjournaloftheSocietyforNeuroscience,vol.26,no.25,pp.6885‐92,2006.

[280] N.Shabihi,F.Taghiyareh,andM.H.Abdoli,"Analyzingtheeffectofgame‐elementsine‐learningenvironmentsthroughMBTI‐basedpersonalization,"inTelecommunications(IST),20168thInternationalSymposiumon,2016,pp.612‐618:IEEE.

[281] M.R.MitchellandM.N.Potenza,"AddictionsandPersonalityTraits:ImpulsivityandRelatedConstructs,"CurrentBehavioralNeuroscienceReports,vol.1,no.1,pp.1‐12,2014.

[282] G.F.Tondello,R.R.Wehbe,L.Diamond,M.Busch,A.Marczewski,andL.E.Nacke,"TheGamificationUserTypesHexadScale,"presentedattheProceedingsofthe2016AnnualSymposiumonComputer‐HumanInteractioninPlay,Austin,Texas,USA,2016.

[283] R.Kotov,W.Gamez,F.Schmidt,andD.Watson,"Linking“big”personalitytraitstoanxiety,depressive,andsubstanceusedisorders:Ameta‐analysis,"PsychologicalBulletin,vol.136,no.5,pp.768‐821,2010.

[284] B.R.Braams,S.Peters,J.S.Peper,B.Guroglu,andE.A.Crone,"Gamblingforself,friends,andantagonists:differentialcontributionsofaffectiveandsocialbrainregionsonadolescentrewardprocessing,"Neuroimage,vol.100,pp.281‐289,2014.

[285] N.Vegt,V.Visch,A.Vermeeren,andH.Ridder,ACaseStudyonGamifiedInterventionsforTeamCohesioninFactoryWork.2018,pp.176‐208.

[286] I.Ajzen,"PerceivedBehavioralControl,Self‐Efficacy,LocusofControl,andtheTheoryofPlannedBehavior1,"JournalofAppliedSocialPsychology,vol.32,no.4,pp.665‐683,2002.

[287] S.Gilbert,DesigningGamifiedSystems:Meaningfulplayininteractiveentertainment,marketingandeducation.FocalPress,2015.

[288] J.Schell,"Theartofgamedesign:abookoflenses,"ed:SanFrancisco,Calif.:MorganKaufmann;Oxford:ElsevierScience[distributor],2008.

[289] K.SalenandE.Zimmerman,RulesofPlay:GameDesignFundamentals.MITpress,2004.

[290] T.Hainey,T.Connolly,M.Stansfield,andE.Boyle,"Thedifferencesinmotivationsofonlinegameplayersandofflinegameplayers:Acombinedanalysisofthreestudiesathighereducationlevel,"Computers&Education,vol.57,no.4,pp.2197‐2211,2011.

[291] L.VanLeijenhorst,K.Zanolie,C.S.VanMeel,P.M.Westenberg,S.A.R.B.Rombouts,andE.A.Crone,"WhatMotivatestheAdolescent?BrainRegionsMediatingRewardSensitivityacrossAdolescence,"CerebralCortex,vol.20,no.1,pp.61‐69,2010.

References

182

[292] K.N.Spreckelmeyeretal.,"Anticipationofmonetaryandsocialrewarddifferentlyactivatesmesolimbicbrainstructuresinmenandwomen,"Socialcognitiveandaffectiveneuroscience,vol.4,no.2,pp.158‐165,2009.

[293] B.Knutson,G.W.Fong,C.M.Adams,J.L.Varner,andD.Hommer,"Dissociationofrewardanticipationandoutcomewithevent‐relatedfMRI,"Neuroreport,vol.12,no.17,pp.3683‐3687,2001.

[294] E.Demurie,H.Roeyers,D.Baeyens,andE.Sonuga‐Barke,"CommonalterationsinsensitivitytotypebutnotamountofrewardinADHDandautismspectrumdisorders,"JournalofChildPsychologyandPsychiatry,vol.52,no.11,pp.1164‐1173,2011.

[295] A.A.Scott‐VanZeeland,M.Dapretto,D.G.Ghahremani,R.A.Poldrack,andS.Y.Bookheimer,"Rewardprocessinginautism,"Autismresearch:officialjournaloftheInternationalSocietyforAutismResearch,vol.3,no.2,pp.53‐67,2010.

[296] J.A.Richeyetal.,"Commonanddistinctneuralfeaturesofsocialandnon‐socialrewardprocessinginautismandsocialanxietydisorder,"Socialcognitiveandaffectiveneuroscience,vol.9,no.3,pp.367‐377,2014.

[297] H.Oinas‐Kukkonen,"Behaviorchangesupportsystems:aresearchmodelandagenda,"inProc.ofthe5thInternationalConferenceonPersuasiveTechnology,PERSUASIVE2010,Copenhagen,Denmark,2010,pp.4‐14:SpringerBerlinHeidelberg.

[298] M.M.M.v.Dooren,R.Spijkerman,R.H.M.Goossens,V.M.Hendriks,andV.T.Visch,"PLEXasinputandevaluationtoolinpersuasivegamedesign:pilotstudy,"presentedattheProceedingsofthefirstACMSIGCHIannualsymposiumonComputer‐humaninteractioninplay,Toronto,Ontario,Canada,2014.

