RESEARCH Open Access

How are evidence and knowledge used inorthopaedic decision-making? Threecomparative case studies of differentapproaches to implementation of clinicalguidance in practiceAmy Grove1* , Aileen Clarke1 and Graeme Currie2

Abstract

Background: The uptake and use of clinical guidelines is often insufficient to change clinical behaviour and reducevariation in practice. As a consequence of diverse organisational contexts, the simple provision of guidelines cannotensure fidelity or guarantee their use when making decisions. Implementation research in surgery has focused onunderstanding what evidence exists for clinical practice decisions but limits understanding to the technical,educational and accessibility issues. This research aims to identify where, when and how evidence and knowledgeare used in orthopaedic decision-making and how variation in these factors contributes to different approaches toimplementation of clinical guidance in practice.

Methods: We used in-depth case studies to examine guideline implementation in real-life surgical practice. Weconducted comparative case studies in three English National Health Service hospitals over a 12-month period.Each in-depth case study consisted of a mix of qualitative methods including interviews, observations anddocument analysis. Data included field notes from observations of day-to-day practice, 64 interviews with NHSsurgeons and staff and the collection of 121 supplementary documents.

Results: Case studies identified 17 sources of knowledge and evidence which influenced clinical decisions inelective orthopaedic surgery. A comparative analysis across cases revealed that each hospital had distinctapproaches to decision-making. Decision-making is described as occurring as a result of how 17 types ofknowledge and evidence were privileged and of how they interacted and changed in context. Guidelineimplementation was contingent and mediated through four distinct contextual levels. Implementation could beassessed for individual surgeons, groups of surgeons or the organisation as a whole, but it could also differbetween these levels. Differences in how evidence and knowledge were used contributed to variations in practicefrom guidelines.

(Continued on next page)

* Correspondence: A.L.Grove@warwick.ac.uk1Division of Health Sciences, Warwick Medical School, University of Warwick,Coventry, UKFull list of author information is available at the end of the article

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Grove et al. Implementation Science (2018) 13:75 https://doi.org/10.1186/s13012-018-0771-4

(Continued from previous page)

Conclusion: A range of complex and competing sources of evidence and knowledge exists which influence theworking practices of healthcare professionals. The dynamic selection, combination and use of each type ofknowledge and evidence influence the implementation and use of clinical guidance in practice. Clinical guidelinesare a fundamental part of practice, but represent only one type of evidence influencing clinical decisions. In theorthopaedic speciality, other distinct sources of evidence and knowledge are selected and used which impact onhow guidelines are implemented. New approaches to guideline implementation need to appreciate andincorporate this diverse range of knowledge and evidence which influences clinical decisions and to take accountof the changing contexts in which decisions are made.

Keywords: Guidelines, Implementation, Orthopaedic surgery, Evidence-based medicine, Comparative case study

BackgroundAcross the world, policymaking organisations exist to pro-duce clinical guidance and recommendations for health-care, which are based on scientific evidence. In the UK,these organisations are the National Institute for Healthand Care Excellence (NICE) and the Scottish Intercollegi-ate Guidelines Network (SIGN) [1, 2]. Together, they pro-vide evidence-based guidance on the most effective waysto diagnose, treat and prevent poor health for NationalHealth Service (NHS) patients. One of the aims of clinicalguidance is to reduce variation in practice, and therefore,limit inequalities in service delivery [3]. However, previousresearch has established that healthcare organisations faceseveral challenges to implementation [4–6].In 2014, a review of systematic reviews explored factors

which influence guideline implementation and uptake [7].The findings report a range of factors including those as-sociated with guidelines themselves (e.g., complexity), withhealthcare professionals (e.g., lack of awareness) and withthe working environment such as limited time, personneland resources devoted to support guideline adherence [7].Limited information was provided to describe the contextof guideline implementation in detail or the differences inhealthcare environments which might impact implemen-tation, such as differences in the processes of healthcaredelivery and national policy decisions. Evidence existsdescribing the barriers to implementation of clinicalguidance and the rates of guideline uptake across arange of diseases [3–5, 8, 9]. For example, lack of timewas identified as the most commonly reported barrierto implementation [5]. However, too few studies rigor-ously assess the effectiveness of approaches to improveimplementation of clinical guidelines or explicitly de-scribe factors which enable their uptake and use [6, 7].Rates of guideline uptake provide a proxy for guidelineimplementation, but fail to demonstrate the realisticuptake and actual use of clinical guidelines in realworld healthcare practice.There is a need for empirical research which examines a

wide range of contextual characteristics influencing theuptake and use of clinical guidelines. This includes the

professional and environmental factors which have beendescribed previously, but also requires a description of thelocal practice context and wider healthcare sector insufficient detail to facilitate more effective guidelineimplementation. Therefore, the aim of this research wasto identify where, when and how evidence and knowledgeare used in healthcare decision-making and how variationin these factors contributes to different approaches toimplementation of clinical guidance in practice.

Over and above clinical guidelinesAchieving effective guideline implementation reflectsnot only the people involved, but also their professionalroles, their positions in the organisation and the epi-stemic communities to which they belong [10, 11].Producers and users of evidence and guidelines often siton different sides of the social, scientific and clinicalboundaries. Therefore, they possess varying types ofknowledge which may mean that they privilege evidencedifferently. This can make integration across these com-munities challenging [11]. Day-to-day decision-makingby healthcare professionals requires the selection of manytypes of knowledge and evidence, situated within local,contextual and social circumstances [12]. For example,research knowledge coexists with the lived experience ofpatients and the macro healthcare initiatives and incen-tives from policymakers and regulatory bodies.Preserving clinical autonomy and medical judgement by

these professional groups is also recognised as important,particularly when guidelines challenge traditional practices[13, 14]. In this study, we examine guideline implementa-tion in the context of real-life NHS practice. We investigateguideline implementation problems through the applicationof a range of qualitative methods to explore where, whenand how evidence and knowledge were used in clinicaldecision-making in elective orthopaedic surgery.