[299] P.Siriaraya,V.T.ValentijnVisch,M.M.M.vanDooren,andR.Spijkerman,"Learningsandchallengesindesigninggamificationsformentalhealthcare:thecasestudyoftheReadySetGoalsApplication,"in10thInternationalConferenceonVirtualWorldsandGamesforSeriousApplications(VSGames2018),2018:IEEE.

[300] M.Krank,S.H.Stewart,R.O'Connor,P.B.Woicik,A.‐M.Wall,andP.J.Conrod,"Structural,concurrent,andpredictivevalidityoftheSubstanceUseRiskProfileScaleinearlyadolescence,"Addictivebehaviors,vol.36,no.1,pp.37‐46,2011.

[301] C.W.LangfredandN.A.Moye,"Effectsoftaskautonomyonperformance:anextendedmodelconsideringmotivational,informational,andstructuralmechanisms,"TheJournalofappliedpsychology,vol.89,no.6,pp.934‐945,2004.

[302] H.M.Lau,J.H.Smit,T.M.Fleming,andH.Riper,"SeriousGamesforMentalHealth:AreTheyAccessible,Feasible,andEffective?ASystematicReviewandMeta‐analysis,"(ineng),FrontPsychiatry,vol.7,p.209,2016.

[303] R.C.Kessler,P.Berglund,O.Demler,R.Jin,K.R.Merikangas,andE.E.Walters,"Lifetimeprevalenceandage‐of‐onsetdistributionsofDSM‐IVdisordersintheNationalComorbiditySurveyReplication,"(ineng),ArchGenPsychiatry,vol.62,no.6,pp.593‐602,Jun2005.

[304] R.W.Wiers,C.Eberl,M.Rinck,E.S.Becker,andJ.Lindenmeyer,"Retrainingautomaticactiontendencieschangesalcoholicpatients'approachbiasfor


183

alcoholandimprovestreatmentoutcome,"(ineng),PsycholSci,vol.22,no.4,pp.490‐7,Apr2011.

[305] S.Nicholson,"Auser‐centeredtheoreticalframeworkformeaningfulgamification,"Games+Learning+Society,vol.8,no.1,pp.223‐230,2012.

[306] R.Tate,J.Haritatos,andS.Cole,HopeLab'sApproachtoReMission.2009,pp.29‐35.

[307] D.Thompson,T.Baranowski,andR.Buday,"Conceptualmodelforthedesignofaseriousvideogamepromotingself‐managementamongyouthwithtype1diabetes,"Journalofdiabetesscienceandtechnology,vol.4,no.3,pp.744‐749,2010.

[308] P.Cuijpers,M.Berking,G.Andersson,L.Quigley,A.Kleiboer,andK.S.Dobson,"Ameta‐analysisofcognitive‐behaviouraltherapyforadultdepression,aloneandincomparisonwithothertreatments,"(ineng),CanJPsychiatry,vol.58,no.7,pp.376‐85,Jul2013.

[309] A.C.Butler,J.E.Chapman,E.M.Forman,andA.T.Beck,"Theempiricalstatusofcognitive‐behavioraltherapy:areviewofmeta‐analyses,"ClinicalPsychologyReview,vol.26,no.1,pp.17‐31,2006.

[310] J.Zimmerman,J.Forlizzi,andS.Evenson,"ResearchthroughdesignasamethodforinteractiondesignresearchinHCI,"presentedattheProceedingsoftheSIGCHIConferenceonHumanFactorsinComputingSystems,SanJose,California,USA,2007.

[311] W.Gaver,"Whatshouldweexpectfromresearchthroughdesign?,"presentedattheProceedingsoftheSIGCHIConferenceonHumanFactorsinComputingSystems,Austin,Texas,USA,2012.

[312] C.E.Wiersetal.,"Effectsofcognitivebiasmodificationtrainingonneuralalcoholcuereactivityinalcoholdependence,"(ineng),AmJPsychiatry,vol.172,no.4,pp.335‐43,Apr2015.

[313] P.J.ConrodandK.Nikolaou,"AnnualResearchReview:Onthedevelopmentalneuropsychologyofsubstanceusedisorders,"(ineng),JChildPsycholPsychiatry,vol.57,no.3,pp.371‐94,Mar2016.

[314] M.Malmberg,G.Overbeek,K.Monshouwer,J.Lammers,W.A.Vollebergh,andR.C.Engels,"Substanceuseriskprofilesandassociationswithearlysubstanceuseinadolescence,"(ineng),JBehavMed,vol.33,no.6,pp.474‐85,Dec2010.

[315] N.Castellanos‐RyanandP.Conrod,"Personalityandsubstancemisuse:Evidenceforafour‐factormodelofvulnerability,"inDrugabuseandaddictioninmedicalillness:Causes,consequencesandtreatment.NewYork,NY,US:SpringerScience+BusinessMedia,2012,pp.47‐62.

[316] P.A.Woicik,S.H.Stewart,R.O.Pihl,andP.J.Conrod,"Thesubstanceuseriskprofilescale:Ascalemeasuringtraitslinkedtoreinforcement‐specificsubstanceuseprofiles,"Addictivebehaviors,vol.34,no.12,pp.1042‐1055,2009.

[317] M.Zuckerman,"SensationSeekingandSubstanceUseandAbuse:Smoking,Drinking,andDrugs,"inSensationseekingandriskybehavior.Washington,DC,US:AmericanPsychologicalAssociation,2007,pp.107‐143.

References

184

[318] M.M.vanDooren,V.T.Visch,andR.Spijkerman,"TheDesignandApplicationofGameRewardsinYouthAddictionCare,"Information,vol.10,no.4,p.126,2019.

[319] Y.Francillette,A.Gouaïch,N.Hocine,andJ.Pons,"Agameplayloopsformallanguage,"in201217thInternationalConferenceonComputerGames(CGAMES),2012,pp.94‐101.

[320] L.Andreas,M.Max,andM.AndreeaCatalina,"DeepandShallowGamificationinMarketing:ThinEvidenceandtheForgottenPowersofReallyGoodGames,"inEngagingConsumersthroughBrandedEntertainmentandConvergentMediaHershey,PA,USA:IGIGlobal,2015,pp.110‐126.

[321] W.IJsselsteijn,Y.DeKort,andK.Poels,"{TheGameExperienceQuestionnaire:Developmentofaself‐reportmeasuretoassessthepsychologicalimpactofdigitalgames.ManuscriptinPreparation},"2013.

[322] N.Gillet,R.J.Vallerand,andM.‐A.K.Lafrenière,"Intrinsicandextrinsicschoolmotivationasafunctionofage:themediatingroleofautonomysupport,"SocialPsychologyofEducation,vol.15,no.1,pp.77‐95,2012/03/012012.

[323] T.M.KellyandD.C.Daley,"Integratedtreatmentofsubstanceuseandpsychiatricdisorders,"Socialworkinpublichealth,vol.28,no.3‐4,pp.388‐406,2013.

[324] P.J.Prins,Dovis,S.,Ponsioen,A.,TenBrink,E.,&VanderOord,S.,"DoescomputerizedworkingmemorytrainingwithgameelementsenhancemotivationandtrainingefficacyinchildrenwithADHD?,"Cyberpsychology,behavior,andsocialnetworking,,vol.14,no.3,pp.115‐122,2011.

[325] S.Dovis,S.VanderOord,R.W.Wiers,andP.J.Prins,"Canmotivationnormalizeworkingmemoryandtaskpersistenceinchildrenwithattention‐deficit/hyperactivitydisorder?Theeffectsofmoneyandcomputer‐gaming,"(ineng),JAbnormChildPsychol,vol.40,no.5,pp.669‐81,Jul2012.

[326] A.Sungetal.,"Fromthesmallscreentothebigworld:mobileappsforteachingreal‐worldfacerecognitiontochildrenwithautism,"AdvancedHealthCareTechnologies,vol.1,pp.37‐45,2015.

[327] K.M.Carrolletal.,"Motivationalinterviewingtoimprovetreatmentengagementandoutcomeinindividualsseekingtreatmentforsubstanceabuse:amultisiteeffectivenessstudy,"Drugandalcoholdependence,vol.81,no.3,pp.301‐312,2006.

[328] T.Susi,M.Johannesson,andP.Backlund,"Seriousgames:Anoverview,"ed:Institutionenförkommunikationochinformation,2007.

[329] R.M.Entman,"Framing:TowardClarificationofaFracturedParadigm,"JournalofCommunication,vol.43,no.4,pp.51‐58,1993.

[330] J.Juul,"TheOpenandtheClosed:Gameofemergenceandgamesofprogression,"inComputerGamesandDigitalCulturesConferenceProceedings,TampereUniversityPress,2002,pp.323‐329:Tampere.

[331] E.AdamsandJ.Dormans,GameMechanics:AdvancedGameDesign.NewRidersPublishing,2012,p.360.


185

[332] S.Göbel,V.Wendel,C.Ritter,andR.Steinmetz,"Personalized,AdaptiveDigitalEducationalGamesUsingNarrativeGame‐BasedLearningObjects,"Berlin,Heidelberg,2010,pp.438‐445:SpringerBerlinHeidelberg.

[333] G.Andrade,G.Ramalho,H.Santana,andV.Corruble,"Automaticcomputergamebalancing:areinforcementlearningapproach,"presentedattheProceedingsofthefourthinternationaljointconferenceonAutonomousagentsandmultiagentsystems,TheNetherlands,2005.

[334] K.vanderKooij,E.Hoogendoorn,R.Spijkerman,andV.Visch,"Validationofseriousgames,"InternationalJournalofSeriousGames,vol.2,no.3,2015.

[335] J.E.VanGemert‐Pijnen,S.M.Kelders,andE.T.Bohlmeijer,"UnderstandingtheUsageofContentinaMentalHealthInterventionforDepression:AnAnalysisofLogData,"(inEnglish),JMedInternetRes,OriginalPapervol.16,no.1,p.e27,2014.

[336] C.Granja,W.Janssen,andM.A.Johansen,"FactorsDeterminingtheSuccessandFailureofeHealthInterventions:SystematicReviewoftheLiterature,"(ineng),JournalofmedicalInternetresearch,vol.20,no.5,pp.e10235‐e10235,2018.