Guideline implementation in orthopaedic surgeryA mixed methods systematic review has identified vari-ous sources of evidence and knowledge which influencedecision-making within orthopaedic surgery [15]. The

Grove et al. Implementation Science (2018) 13:75 Page 2 of 14

findings revealed several factors which impact uponguideline implementation specific to the orthopaedic spe-cialty. For example, compared to other clinical specialties,orthopaedic surgery represents a highly professionalisedarea of clinical work where elite communities of practiceare strongly embedded [16]. There is a tendency towardsdecisive and authoritative patterns of decision-making,and surgeons are able to retain substantial autonomy overtheir work practices to resist external intervention [15].Enhanced implementation of guidelines is an unex-

plored area of research in the highly professionalised,intensely networked group that is orthopaedic surgery.This study provides an in-depth understanding ofdecision-making using clinical guidance. It generates aunique perspective on the challenges of translatingresearch into clinical practice and moves away from theapproach of assessing compliance or auditing guidelineuptake. A key part of understanding complex problems,such as where, when and how evidence and knowledgeare used in practice, requires examining the values, beliefsand norms of individuals who are responsible for makingdecisions in context. This complements a broader investi-gation of an organisation’s capacity to support clinicalguideline implementation and of other contextual factorswithin the healthcare sector which influence the use ofknowledge and evidence in practice.

MethodsWe examined the implementation of an example of NICEguidance within orthopaedic surgery, to identify similar-ities and differences in the way this type of evidence isused in practice. In 2014, NICE released updated guidanceon hip implants for total hip replacement for end-stagearthritis (see Additional file 1: Appendix SP1 for anexample of the guideline recommendations) [17]. At thetime of our study, it had been over a decade since the

previous version of the guidance was disseminated topractitioners. Therefore, our study was ideally timed toinvestigate the implementation of this updated guidancein the NHS to facilitate more general exploration ofguidelines.Three comparative case studies were conducted, using

qualitative methods with multiple levels of analysis [18].The case studies were conducted in UK hospital Trustsin the midlands, north and south west of England (seeTable 1). A hospital Trust is an organisation that pro-vides secondary healthcare services to a locality withinthe English NHS system. The protocol for the entirestudy has been described in detail elsewhere [19]. Eachcase study traced the implementation of NICE guidancein practice to explore the understanding and use of evi-dence and knowledge in orthopaedic surgery. We selectedcases to represent maximal variation in orthopaedic ser-vices in England [20]. For example, an orthopaedic depart-ment in a teaching hospital, one in a non-teachinghospital and a third in a designated academic orthopaedicdepartment where staff members hold hybrid academic/clinical roles in both the hospital and affiliated University.We followed the roadmap for a case study research de-

veloped by Eisenhardt [21] where each case study startedas close as possible to the ideal of no theory under consid-eration. This method prevents any pre-selected theoreticalperspectives limiting the data collection process [21]. Theresearch was abductive in nature. We were informedby previous literature and theory but also by the datacollected [22].Data collection in the field allowed for concepts of

interest to develop as the case studies progressed [23].This flexible approach is a key feature of case study de-signs, which enabled us to adjust data collection processesto further investigate emergent themes and to take advan-tage of opportunities as they arose [24, 25]. For example,

Table 1 Case study setting and participants

Descriptor Case study A Case study B Case study C

Setting Orthopaedic trauma centre Small hospital Trust split betweentwo geographical sites. Thereforethe orthopaedic services wereseparated across two hospitalbuildings

Large orthopaedic departmentwith specialist trauma centrewhich received national referralsfor complex hip implant revisionsurgery

University link Teaching hospital with a designatedacademic orthopaedic departmentlocated in a university owned buildingwithin the NHS hospital

None Teaching hospital

Participants A majority of surgeons held jointposts between the NHS and thesame university department. Thestaff conducted clinical effectivenessand cost-effectiveness research,mainly national randomisedcontrolled trials of varioustechniques and treatmentswithin orthopaedic surgery

Surgeons provided generalorthopaedic services to thelocal population supportedby a designated group ofallied health professionals

Surgeons in the teaching hospitalheld contracts with the NHS hospital.A minority of surgeons held honorarycontracts with one of the four universitiesin the region, i.e. they were not fromthe same academic department

Grove et al. Implementation Science (2018) 13:75 Page 3 of 14

the importance of groups of surgical colleagues acting ascommunities of practice (as a potential theme) grew, asmore data was collected and as case studies progressed inseries. This enabled us to search for specific instances ofsurgeons in communities in the later cases and thusformed part of the data collection process.Across the three cases, we sampled orthopaedic surgeons

and Allied Health Professionals (AHPs) who conducted orfacilitated joint replacement surgery, i.e., we sampled pur-posively aiming for heterogeneity of professional back-ground, level of training and years in practice. Snowballsampling enabled us to follow direct recommendationsfrom participants. We aimed to explore guideline im-plementation from all perspectives, so we also invitedadministrators and managers involved in guideline im-plementation and in the decisions made for patientsundergoing hip replacement surgery.

Data collectionWe selected a combination of qualitative methods in-cluding document analysis, observation and interviews.One author (AG) was responsible for all data collectionin each of the cases. Data collection remained systematicand transparent, and decisions were recorded in casesummaries. We continued data collection until no newinformation was obtained and theoretical saturationwithin each case was reached [26]. Within each case,3 months of observation took place between 1 December2014 and 11 December 2015. Observations consisted ofopportunistic shadowing involving, watching clinic andteaching sessions and attendance at planned operatingsessions, particularly pre-theatre preparation time. Obser-vations enabled informal discussions with surgeons andclinical staff and provided an opportunity to describeactions and decisions in real time. Each observation wasrecorded in a field journal using a predeterminedtemplate.Document analysis involved collection of key organisa-

tional documents such as clinical pathways describingstructured multidisciplinary plan of care and hospitalprotocols (see Table 2) [27]. The documents helped us

to understand and frame intentions to change practicewithin the orthopaedic departments. Analysis of the doc-uments enabled us to gain a wider understanding of thecontext within which decisions were made.We interviewed 64 participants between December

2014 and December 2015. During the interviews, wesought to understand the approaches and beliefs of par-ticipants regarding knowledge and evidence, in order toreveal the strategies used by professionals when makingdecisions. Questions explored the extent of professionals’beliefs regarding NICE and the involvement and impact ofclinical guidance on surgical practice within their hospital.The open interview format enabled participants to expandon topics of interest freely. We set out to discover whatprofessionals considered to be evidence and knowledge inpractice, rather than focusing on any pre-existing defin-ition which may have restricted the findings. Each inter-view was labelled with location and timing and ananonymised identification number.A copy of the interview topic guide is presented in the

Additional file 1: Appendix SP2. Table 3 displays thedifferent professional groups interviewed (‘C’ clinical, ‘A’allied health professionals and ‘M’ managers). To obtain anational perspective, we conducted eight key informantinterviews with stakeholders from NICE, The RoyalCollege of Surgeons, and Clinical Commissioning Groups.