[337] R.Orji,G.F.Tondello,andL.E.Nacke,"Personalizingpersuasivestrategiesingamefulsystemstogamificationusertypes,"inProceedingsofthe2018CHIConferenceonHumanFactorsinComputingSystems,2018,p.435:ACM.

[338] A.CadamuroandV.Visch,"‘WhatRemains?’:APersuasiveStoryTellingGame,"inGamesforHealth,B.Schouten,S.Fedtke,T.Bekker,M.Schijven,andA.Gekker,Eds.:SpringerFachmedienWiesbaden,2013,pp.153‐160.

[339] G.Dunnetal.,"Effectiveelementsofcognitivebehaviourtherapyforpsychosis:resultsofanoveltypeofsubgroupanalysisbasedonprincipalstratification,"(ineng),Psychologicalmedicine,vol.42,no.5,pp.1057‐1068,2012.

[340] M.Y.NgandJ.R.Weisz,"Buildingascienceofpersonalizedinterventionforyouthmentalhealth,"Journalofchildpsychologyandpsychiatry,andallieddisciplines,vol.57,no.3,pp.216‐236,10/152016.

[341] A.Ghaderi,"Doesindividualizationmatter?Arandomizedtrialofstandardized(focused)versusindividualized(broad)cognitivebehaviortherapyforbulimianervosa,"BehaviourResearchandTherapy,vol.44,no.2,pp.273‐288,2006.

[342] K.Vernmark,Lenndin,J.,Bjärehed,J.,Carlsson,M.,Karlsson,J.,Öberg,J.,...&Andersson,G.,"Internetadministeredguidedself‐helpversusindividualizede‐mailtherapy:ArandomizedtrialoftwoversionsofCBTformajordepression,"Behaviourresearchandtherapy,vol.48,no.5,pp.368‐376,2010.

[343] J.L.Hudsonetal.,"Childinvolvement,alliance,andtherapistflexibility:processvariablesincognitive‐behaviouraltherapyforanxietydisordersinchildhood,"(ineng),BehavResTher,vol.52,pp.1‐8,Jan2014.

[344] S.E.Knowlesetal.,"Qualitativemeta‐synthesisofuserexperienceofcomputerisedtherapyfordepressionandanxiety,"(ineng),PloSone,vol.9,no.1,pp.e84323‐e84323,2014.

[345] W.TangandD.Kreindler,"SupportingHomeworkComplianceinCognitiveBehaviouralTherapy:EssentialFeaturesofMobileApps,"(ineng),JMIRMentHealth,vol.4,no.2,p.e20,Jun82017.

References

186

[346] K.D.Ersche,A.J.Turton,S.Pradhan,E.T.Bullmore,andT.W.Robbins,"DrugAddictionEndophenotypes:ImpulsiveVersusSensation‐SeekingPersonalityTraits,"BiologicalPsychiatry,vol.68,no.8,pp.770‐773,2010/10/15/2010.

[347] R.Orji,L.E.Nacke,andC.D.Marco,"TowardsPersonality‐drivenPersuasiveHealthGamesandGamifiedSystems,"presentedattheProceedingsofthe2017CHIConferenceonHumanFactorsinComputingSystems,Denver,Colorado,USA,2017.

[348] E.Murrayetal.,"EvaluatingDigitalHealthInterventions:KeyQuestionsandApproaches,"(ineng),AmJPrevMed,vol.51,no.5,pp.843‐851,Nov2016.

[349] C.W.LangfredandN.A.Moye,"Effectsoftaskautonomyonperformance:Anextendedmodelconsideringmotivational,informational,andstructuralmechanisms,"JournalofAppliedPsychology,89,pp.934‐945,2004.

[350] L.R.Martin,S.L.Williams,K.B.Haskard,andM.R.DiMatteo,"Thechallengeofpatientadherence,"TherapeuticsandClinicalRiskManagement,vol.1,no.3,pp.189‐199,09/2005.

[351] A.Loh,R.Leonhart,C.E.Wills,D.Simon,andM.Harter,"Theimpactofpatientparticipationonadherenceandclinicaloutcomeinprimarycareofdepression,"(ineng),PatientEducationandCounseling,vol.65,no.1,pp.69‐78,Jan2007.

[352] K.Ryan,S.Dockray,andC.Linehan,"AsystematicreviewoftailoredeHealthinterventionsforweightloss,"(ineng),Digitalhealth,vol.5,pp.2055207619826685‐2055207619826685,2019.

[353] A.J.Linn,M.Vervloet,L.vanDijk,E.G.Smit,andJ.C.VanWeert,"EffectsofeHealthInterventionsonMedicationAdherence:ASystematicReviewoftheLiterature,"(inEnglish),JMedInternetRes,Reviewvol.13,no.4,p.e103,2011.

[354] B.K.Campbell,J.Guydish,T.Le,E.A.Wells,andD.McCarty,"Therelationshipoftherapeuticallianceandtreatmentdeliveryfidelitywithtreatmentretentioninamultisitetrialoftwelve‐stepfacilitation,"(ineng),Psychologyofaddictivebehaviors:journaloftheSocietyofPsychologistsinAddictiveBehaviors,vol.29,no.1,pp.106‐113,2015.

[355] M.K.Nock,B.D.Michel,andV.I.Photos,"Single‐caseresearchdesigns,"Handbookofresearchmethodsinabnormalandclinicalpsychology,pp.337‐350,2007.

[356] H.Ossebaard,L.vanGemert‐Pijnen,A.deBruijn,andR.Geertsma,"Magnitudeofehealthtechnologyriskslargelyunknown,"IntJAdvSystMeas,vol.6,no.1,pp.57‐71,2013.

[357] C.Hallensleben,S.vanLuenen,E.Rolink,H.Ossebaard,andN.Chavannes,"eHealthforpeoplewithCOPDintheNetherlands:ascopingreview,"InternationalJournalofChronicObstructivePulmonaryDisease,vol.Volume14,pp.1681‐1690,07/012019.


187

188


189

SUMMARYThisdissertationfocusedontheaddedvalueofpersonalizedgamificationasafactortoenhanceimplementationpotentialofeHealthinterventionsinyouthmentalhealthcare.Mentalhealthdisordersaretheleadingcauseofdisabilityin adolescents. It is important for these adolescents to go into therapy, asadolescence is a period in live in which essential developments occur onwhichmentalhealthdisordershaveanegativeimpact.Althoughpsychosocialtherapiesareeffective inreducingpsychiatricsymptomsinadolescentswithmentaldisorders,thereisstillroomforimprovement.Forexample,becauseofpremature termination of treatment, poor attendance of treatment‐sessionsandalowornon‐adherencetohomeworkassignments.

One way of improving psychological treatment is the use of the use ofInformation and Communication Technologies combined with face‐to‐facetherapy (also called “blended eHealth”). It can extend the reach ofpsychologicaltherapybeyondtheclinicalsetting,astechnologiescanbeusedanytimeandanywhere. It isespeciallysuitable foradolescents,asamajorityownsasmartphone.

Current eHealth interventions inmentalhealthcareareoften focusedon thetherapeuticcontentandprovidelimitedinteractionmotivationfortheusers,causing a high drop‐out rate. Users of therapeutic eHealth should thus bemotivatedtostartandcontinuetousetheonlinemodulesfortherapy‐relatedactivities,especiallywhentheyhavetoperformtheseonlinemodulesintheirownenvironmentandtime.

Gamification seems a suitable design technique to enhance this motivationwithineHealth interventions. Itaims tochange thebehaviorofauser in therealworld (this change is also called the (health related) transfer effect) bycreatinga gameworldexperience that ismoreengaging, free andenjoyablecomparedtoarealworldexperience,byusinggame‐elementsinanon‐gamecontext. However, some game‐elements can bemoremotivating for specificindividualsthanothersandshouldthereforebepersonalized.

Summary

190

Wefirstconducteda literaturestudy that focusedonhowpersonalization isappliedingamedesignforhealthcareandhowthesegamesinfluencedhealthrelatedoutcomes(Chapter2).Thiswasfollowedbyafocusgroupstudythatfocused on the therapeutic practice of personalization in youth mentalhealthcare (Chapter 3) and experiments that focused on the game designrelevance of personalization in youthmental healthcare (Chapter4). Basedon this information, we designed an eHealth application for youth mentalhealthcare and implemented gamification and personalization in the designandexplainthisprocessinChapter5.Concluding,theaimofthestudyinthisdissertation was to study the added value of personalized gamification toenhance implementation potential of eHealth interventions in youthmentalhealthcare.

Sincepersonalizationingamificationhadneverbeensystematicallystudied,Iexecuted a literature study and developed a model to study the effects ofpersonalization in game design for healthcare (Chapter 2). Based on theliterature we proposed a model for different types of personalization ineHealth development and design. We defined ‘personalization’ as theinvolvement of stakeholders across ProblemDefinition, Product Design andTailoring(thePersonalizedDesignProcess (PDP)phases). In the firstphase,information is generated to identify, establish and analyze the problem andgenerate related ideas. In the next ProductDesign phase, possible solutionsareproduced,resultinginproductideasordesignproposal(s)thataretestedandevaluatedbyusers, and further improved through iterations. In the lastTailoring phase, the final product can be tailored to the needs of individualend‐users. The studies generally found positive effects on interactionexperience,interactionbehaviorandhealthrelatedtransfereffects.However,sinceamajorityof thestudieswereof lowmethodologicalquality,wecouldonly suggest that it is important to involve stakeholders across the PDP‐phases.Itwilllimittheamountofiterationsneeded,asthechanceisincreasedthattheeHealthinterventionisalignedtotheusers.Consequently, theuserswill potentially use the product to its full extend which will positivelyinfluencethehealthrelatedtransfereffect.