Data analysisAs outlined in the roadmap method, all data were ana-lysed, integrated and triangulated within case beforecomparative case analysis was undertaken across thecases. The three data sources were processed into textformat to allow for thematic analysis through data famil-iarisation, coding and development of categories fromcodes [28]. The first stage of data analysis was conductedby one author (AG). Second round coding was per-formed jointly by all authors during two data analysissessions. Coding differences were reconciled throughdiscussion by all authors and refinement of first- andsecond-order codes was performed to generate categor-ies and themes. The three types of qualitative data wereintegrated using the Pillar Integration Process [29]. Thisis a matrix integration technique for mixing data whichhas been collected using different methods.

Table 2 Document type and quantity by case study site

Document type Case A Case B Case C

Clinical pathways 5 3 6

Protocols 17 2 4

Meeting notes 7 5 11

Strategy documents 2 1 0

Quarterly and annual reports 14 18 17

Internal presentations 2 5 2

Sub-total 47 34 40

Total 121

Table 3 Participant numbers detailed by case study site and byprofessional group

Professional group Case A Case B Case C Key informant interviews

Clinical (C) 12 10 8 4

AHP (A) 4 5 6 2

Managers (M) 2 4 5 2

Sub-total 18 19 19 8

Total 64

Grove et al. Implementation Science (2018) 13:75 Page 4 of 14

We compared data across data collection methods.Triangulation between data collection methods facili-tated the validation of cases as we searched forconvergence among the multiple data sources. We tri-angulated data from the three data sources and inter-preted them together to find common themes byeliminating overlapping areas and identifying areas of con-vergence [26]. We noted, for example, if a ‘guidelineimplementation process’ document in a hospital didnot match with the data collected in the observationsand interviews, this perhaps demonstrates that partici-pants were not aware of this document or that it wasnot an important factor into decision-making. Eachcase was written up to provide narrative descriptionsof the current situation at each hospital. This processwas central to generating familiarity and insight [30].It enabled us to see patterns in each case as theyemerged and accelerated our cross case comparisons[21]. The goal of the cross case comparison was tosearch for further patterns in the data and to explorehow these were represented or played out differentlyin the three cases.

ResultsThe overarching themes displayed in Fig. 1 representbroader narratives to describe the structural levelswhich influenced guideline implementation in electiveorthopaedic practice. The evidence and knowledge ofindividual surgeons, groups of healthcare professionals,healthcare organisations and the regulatory environ-ment interacted to produce the context for guidelineimplementation.

Types of knowledge and evidence which influence theimplementation of clinical guidance in electiveorthopaedic surgeryCase studies revealed that a significant number and a di-verse range of knowledge and evidence sources wereused in decisions made by orthopaedic surgeons regard-ing hip replacement surgery. We characterised thesesources into micro, meso or macro levels of influence, asdisplayed in Fig. 2. The sources of evidence and know-ledge have been categorised this way to demonstrate thestructural level at which they were enacted in practice[31]. The multi-level approach to synthesis helped torecognise the interdependence between the variouslevels. Micro knowledge and evidence that tended toinfluence individual decision makers, meso evidence andknowledge sources appeared to act at the level of theorganisation, whereas macro knowledge and evidenceexisted in the higher domain of the wider healthcareenvironment. Narrative descriptions and an example ofeach source of evidence and knowledge are providedin Table 4.Evidence in the form of NICE guidance, i.e., formal co-

dified knowledge, was just one of the 17 types of know-ledge and evidence identified in our study. Therefore,the additional 16 sources of evidence, such as the struc-ture and location of the hospital or the opinions ofleaders and professional societies, influenced the uptakeand use of clinical guidelines in orthopaedic surgery.A determining factor of guideline implementation

was how these knowledge and evidence types wereamalgamated together in the different contexts of prac-tice. The amalgamation process was flexible, adaptable

Fig. 1 Visual representation of the four thematic findings which describe the influence of evidence and knowledge on decision-making

Grove et al. Implementation Science (2018) 13:75 Page 5 of 14

and on-going, and therefore, the dominant source ofevidence and knowledge in each case would oftenchange. What was important to the decision-maker atone point in time was not always the presence or con-tent of evidence-based guidelines (macro). It couldequally be any other type of evidence or knowledge,interacting with, for example, a surgeon’s training andformal education (micro) or the contingencies of prac-tice such as the structure of the hospital (meso).

Distinct approaches to decision-makingComparison across cases was essential to look beyondthe initial impressions and see the findings throughmultiple lenses. We framed our cases according totheir general approach to decision-making and howguidelines were implemented in each hospital. Thedifferences in decision-making contributed to vari-ation from the codified evidence contained in guide-lines across all three cases. The analysis revealed thatthe individuals, groups and organisations in each casehad a distinct approach to decision-making. The approachwas dependent on how the 17 types of knowledge and evi-dence interacted, changed and were used in orthopaedicpractice. Guideline implementation was contingentand mediated through the distinct contextual environ-ments, which were subject to forces of the regulatoryenvironment.

Case ACase A was an academic centre located within a traumaand orthopaedic department. More than the other twocases, case A appeared to have a positive view of the for-mal codified knowledge contained in clinical guidelinesand what guidelines set out to achieve in the healthcaresector. The surgeons working in case A took a popula-tion perspective on clinical decisions. The culture, normsand political influence of the sector acted on case A in apositive fashion, as surgeons working here valued theprocesses and aims of generating clinical guidelines andthe goals of NICE as an organisation. Surgeons were ac-customed to answering questions using a larger popula-tion frame of reference and suggested that they “maysubconsciously be following NICE guidance” (Junior sur-geon) as it was indoctrinated in the organisational know-ledge and processes. One surgeon states that “NICEguidelines help you look at cost effectiveness and evi-dence a lot more than you would think about in normaldaily practice as an individual orthopaedic surgeon”(Consultant surgeon).Case A had the most advanced and formalised guide-

line implementation processes i.e., their managerialknowledge. This was reflected in case A’s extensiveprotocol documentation compared to the other twocases (17 versus 4 and 2). Each protocol was linked to apiece of clinical guidance or an internal evidence

Fig. 2 Summary of the of knowledge and evidence identified in case studies of guideline implementation in orthopaedic practice

Grove et al. Implementation Science (2018) 13:75 Page 6 of 14

Table 4 Narrative descriptions and an example of each source of evidence and knowledge depicted in Fig. 2

Evidence and knowledge types Narrative description Example from the data

Macro

External evidence created by healthcareregulators, e.g. CQC and GMC

The wider delivery of healthcare in Englandis governed by the UK Health and SocialCare Regulators such as the Care QualityCommission. In orthopaedics, surgeonshave to be registered with the GeneralMedical Council and with their RoyalCollege. Regulators are responsible forensuring that surgeons are included inan up-to-date registry of qualified doctorsand practice according to establishedstandards