191

Personalization is not only often applied in game design for healthcare, butalso in a therapeutic process. Therapists and patients often adapt therapyprotocols,toalignittotheirpersonalpreferencesandsituation.IfdesignersofeHealth formentalhealthcaredonot take this into account it ismore likelythatthefinaleHealthdesigndoesnotsuitthetherapeuticpractice.Thiswillinturn negatively influence the implementation. We conducted focus‐groupdiscussionswith patients and therapists in youth addiction careon therapyprotocol application and personalization (Chapter 3) and generatedrecommendations for eHealth designers to enhance alignment of eHealth tothe therapeutic practice and implementation: a) study and copy at least theactualappliedpartsofatherapyprotocolineHealth,b)co‐designeHealthinsuchawaythatboththerapistsandpatientscanpersonalizespecificpartsofthefinaleHealthdesign,andc)investigateifpartsofthetherapyprotocolthatare not presently applied by therapists or patients should be part of theeHealthapplication.

Even when an eHealth product is aligned to therapeutic practice, it isimportant to enhance the motivation of patients to use eHealth and tofacilitate the achievement of aimed‐for real‐world goals such as behavioralchange. This can be done by making it more appealing by applyinggamificationdesign.WefirstusedadesignmethodwiththeoftenusedPLEXcards that represent 22 playful experiences that can motivate users to(continue)toplayagame.Wewantedtoexaminewhethertheinputofplayfulexperienceswasalsoexperiencedbyotherend‐users fromthesamecontextin theactualdesign itself(Chapter4A).However theexperiences thatwereused in thedesignof theprototypedidnotcorrespondone‐on‐onewith theexperiencesthatwerereportedbyotheruserswhoplayedtheprototype.Toensure that the product is still aligned to preferred experiences and limitpossible individual preferences of stakeholder types that cannot begeneralizable to the specific stakeholder type, it is important to involvestakeholders inmultiplemomentsandphasesofaPDP,andnotonly inone.Next to the specific designmethod,we alsowanted to study the effect of aspecificdesignelementinayouthmentalhealthcarecontext.Rewardsarethemost typically used game‐elements to foster motivation in entertainmentgaming.However, it isunclearwhethergame‐rewardsarealsoeffective ina

Summary

192

healthcare context. For example, patients with substance‐related disordersmay be less sensitive to non‐drug‐related rewards compared to patientswithoutasubstanceusedisorder.Resultsofourstudy(Chapter4B)showedthat, incontrast toourexpectations, substancedependentparticipantsweremore motivated by the types of rewards compared to non‐substancedependentparticipants.

ThepreviouschaptersprovidedargumentationforpersonalizationtoenhanceimplementationofgamifiedeHealth.Thiswastakenintoaccountinthedesignprocess of an eHealth application for youthmental healthcare (Chapter5).Wewantedtotestthegeneraleffectofthegamification,byconductinganon‐randomized pre‐post (eight weeks) study. In this study we contrasted twoconditions:oneeHealthinterventionthatwasgamifiedandonethatwasnotgamified.However,theinflowofpatientsintheLucastudywaslow,andatthetime of writing it was uncertain whether the study could be finalized asplannedaccordingtoourprotocol.Themainreasonwasthatthestudyset‐upwasnotinlinewiththecurrenttherapeuticpractice,whichwasareasonfortherapistsandpatientstonotbefullywillingtoparticipateintheexperiment.Thus, evenwhen eHealth is personalized and gamified, implementation canstill be influenced by negative expectations about the effect, a limitedintegrationwithincurrenttherapy.

Concluding, when stakeholders are more actively involved in the designphases of a gamification, the motivational effect of the gamification can beenhanced. It is important to align an eHealth product to the context ofapplicationand toalign thedesign to thepreferredexperience,capacityandcontextofauser,toenhancetheimplementationpotential.Secondly,theset‐up of effect‐studies should be adapted to the context of application to limitinvasivenessintherapeuticpracticeandtoenhancethefeasibilityofthestudy.Ifastudyset‐upisnotinlinewiththecurrenttherapeuticpractice,therapistsandpatientsmaythusnotbefullywillingtoparticipateinanexperiment.Thiswould make it difficult or impossible to test the effect of personalizedgamificationineHealth,whichisusefulinformationforfutureeHealthdesignsandstudiestoenhanceimplementationpotential.


193

194


195

SAMENVATTINGDitproefschriftrichttezichopdetoegevoegdewaardevangepersonaliseerdegamificatiealseenfactoromdeimplementatievaneHealthinterventiesindejeugd geestelijke gezondheidszorg (jeugd‐GGZ) te vergroten. Psychischestoornissen zijn de grootste oorzaak vanbelemmeringen in de adolescentie.Voordezeadolescentenishetbelangrijkominbehandelingtegaan,omdatdeadolescentieeenperiodeinhetleveniswaarbinnenessentiëleontwikkelingenplaatsvindenwaar psychische stoornissen een negatieve impact op hebben.Ondanks effectieve psychosociale behandelingen om symptomen bijadolescenten met psychische stoornissen te verminderen, is er nog steedsruimtevoorverbetering.Bijvoorbeeld,omdatveelbehandelingenvroegtijdigstoppen, men vaak niet aanwezig is op behandelsessies, en behandelopdrachtenvaaknietofnietvolledigwordenuitgevoerd.

Eenmogelijkheidompsychologischebehandelingteverbeteren,ishetgebruikvan Informatie en Communicatie Technologieën in combinatie met face‐to‐facetherapie(ook“blendedbehandeling”genoemd).Hetkanzorgenvooreengroterbereikvande face‐to‐facebehandelingdanalleenbinnendeklinischesetting,omdatdergelijketechnologieënaltijdenoveralgebruiktkanworden.Daarbijishetvooralgeschiktvooradolescenten,omdateenmeerderheidvanheninhetbezitisvaneensmartphone.