An inspection report from theCare Quality Commission

The media and the influence of‘the press’

The mass media (or press) is a diversifiedcollection of resources who reach a largeaudience via mass communication

An article in a newspaper describing‘good’ or ‘bad’ hip implants

The opinion of leaders andprofessional societies

An opinion leader was an eminent individualwho had the ability to influence the opinionof the orthopaedic community on a subjectmatter for which they well known. Theprofessional societies were larger organisationswho represented the groups and sub-groupsof surgeons

An opinion leader could be a principalinvestigator of a large clinical trial inorthopaedics. The professional societieswere the British Orthopaedic Associationand the Hip Society

Formal codified knowledge Evidence or knowledge that is writtendown can be shared and is easy to accessand available to the public

A NICE guideline or article published ina journal

Culture, norms and political influenceof the sector

The standards and accepted way of practicingin the UK healthcare context. Including thepublic delivery of services and formal andinformal methods in which healthcare isorganised in the NHS

The hierarchical structure of the healthcaresystem. Political factors included strategiesenforced by government and the medico-legal challenges to practice

Meso

Managerial knowledge Each hospitals’ business organisationalprocesses which underpin day to dayroutines and capabilities of the Trust

NHS hospital resource issues such as time,cost and safety or quality of services

Organisational knowledge An extension of managerial knowledgewhich has a wider structural emphasis. Itis embedded in the processes of healthcareorganisations and influences the behaviourof staff

A hospitals’ internal processes which arenot written down. Anecdotally referredto as “the way we do things around here”

The structure and location of the hospital The physical location of the hospital buildingsand departments and the structure of thehospital wards

The number of elective orthopaedictheatres available to use

Evidence from implant manufacturingcompanies

Information that came directly frommanufacturer’s representatives located in the hospital or indirectly through marketing

Leaflets about a hip implant from amanufacturer’s representative

Socialisation and association with colleagues Knowledge that came from the inside andspread within the defined clinical group, inthis case the orthopaedic community

Evidence of the outcome of a surgeryfrom a colleague or knowledge that amentor had passed on

Micro

Informal experiential knowledge Tacit knowledge that surgeons ‘know’regarding how to behave and perform asan orthopaedic surgeon

Represents a surgeon’s lifetime’s work,and in turn their identity as a surgeon

Informal experiential knowledge builtup over time

The tacit knowledge that surgeons ‘know’which has built up over time working in thespecific hospital but which can be difficultto describe

Knowing which colleague to refer adifficult case to when the surgeondoes not have the specific expertiseor experience

Evidence from the professional hierarchy The layered social structure within thehospital which conceptualised thesuperior and inferior relationshipsbetween clinical staff

Described as the ‘clinical peckingorder’ with the consultant surgeonat the top

Grove et al. Implementation Science (2018) 13:75 Page 7 of 14

summary which had been produced by medical librar-ians. The implementation process was described by theChair of the Guideline Committee:

We started monitoring NICE implementation in thetrust in 2011/12 when we set up the NICEimplementation group…when NICE issues theirguidance, I forward a list to the NICE administratorwho would send it out to lead clinician inorthopaedics…If it’s a clinical guideline, or qualitystandard…there’s recommendations in there, and aform of a baseline assessment with therecommendations in…they (the clinician) have toindicate if we’re compliant or not. (Hybrid surgeon)

The surgeons in case A demonstrated a distinctlydifferent trend in their confidence and appreciation ofclinical guidance (individual surgeon factors). We con-sider that this was due to their departmental focus onresearch (apprenticeship style training) and academicoutput developed through training and formal educationin Evidence Based Medicine (EBM), and beliefs regard-ing the importance of NICE and the EBM approach. Aquote from a surgeon illustrates this:

“We do try as much as possible to follow the basis ofevidence...So the typical patients would be where wewould need to use evidence. There is not much of adilemma about someone with end stage osteoarthritisas the guidance shows the pathway that they shouldtake.” (Consultant surgeon)

The distinct approach to decision-making in case Awas ‘pragmatic EBM decision-making’, where the trad-itional approach to evidence-based decision-making rec-ommended in clinical guidelines was suitable, but notsufficient for their practice. Surgeons in case A acknow-ledge traditional EBM and focused on pragmatic EBM.This practice-based approach to EBM considered theimportant point that knowing what to do, how to do it,

and the likely outcome of an orthopaedic interventiondescribed in NICE guidelines were only part of theknowledge picture which had to fit into the ever chan-ging context of practice. However, external evidence cre-ated by regulators and managerial and organisationalknowledge could limit the behaviour and decisions ofsurgeons. For example, surgeons working here wererestricted in the hip replacement implants, they couldselect based on cost and procurement contracts (meso)and thresholds established by the Orthopaedic DeviceEvaluation Panel in the UK (macro).

Case BCase B was a small hospital Trust split between twogeographical sites. Case B demonstrated an ‘“it depends”approach to decision-making produced by the binarycharacteristics of the Trust’; this was bought about bythe structure and location of the hospital. The binarycharacteristics reflect the two distinct hospital locationsin case B. By name, the hospitals were one Trust; how-ever, the day-to-day operations and decision-makingpractices were separated and distinct. Each hospital loca-tion in case B had their own way of doing things (organ-isational knowledge) and staff acted protectively tomaintain them. The ‘it depends approach’ signifies theparticipants views that what happened in practicedepended on which of the two case B hospital sites theywere located at when the decision was made.In case B, external evidence created by regulators was

not regarded as important. Formal codified evidence inNICE guidelines was not valued by the clinical staff.Guidelines were often considered in a negative light andconsidered to be the responsibility of hospital admin-istrators and managers (i.e., managerial knowledge).Described by a hospital board member below:

“I think NICE guidance is very much just seen asanother layer of administration for clinicians. If noone’s looking at whether you’re following NICEguidance or not they just sit on a shelf, unless you

Table 4 Narrative descriptions and an example of each source of evidence and knowledge depicted in Fig. 2 (Continued)

Evidence and knowledge types Narrative description Example from the data

Training and formal education The training and formal education ofhealthcare professionals which arerecognised through standard academicqualifications

A Master’s degree in Evidence-BasedMedicine

Apprenticeship style training andinformal education

Personal training which occurs duringeach working day with senior colleagues

Training gained through fellowshipprogrammes and practice-based learning

Individual patient and surgeon factors Characteristics of the patient or surgeonthat influenced clinical practice decisions

Patients age or a surgeons years in practice

Evidence linked to the innate ‘feel’of surgery

A description of the surgeon’s judgement,skill, craft and instinct

A surgeon not knowing exactly whatwill occur during an operation until theystarted the surgery and can see and feelthe operation takes place

Grove et al. Implementation Science (2018) 13:75 Page 8 of 14

have a very active team of clinicians who take this onboard. But that’s not a consistent.” (Hospital boardmember)

The negative value attached to NICE and guidelineswas echoed by the surgeon in the quote below, in whichthey demonstrate the importance of socialisation and as-sociation with colleagues when making clinical decisions:

“NICE are not being proactive enough, I would say, interms of making recommendations on prostheses andthey could do a lot more. There is very little sort ofrobust evidence to guide practice so you rely on otherpeoples’ anecdotal experience and normal practice tohelp guide what, what works and what doesn’t”(Consultant surgeon)

Within the organisational processes, guidelines werenot considered an important part of practice in case B.Compliance to NICE guidelines as evidence of externalregulation did not appear to be valued and hence imple-mentation was haphazard, as described by a hospitaladministrator:

I send (guidance) out to the General Manager andthen they will send it to the most appropriate person.We used to meet to discuss if we were compliant…but now we send a questionnaire out…so they have todo is tick ‘Yes’ and ‘No’ but they do not alwaysrespond. (Administrator)

What mattered for implementation in case B was thedynamics of organisational change and leadership enactedthrough evidence from the professional hierarchy in thetwo orthopaedic departments within the Trust. The indi-vidual surgeon factors and the differing characteristics andprocesses of the two groups of surgeons headed up byopinion leaders meant that guideline implementationacross the organisation was difficult. The distinct groupsstruggled to work together and share organisational know-ledge because the specific contexts that clinicians weresocialised into differed. For example, evidence from im-plant manufactures carried greater weight in one locationcompared to the other. The socialisation and associationwith colleagues reflected the behaviour and norms thatguided or regulated the action and decision-making ofindividuals at case B.

Case CFinally, case C was an orthopaedic department located in ateaching hospital. In case C, ‘socialised decision-makingwas prominent and evidence was discretionary’. The ortho-paedic department in case C was closed to the influence ofadministrators and managerial knowledge emerging from

elsewhere in the Trust. They were also resistant to pressurefrom external policymakers and evidence from regulators.This divide across professional boundaries compounded bythe professional hierarchy and cultural norms of the sectormade guideline implementation challenging. Nevertheless,knowledge and evidence generated through managerialand organisational knowledge was present in the widerhospital organisation:

We have standard operating procedures for NICE,which encourage [surgeons] to write their own actionplan (Trauma & Orthopaedic Manager)

However, observations of practice revealed that man-agerial and organisational knowledge enacted throughprocesses attached to NICE guidelines and governancebelonged in the managerial and administrative domain,not within surgical staff. Surgeons reported having “neverseen” the organisation’s NICE process (Consultantsurgeon). This is reflected in the observation note below:

NICE was rarely noted as an influential factor in theday-to-day activities of surgeons. Surgeons I spoke towere unaware their hospital had a NICE process, theywould respond “do we have one?” “I’ve never seen it”.(Observation note)

Clinical practice decisions were made using knowledgeand evidence gained through socialisation and associ-ation with colleagues. Surgeons possessed resilient ex-periential knowledge built up over time as the majorityhad been working in this hospital for their entire careersand were relatively separate from ‘outside’ knowledge.Formal codified evidence in clinical guidelines had tocompete with the complex social systems that existed inthe hospital. For example, case C was a referral centrewhich specialised in performing complex hip replace-ment revision surgery. In this context, clinical guidelinesappeared to be less important because the approaches,techniques and implants needed to perform hip revisionswere specialist (the innate feel of surgery) and thereforenot included in guideline recommendations. One sur-geon noted “NICE, is irrelevant. They don’t tell meanything I NEED to know” (Consultant surgeon). Thespecialist surgeons working in case C referred to a stand-ard hip replacement operation as “boring” and, hence thework of other surgeons “on the treadmill” who are outsideof their community or social system (Consultant surgeon).In this sense, guidelines were not valued or applicable totheir specialist work. The informal experimental know-ledge and informal experiential evidence built up over timeappeared to take precedence in clinical decision-makingprocesses and therefore restricted guideline implementa-tion of guidelines for this group of surgeons. These

Grove et al. Implementation Science (2018) 13:75 Page 9 of 14

dominate types of knowledge and evidence were groundedin the experience of surgical work in practice and legacyknowledge about the organisational functions which sur-geons developed as a consequence of working there for along period of time.

DiscussionIn orthopaedic surgery, clinical guidelines are an import-ant part of practice as they not only help guaranteesafety and encourage quality improvement but also en-sure that NHS resources are used appropriately [32].The aim of this study was to identify where, when andhow evidence and knowledge are used in healthcaredecision-making in orthopaedic surgery and how vari-ation in evidence and knowledge contributes to differ-ences in the implementation of clinical guidance inpractice. Previous scholars have counted and categorisedevidence using broader taxonomies, which describeknowledge as individual, group, tacit or explicit [33–35].Others emphasise the role of the person in the activityto distinguish between action, doing and practice, andknowledge facts and processes [36]. It is important tohighlight that what is considered evidence and know-ledge is highly contested and influenced by the environ-ment in which it is used [10].In fulfilling our aim, we discovered 17 different types

of knowledge and evidence which were used in ortho-paedic surgery. During our case comparison, it becameclear that the dynamic selection, combination and use ofeach type of knowledge and evidence influenced the im-plementation and use of clinical guidance in practice. Atthe time of study, none of the cases could definitivelyprovide evidence to demonstrate that they were acces-sing, using and monitoring guideline recommendations.We examined implementation of guidelines by individ-ual surgeons, groups of surgeons and the Trusts as awhole. Interestingly, implementation could differ be-tween these levels.In the context of orthopaedic surgery, the process of

privileging different types of knowledge and evidencein the context of surgery resulted in three distinct ap-proaches to decision-making in orthopaedic surgery.These include ‘pragmatic EBM decision-making’ (case A),where NICE guidelines failed to deliver all the knowledgeand evidence needed to make clinical decisions whenorganisational context restricted surgeon choice. An ‘“itdepends” approach produced by the binary characteristics’of case B linked to the professional hierarchy of surgeonsworking in separate geographical locations and ‘socialiseddecision-making’ where evidence was discretionary due tothe strong influence of socialisation and informal know-ledge sharing between surgeons in case C.The similarities and differences between the three