HuidigeeHealthinterventiesindeGGZzijnvaakgerichtopdetherapeutischeinhoudenzorgenvooreenbeperkteinteractiemotivatievoordegebruikers,watvoorveeluitval ingebruikzorgt.GebruikersvantherapeutischeeHealthmoeten dus gemotiveerd worden om de online modules te gebruiken enblijven tegebruikenvoor therapeutische‐gerelateerdeactiviteiten, vooralalsditmoetgebeureninhuneigenomgevingentijd.

Gamificatie lijkt een gepaste ontwerptechniek om deze motivatie binneneHealth interventies te verhogen. Het heeft als doel het gedrag van degebruikerinzijnofhaardagelijkseleventeveranderen,dooreenspel‐wereld‐ervaringtecreërendieaantrekkelijker,vrijer,enleukerisdandeervaringenin de echte wereld, door het gebruik van spelelementen in een niet‐spelcontext. Maar sommige spelelementen kunnen motiverender zijn voor

Samenvatting

196

specifieke individuen dan anderen en zouden daarom gepersonaliseerdmoetenworden.

Ikhebeersteen literatuuronderzoekuitgevoerd,waarbijgekekenwerdhoepersonalisatieistoegepastingameontwerpvoordegezondheidszorgenhoedezegamesgezondheids‐gerelateerdeuitkomstenbeïnvloedde(Hoofstuk2).Dit werd vervolgd door een focusgroep onderzoek, dat zich richtte op detherapeutische praktijk van personalisatie in de GGZ voor adolescenten(Hoofdstuk 3) en experimenten die zich richtten op de game ontwerprelevantie van personalisatie in de GGZ voor adolescenten (Hoofdstuk 4).Gebaseerdopdeze informatie,hebbenweeeneHealthapplicatieontwikkeldvoor de jeugd‐GGZ, waarbij gamificatie en personalisatie in het ontwerpbetrokkenwerd. Dit proceswordt uitgelegd inHoofdstuk5. Concluderend,was het doel van de onderzoeken in deze dissertatie om de toegevoegdewaarde van gepersonaliseerde gamificatie te onderzoeken om deimplementatiepotentieelvaneHealthindejeugd‐GGZteverhogen.

Omdatpersonalisatieingamificatienooitsystematischonderzochtwas,werder een literatuur onderzoek uitgevoerd en een model ontwikkeld om deeffecten van personalisatie in game ontwerp voor gezondheidszorg teonderzoeken(Hoofstuk2).Wedefinieerden‘personalisatie’alshetbetrekkenvanstakeholdersindeProbleemDefinitie,ProductOntwerp,enTailoring(dePersonalizedDesignProces(PDP)‐fases).Indeeerstfase,wordterinformatiegegenereerdomhetprobleemteidentificerenenteanalyserenendoorideeënte genereren. In de volgende product ontwerp fase worden er mogelijkeoplossingen geproduceerd, welke resulteren in product ideeën of ontwerpvoorstellen die getest en geëvalueerdwordendoor de gebruikers en verderontwikkeld worden door iteraties. In de laatste Tailoring fase, kan hetuiteindelijkeproductaangepastwordennaardebehoeftesvandeindividuelegebruikers. De onderzoeken vonden in het algemeen positieve effecten opervaringen, interactie gedrag en gezondheid‐gerelateerde effecten. Echter,omdat een meerderheid van de onderzoeken een lage methodologischekwaliteit had, kunnen we alleen suggereren dat het belangrijk is omstakeholders in de PDP‐fases te betrekken. Dit zal naar verwachting dehoeveelheid iteratiesdie nodig zijnbeperkenomdatde eHealth interventies


197

naarverwachtingmeerin lijn liggenmetdegebruikers.Alsgevolg,zullendegebruikers het product volledig benutten wat weer een positief effect zalhebbenopgezondheid‐gerelateerdeuitkomstendiemenwilbeïnvloeden.

Voor ontwerpers van eHealth is het belangrijk om te weten wat ergepersonaliseerd is in de therapeutische praktijk, omdat het product moetpassen in de context waarin het gebruikt gaat worden. Behandelaren encliëntenpersonaliseren vaak therapieprotocollen, zodatdezepassen in hunpersoonlijkevoorkeurenensituaties.AlsontwerpersvaneHealthvoordeGGZdit niet meenemen, is het waarschijnlijker dat het uiteindelijke eHealthproduct niet past binnen de therapeutische praktijk. Dit heeft weer eennegatief effect op de implementatie en ook op het uiteindelijk gebruik. Wijvoerdenfocus‐groepdiscussiesmetbehandelarenencliëntenindejeugd‐GGZover protocol toepassing enpersonalisatie (Hoofdstuk3).Opbasis hiervankwamenwemetaanbevelingenvooreHealthontwerpersomzoeHealthmeerop een lijn te brengen met de therapeutische praktijk en daarmee deimplementatie kans te verhogen: a) onderzoek en kopieer in ieder geval detoegepaste onderdelen van een behandelprotocol in eHealth, b) co‐designeHealth op zo’n manier dat zowel behandelaren en cliënten specifiekeonderdelenvanhetuiteindelijkeeHealthontwerpkunnenpersonaliseren,enc)onderzoekoferdelenvanhettherapieprotocolnunietwordentoegepastdoor behandelaren en cliëntenmaar diewel onderdeel zoudenmoeten zijnvandeeHealthapplicatie.