approaches generate key contextual dimensions specific

to guideline implementation in orthopaedic surgery.These reflect the relationship between guideline imple-mentation and knowledge and evidence that is actuallyused in orthopaedic practice. Orthopaedic surgeons inour study held ambivalent or negative attitudes to-wards clinical guidelines. They did not privilege thisformal codified evidence because it originated outsideof orthopaedics and did not contain the micro sourcesof knowledge and evidence (e.g., experience, training,individual characteristics, and the innate feel of sur-gery) that were considered more important for sur-geons’ decision-making. However, the culture andnorms of EBM, identified in case A, demonstrate thatit was possible to positively influence guideline imple-mentation. Consideration of the power of professionalhierarchies is vital, as surgeons working at the top ofthe hierarchy, such as clinical leaders can restrict ordiminish the influence of mangers and policymakers.Organisational constraints linked to financial restric-tions, regulation and procurement-influenced imple-mentation and a lack of focus on clinical guidelines inorthopaedic practice. The presence of organisationalprocesses and protocols could not ensure what guide-lines were valued and used. What was more importantis what evidence and knowledge transferred from surgicalcolleagues and professional societies.It is likely that many of the 17 types of knowledge and

evidence identified in this research would not have beendiscovered without the structured, comparative casestudy approach used in our study. One of the aims ofour research was to go beyond reports in the previousguideline implementation literature [3–9, 37]. It is sig-nificant that a large number and diverse range of know-ledge and evidence types acting across the entire domainof healthcare emerged from the three case studies ofNHS practice. Previous research has identified a consid-erable number of barriers and facilitators to the imple-mentation of clinical guidelines but often they are toogeneric or limited in scope, for example, the difficultly ofengaging individual clinicians or problems with processand resource issues within hospitals [7, 13, 14].We have demonstrated that guideline implementation

and subsequent evidence-based practice were not alwayspossible or preferable in the three cases. Comparativecase study analysis revealed dynamic contextual differ-ences and variation in practices and processes betweenthe three hospitals. For example, some surgeons hadstrict limits placed on the orthopaedic implants theycould order within their hospital. This restricted theirimplant decisions. However, this varied across the casesand differences were found in implant selection practicesand processes. This variation had a direct impact on theuptake and use of evidence in practice and demonstratesthe problems of effectively implementing standardised

Grove et al. Implementation Science (2018) 13:75 Page 10 of 14

guidance in surgery. One of the aims of clinical guidanceis to reduce unjustified variation in practice, but wefound that this was not achievable or appropriate in allcontexts.Our findings encourage more research into how to en-

gage individuals and groups of healthcare professionalsto consider the content of clinical guidelines and how itmight add to their decision-making processes. The over-arching view that guidelines are the responsibility ofmanagers and administrators demonstrates a lack ofownership of the guideline in general in orthopaedics.Also, in the context of orthopaedics, our findings pro-vide insight into approaches to knowledge mobilisationtargeted at communities of practice as an area for inves-tigation and improvement. Improvements in the uptakeand use of NICE guidance in orthopaedic practice willrequire the development, presentation and disseminationof evidence-based guidelines in surgery to be bettertailored to the orthopaedic community.

Strengths and limitationsThe key strength of our research is the use of case studiesto examine the context of guideline implementation. Inachieving our aim, we were able to discover and explainthe gap in implementation of clinical guidelines in ortho-paedics. Uptake and use of guidelines, even whengrounded on the findings of empirical research includinggold standard randomised controlled trials, were not guar-anteed in practice. This suggests that well-developedguidelines are necessary but not sufficient to achieve thegoals of policymaking organisations such as NICE andScottish Intercollegiate Guidelines Network (SIGN). Casestudy methods allowed us to describe the decision-makingcontext in detail. This method has provided a greaterdepth of description of the evidence and knowledgesources than have been outlined in previous literature, forexample, expanding the tacit-explicit-group-individualcategorisations. The way in which evidence and know-ledge interacted with context produced variation in theextent to which guidelines are implemented and used inorthopaedics.A second strength of our research is the use of

multiple data sources (interviews, observations anddocument analysis) to study the same phenomena. Thecombination of methods facilitated us to overcome theweaknesses that emanate from selecting a single methodto study the complex process and practice of guidelineimplementation [38]. We triangulated data to enhancethe credibility of our analysis and findings. When datafrom one source substantiates a pattern from another,the findings are stronger and better substantiated [26].Comparison across our three comparative cases enabledus to generate a more sophisticated understanding ofthe data we collected [22, 23].

Our study has limitations. The direct observation ofhealthcare professionals in their practice may havedriven a change to ‘good’ or ‘better’ behaviour by partici-pants; a phenomenon known as observer bias [39]. Toensure the quality and rigour of our data, we extendedour access and observation as much as possible, whilstalso conducting crosschecks and validation during inter-views and between different individuals and professionalgroups. We sampled three hospitals from the populationof 135 hospital Trusts in England which deliver hip re-placement services [40]. Although the sample was small,we aimed to achieve a broad representation of the typesof elective orthopaedic services available in England andaimed to produce in-depth rather than a breadth in ourcase study design and data.

Policy and practice implicationsOver the last 20 years, there have been significantchanges in the way policy-making organisations such asNICE create and disseminate guidance to improve healthand social care. What appears to have remained constantis the way in which codified knowledge in guidance isproduced with an assumption that a linear ‘push’ fashionwill ensure that it is received and acted on by cliniciansworking in healthcare organisations. The findings of ourstudy reconfirm and extend our knowledge of the limitsof this approach for improving the use of guidance andfor reducing variation in practice for orthopaedics.In this study, we have raised the issue of whether

NICE guidelines are ever likely to be appropriate for thefield of orthopaedics. This is due to the wide range ofknowledge and evidence identified as influential todecision-making, coupled with the differences in guide-line implementation across the structural levels. How-ever, we do not consider that the evidence contained inguidelines is inappropriate. Instead, the ways in whichknowledge and evidence were privileged differently byindividual practitioners, groups of surgeons and organi-sations meant that guidelines were rarely accessed as abeneficial evidence source. Surgeons were not concernedabout what guidelines recommended. What was import-ant was their definition of knowledge and evidence andhow this interacted with understandings of knowledgeand evidence in their group and wider organisation. Inthis study, ‘one size’ guidance could never ‘fit all’ the sur-geons’ requirements and therefore, the guidance hadlimited value in their specific circumstances.Nevertheless, evidence in guidelines represents best prac-

tice, and NICE and SIGN must produce recommendationsfor healthcare. We have provided evidence to suggest thatthe current modes of transfer and implementation are inef-fective. Changes could be made to the process of guidelinecreation, dissemination or even regulation to move towardseffective knowledge mobilisation. For example, a more

Grove et al. Implementation Science (2018) 13:75 Page 11 of 14

inclusive process of involving clinicians in guideline devel-opment would be welcomed. The current system relies onclinicians being aware of guideline updates, rather thanbeing enabled to actively volunteer their contributions. Tar-geting the orthopaedic community through professionalmeetings, networks and clinical leaders would communi-cate the need for involvement and increase discussionabout new guidelines in contrast to the current process ofone-way dissemination by policymakers.Regulation was a valuable mechanism as it achieved a

desired outcome for achieving targets and controlling thebehaviour of the surgeons. However, moving towardsregulation as the norm did not appear to be a desirableoption for most of the professional groups in this study.Knowledge and evidence from external regulators did nothold the same positive status achieved by the knowledgeand evidence emanating from colleague and professionalnetworks. Restricting the discretion and authority of clin-ical professionals by increasing regulatory power wouldnot be recommended. Instead, interventions which takeadvantage of the positive knowledge mobilisation betweenorthopaedic colleagues are encouraged, as they may im-prove the sharing of evidence-based practice.Improvements need to be made to how healthcare

professionals working in hospitals see and think aboutevidence from guidelines in combination with otherknowledge and evidence sources. Improvement interven-tions are required to help users of guidance identify‘where they are at’ in their decision-making processes.Clinical guidelines were often unable to provide a solu-tion to a decision problem; therefore, it is important tounderstand what other types of evidence and knowledgeare available or used by others. Practitioners need toacknowledge the difference between certain types ofknowledge as positive or negative to patient care. Wherepossible, those working in healthcare should focus onreducing undesirable types of evidence and knowledgepresent in their organisation. This could be an area forimprovement work. However, surgeons in this studywere often unaware of or ambivalent about the conse-quence of their decisions because processes were notopen, transparent or subject to feedback loops.Healthcare practitioners could take a more transparent

approach in understanding the evidence that is drivingtheir decisions and how guidelines may fit into the pic-ture. This will facilitate practitioners in deciding whetherguideline recommendations are appropriate in their con-text. If not, other knowledge sources such as clinical ex-perience could take precedence and be shared, explainedand understood, rather than frowned upon by managersand administrators and recorded as an organisationalrisk. Practice-based knowledge was rarely shared be-tween the professional groups in this study. Encouragingopen decision-making processes might enable those on

all sides of knowledge boundaries to understand and acceptwhy certain options are chosen and actions are taken, espe-cially if they vary from guideline recommendations.

ConclusionThe research aimed to explore guideline implementationthrough the application of comparative case studieswhich investigate the use of NICE clinical guidelines indecisions made in elective orthopaedic surgery in theNHS. The results of our study highlight the range ofcomplex and competing sources of evidence and know-ledge which influence the work practices of healthcareprofessionals. Case study analysis revealed three distinctstyles of orthopaedic practice which represent the waysin which 17 types of knowledge and evidence were usedduring decision-making. The way in which evidence andknowledge were selected and used impacted on howguidelines were implemented in the orthopaedic special-ity. Findings from the case comparison reflect the com-plexity of evidence-based decision-making in the highlyprofessionalised organisationally regulated context ofsurgery. Our results could be used to guide the develop-ment of implementation interventions that are groundedin the findings of this study. New approaches to imple-mentation need to appreciate and incorporate the di-verse range of knowledge and evidence which influencesclinical decisions in orthopaedics and to take account ofthe changing contextual situations in which decisionsneed to be made.

Additional file

Additional file 1: An example of a NICE guidance recommendation.Technology appraisal guidance [TA304] [17]. Total hip replacement andresurfacing arthroplasty for end-stage arthritis of the hip. Interview topicguide. (DOCX 15 kb)

AbbreviationsAHPs: Allied Health Professionals (including Occupational Therapists,Operating Department Practitioners, Physiotherapists, Nurses and HealthcareAssistants); EBM: Evidence-based medicine; NHS: National Health Service;SIGN: Scottish Intercollegiate Guidelines Network

AcknowledgementsWe thank the three hospital departments who agreed to participate in theresearch project, and all the NHS staff who took part in the interviews andassisted with observations.

FundingAmy Grove was supported by a National Institute for Health ResearchDoctoral Fellowship programme project number 2013-06-064. Aileen Clarkeand Graeme Currie were partly supported by the National Institute for HealthResearch Collaboration for Leadership in Applied Health Research and CareWest Midlands at University Hospitals Birmingham NHS Foundation Trust.The views expressed are those of the authors and not necessarily those ofthe NHS, the NIHR or the Department of Health and Social Care.

Grove et al. Implementation Science (2018) 13:75 Page 12 of 14

Availability of data and materialsThe datasets generated and analysed during the current study are notpublicly available due to restrictions of ethical approvals obtained for thisstudy.

Authors’ contributionsAll authors (AG, AC, GC) have made substantial contributions to the designof the study. AG collected the data, and all authors analysed and interpretedthe data. All authors have been involved in drafting the manuscript andhave given final approval of the version to be published. All authors agree tobe accountable for all aspects of the work in ensuring that questions relatedto the accuracy or integrity of any part of the work are appropriatelyinvestigated and resolved.

Ethics approval and consent to participateThe study was approved by the Biomedical and Scientific Research EthicsCommittee of The University of Warwick, England (approved June 2nd 2014;reference number REGO-2014-645) and the Research and Developmentdepartments of each of the three hospital sites (case A approved on 30 June,2014, case B approved on 23 October, 2014, case C approved on 21 August2014). In line with the ethical approval, each participant will be asked to signan informed consent form as will be provided with information outlining thepurpose of the study and their rights to withdraw.

Consent for publicationInformed consent was obtained to include anonymised data in anypublications. Confidentiality is protected as the data is anonymised.

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Author details1Division of Health Sciences, Warwick Medical School, University of Warwick,Coventry, UK. 2Entrepreneurship & Innovation, Organising HealthcareResearch Network, Warwick Business School, University of Warwick, Coventry,UK.

Received: 18 January 2018 Accepted: 23 May 2018

References1. NICE. About. 2017. http://www.nice.org.uk/about. Accessed on 6 Dec 2017.2. SIGN. Who we are. 2017. http://www.sign.ac.uk/who-we-are.html. Accessed

on 6 Dec 2017.3. Lowson K, Jenks M, Filby A, Carr L, Campbell B, Powell J. Examining the

implementation of NICE guidance: cross-sectional survey of the use ofNICE interventional procedures guidance by NHS Trusts. Implement Sci.2015;10:93.