ZelfsalseeneHealthproductaangepastisnaardetherapeutischepraktijk,ishet belangrijk om de motivatie van cliënten om eHealth te gebruiken tevergroten.Ditkangedaanwordendoorhetontwerpaantrekkelijkertemaken,doorhettoepassenvangamificatie.Wehebbeneersteenspecifiekeontwerpmethode,degebruiktePLEXkaarten, onderzocht.PLEXkaartenbevatten22speelseervaringen,diegebruikerskunnenmotiverenomeenspelte(blijven)spelen.Hierbijhebbenweonderzochtofdeinputvanspeelseervaringenookervaren werd door andere eindgebruikers van dezelfde context in hetuiteindelijke ontwerp zelf (Hoofstuk 4A). Resultaten lieten zien dat deervaringen die gebruikt werden voor het ontwerp van de prototypen nietovereenkwamenmetde ervaringendie gerapporteerdwerdendoor andere

Samenvatting

198

gebruikersdiehetprototypespeelden.Omteverzekerendathetproductnogsteeds in lijn ligt met specifieke ervaringen, en mogelijke individuelevoorkeuren van stakeholder types die niet te veralgemenen zijn naar degrotere groep te beperken, is het belangrijk om stakeholders op meerderemomenten en fases van een PDP te betrekken, en niet alleen in een fase ofmoment.Naastdespecifiekeontwerpmethode,wildenweookeenspecifiekontwerp element onderzoeken in de jeugd‐GGZ. Beloningen zijn de meestgebruikte game‐elementen om motivatie te bevorderen in entertainmentgames.Maarhetkanookzozijndatspelelementennietgeschiktzijnvoorallecontexten. Bijvoorbeeld, cliënten met middelen‐gerelateerde stoornissenkunnen minder gevoelig zijn voor niet‐drugs‐gerelateerde beloningen (demeest toegepaste motivator) dan cliënten zonder een middelen stoornis.Resultatenvanonsonderzoek(Hoofstuk4B) lietenziendat, in tegenstelingtot onze verwachtingen, middelen‐afhankelijke deelnemers niet mindergemotiveerd waren door de beloningen dan niet‐middelen afhankelijkedeelnemersenzelfsmeergemotiveerdwarendoorgeldalsbeloning.

Devoorgaandehoofdstukkenzorgdenvoorargumentatievoordetoepassingvan personalisatie om implementatie van gegamificeerde eHealth tevergroten. Dit werd meegenomen in het ontwerp proces van een eHealthapplicatie – de Luca‐app – voor de jeugd‐verslavingszorg (Hoofstuk5).Wewilden het algemene effect van gamificatie onderzoeken, door een niet‐gerandomiseerd onderzoek (acht weken) met een voor en nameting uit tevoeren. In dit onderzoek contrasteerden we twee condities: een eHealthinterventie die gegamificeerd was en een die niet gegamificeerd was. DeinstroomvanpatiënteninhetLucaonderzoekwasechterlaag,ententijdevandit schrijvenwashetonzekerofhetonderzoekvolgendedeplanning inonsprotocol afgerond zou kunnen worden. De voornaamste reden was dat deopzetvanhetonderzoeknietinlijnlagmetdehuidigetherapeutischepraktijk,wateenredenvoorbehandelarenencliëntenwasomniet (volledig)deel tenemen aan het experiment. Dus zelfs als eHealth gepersonaliseerd engegamificeerd is, kan implementatie nog steeds beïnvloed worden doornegatieveverwachtingenoverheteffecteneenbeperkteintegratiebinnendehuidigetherapie.


199

Concluderend, als stakeholders meer actief betrokken zijn in het ontwerpproces van een gegamificeerde interventie, kan het motiverende effect vangamificatievergrootworden.HetisbelangrijkomeeneHealthproductinlijnte brengen met de context van toepassing en met de geprefereerdeervaringen, capaciteit en context van een gebruiker om implementatie tevergroten.Daarnaastmoetdeopzetvanonderzoekenaangepastwordennaardecontextvantoepassingomdestoringbinnendetherapeutischepraktijktebeperkenenhaalbaarheidvanhetonderzoek te vergroten.Alsdeopzet vaneen onderzoek niet in lijn ligt met de therapeutische praktijk, zullenbehandelaren en cliënten niet (volledig) willen meewerken aan eenexperiment wat het moeilijk of onmogelijk maakt om het effect vangepersonaliseerdegamificatiebinneneHealthteonderzoeken.Ditisbruikbareinformatie voor toekomstige eHealth ontwerpen en onderzoeken omimplementatietevergroten.

200

Documents

Delft University of Technology Personalized gamification to … · 2020-03-11 · Personalized gamification to enhance implementation of eHealth therapy in youth mental healthcare