4. Sheldon T, Cullum N, Dawson D, Lankshear L, Lowson K, Watt I, West P,Wright D, Wright J. What’s the evidence that NICE guidance has beenimplemented? Results from a national evaluation using time series analysis,audit of patients’ notes, and interviews. BMJ. 2004;329:999.

5. Weng Y-H, Kuo K, Yang C-Y, Lo H-L, Chen C, Chiu Y-W. Implementation ofevidence-based practice across medical, nursing, pharmacological and alliedhealthcare professionals: a questionnaire survey in nationwide hospitalsettings. Impact Sci Soc. 2013;8:112.

6. Francke A, Smit M, de Veer A, Mistiaen P. Factors influencing theimplementation of clinical guidelines for health care professionals: asystematic meta-review. BMC Med Inform Decis Mak. 2008;8:38.

7. Fitzgerald A, Lethaby A, Cikalo M, Glanville J, Wood H. Review of systematicreviews exploring the implementation/uptake od guidelines. York HealthEconomics Consortium. 2014. https://www.nice.org.uk/guidance/ph56/evidence/evidence-review-2-431762366. Accessed 8 Jan 2018.

8. Royal College of Physicians. Implementing NICE public health guidance forthe workplace: a National Organisational Audit of NHS Trusts in England.2015. https://www.rcplondon.ac.uk/projects/outputs/implementing-nice-public-health-guidance-workplace-2013-round-2. Accessed 8 Jan 2018.

9. Platt C, Larcombe J, Dudley J, McNulty C, Banerjee J, Gyoffry G, Pike K,Jadresic L. Implementation of NICE guidance on urinary tract infections inchildren in primary and secondary care. Acta Paediatr. 2015;104:630–7.

10. Gkeredakis E, Swan J, Powell J, Nicolini D, Scarbrough H, Roginski C, Taylor-Phillips S, Clarke A. Mind the gap: understanding utilisation of evidence andpolicy in health care management practice. J Health Organ Manag. 2011;25:298–314.

11. Currie G, El Enany N, Lockett A. Intra-professional dynamics in translational healthresearch. The perspective of social scientists. Soc Sci Med. 2014;114:81–8.

12. Davies HTO, Powell AE, Nutley SM. Mobilising knowledge to improve UKhealth care: learning from other countries and other sectors—amultimethod mapping study. Health Serv Deliv Res. 2015;3:27.

13. Denny K. Evidence-based medicine and medical authority. J Med Humanit.1999;20:247–63.

14. Timmermans S. From autonomy to accountability: the role of clinical practiceguidelines in professional power. Perspect Biol Med. 2005;48:490–501.

15. Grove A, Johnson R, Clarke A, Currie G. Evidence and the drivers of variationin orthopaedic surgical work: a mixed method systematic review. HealthSyst Policy Res. 2016;3:1.

16. Ferlie W, Wood M, Fitzgerald L. Some limits to evidence-based medicine: acase study from elective orthopaedics. Qual Health Care. 1999;9:99–107.

17. NICE Technology Appraisal (TA304). Total hip replacement and resurfacingarthroplasty for end-stage arthritis of the hip. NICE. 2014. https://www.nice.org.uk/guidance/ta304. Accessed 30 Nov 2017.

18. Stake R. The art of case study research. London: Sage; 1995.19. Grove A, Clarke A, Currie G. The barriers and facilitators to the

implementation of clinical guidance in elective orthopaedic surgery: aqualitative study protocol. Implement Sci. 2015;2:81.

20. Sandelowski M. The problem of rigor in qualitative research. Adv Nurs Sci.1986;8:27–37.

21. Eisenhardt KM. Building theories from case study research. Acad Manag Rev.1989;14:532–50.

22. Mantere S, Ketokivi M. Reasoning in organization science. Acad Manag Rev.2013;38:70–89.

23. Parlett M, Hamilton D. Evaluation as illumination: a new approach to thestudy of innovatory programs. In: Glass G, editor. Evaluation studies reviewannual, vol. 1. Beverly Hills: Sage; 1976.

24. George AL, Bennett A. Case studies and theory development in the socialsciences. Cambridge: MIT Press; 2005.

25. Yin R. Case study research. Beverly Hills: Sage Publications; 1984.26. Creswell J, Clark V. Designing and conducting mixed methods research.

London: Sage Publications; 2010.27. Kinsman L, Rotter T, James E, Snow P, Willis J. What is a clinical pathway?

Development of a definition to inform the debate. BMC Med. 2010;8:131.28. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol.

2006;3:77–101.29. Johnson R, Grove A, Clarke A. Pillar integration process: a technique to

integrate data in mixed methods research. J Mixed Methods Res. 2017:1–20.https://doi.org/10.1177/1558689817743108.

30. Pettigrew AM. Longitudinal field research on change: theory and practice.Organ Sci. 1990;1:267–92.

31. Pope C, Robert G, Bate P, Le May A, Gabbay J. Lost in translation: a multi-level case study of the metamorphosis of meanings and action in publicsector organizational innovation. Public Adm. 2006;84:59–79.

32. NICE. The NICE way. Lessons learned from the National Institute for Healthand Care Excellence. 2014. https://www.nice.org.uk/news/article/using-nice-s-approach-to-base-policy-decisions-on-evidence-could-help-save-billions.Accessed 30 Nov 2017.

33. Cook S, Brown J. Bridging epistemologies: the generative dance betweenorganizational knowledge and organizational knowing. Organ Sci. 1999;10:381–400.

34. Nonaka I. A dynamic theory of organizational knowledge creation. OrganSci. 1994;5:14–37.

35. Lam A. Tacit knowledge, organizational learning and societal institutions: anintegrated framework. Organ Stud. 2000;21:487–513.

36. Orlikowski W. Knowing in practice: enacting a collective capability indistributed organizing. Organ Sci. 2002;13:249–73.

37. Berry K, Haddock G. The implementation of the NICE guidelines forschizophrenia: barriers to the implementation of psychologicalinterventions and recommendations for the future. Psychol Psychother.2008;81:419–36.

Grove et al. Implementation Science (2018) 13:75 Page 13 of 14

38. Dixon-Woods M, Agarwal S, Young B, Jones D, Sutton S. Integrativeapproaches to qualitative and quantitative evidence. London: NHS HealthDevelopment Agency; 2004.

39. Snyder ML, Frankel A. Observer bias: a stringent test of behavior engulfingthe field. J Pers Soc Psychol. 1976;34:857.

40. NHS Confederation. NHS statistics, facts and figures. 2017. http://www.nhsconfed.org/resources/key-statistics-on-the-nhs. Accessed on 6 Dec 2017.

Grove et al. Implementation Science (2018) 13:75 Page 14 of 14