Deferred consent - Erasmus Critical Care · Deferred consent Prof. Jan Bakker MD, PhD Chair dept Intensive Care Adults [email protected] Thursday, March 22, 12

Deferred consent

Prof. Jan Bakker MD, PhD

Chair dept Intensive Care Adults

[email protected]

www.intensivecare.me

Thursday, March 22, 12

mailto:[email protected]

mailto:[email protected]

http://www.intensivecare.me

http://www.intensivecare.me

Emergency research

‣Autonomy of the patient and guaranteeing patient wishes’

‣Protection against discomfort, harm, risk, exploitation

‣Prospect of potential benefit

• Studies comparing interventions with clinical equipoise


0

50

100

150

200

250

300

350

400

1-2h 2-3h 3-4h 4-5h >5h

Injury-Drug admin

Delay in clinical researchDexanabinol in Traumatic Brain Injury

Progress in Brain Research 2007; Vol. 161, Chapter 17: 243-250 § Lancet Neurology 2006;5:38-45

Number of patiens

(861 patients enrolled in the study)

Time between injury and study drug administration

Therapeutic window form animal studies


0

1

2

3

4

5

6

Belgium Netherlands Israel Spain Germany Italy France Others

Injury-Admission

Injury-CT

Injury-Consent

Injury-Drug admin

Delay in clinical researchDexanabinol in Traumatic Brain Injury

Progress in Brain Research 2007; Vol. 161, Chapter 17: 243-250 § Lancet Neurology 2006;5:38-45

Hours

Therapeutic window form animal studies


Practice of informed consent in IC research

‣Prior consent in emergency research can result in selection bias• Canadian Stroke Network: mortality enrolled patients

6.9% vs 21.7% in not enrolled patients (N Engl J Med 2004;350:1414-1421)


Deferred ConsentA New Approach for Resuscitation Research on Comatose PatientsNorman S. Abramson, MD; Alan Meisel, JD; Peter Safar, MD

Methodological constraints inherent in the rapidly growing field ofresuscitation research have created an apparent conflict with newlypromulgated federal regulations, especially those concerning informedconsent requirements. We propose that two new concepts be applied toresuscitation research to satisfy the current federal regulations governingbiomedical research. The first of these concepts is "minimal differentialrisk," ie, in resuscitation medicine, the difference between the risk of anundesirable outcome when standard, commonly accepted therapy may beused and the risk of an undesirable outcome with experimental therapy isminimal. The second concept is "deferred consent," ie, obtaining consent tocontinue with an experimental therapy after administration of that therapyhas already begun. We believe that the emergency exception to informedconsent applies to resuscitation research. Recognition of the applicability ofthese concepts to resuscitation research should allow compliance withfederal and state regulations that otherwise might inhibit or preclude suchresearch endeavors.(JAMA 1986;255:2466-2471)

CONTINUOUSLY increasing publicsupport of biomédical research isreflected by the increased spending ofthe National Institutes of Health(NIH) from $701,800 in 1945 to $436million in 1965' and the projectedNIH budget of $4.5 billion in 1985.2Accompanying this increasing finan¬cial support has been closer publicscrutiny and control. Five years ago,we discussed the emerging conflictbetween newly promulgated federalregulations for the conduct of humanbiomédical research and methodologi-

cal constraints inherent in the rapidlygrowing field of clinical resuscitationresearch, ie, research conducted on

comatose patients in need of immedi¬ate medical treatment.3 Our purposenow is to update this discussion byreviewing recent developments in thegovernment regulation of human bio-medical research and to report on oursolution to the problem of obtaininginformed consent, in an ongoing,NIH-funded, international collabora¬tive randomized clinical trial of car-diopulmonary-cerebral resuscitationafter cardiac arrest, the Brain Resus¬citation Clinical Trial (BRCT).4The first phase, BRCT I, of our

brain resuscitation clinical trials(1979-1983) tested the brain damage-ameliorating effect of thiopental sodi¬um loading after cardiac arrest.Study patients were comatose cardiacarrest survivors whose prognosis for

long-term survival and meaningfulneurologic recovery was dismal. Be¬cause most American institutionalreview boards (IRBs) at that timerequired the usual prospective in¬formed consent from next of kin, andbecause a time constraint was neces¬

sary for the initiation of drug therapy(brain resuscitation therapies, to beeffective, must be administered soonafter the insult), few patients wereentered in participating US institu¬tions. In other words, time was usu¬

ally not sufficient to obtain informedconsent.Based on this experience and on the

legal, ethical, and medical considera¬tions discussed in this article, we

urged the IRBs of hospitals partici¬pating in BRCT II (1984-1987) toconsider the use of "deferred consent"when prior informed consent isimpossible to obtain. This deferredconsent mechanism involves the ran¬dom assignment of patients (whomeet the medical criteria for admis¬sion into the research protocol) toeither the standard or experimentaltherapy group without obtaining in¬formed consent from either thepatients (who are comatose) or theirfamilies (who are not available in theshort, critical period during whichtreatment must be initiated). How¬ever, once family members becomeavailable, an explanation of theresearch protocol is provided, includ¬ing the means by which the patientwas assigned to treatment. At thattime, they are provided with an

opportunity to withdraw the patientfrom the study so that all subsequent

From the Resuscitation Research Center, Depart-ment of Anesthesiology/Critical Care Medicine (DrsAbramson and Safar), and the Western PsychiatricInstitute and Clinic and the School of Law (MrMeisel), University of Pittsburgh.Reprint requests to Resuscitation Research Cen-

ter, 3434 Fifth Ave, Pittsburgh, PA 15260 (DrAbramson).

at Erasmus MC - Univ of Rotterdam on March 14, 2012jama.ama-assn.orgDownloaded from

JAMA 1986;255:2466-2471

After lengthy deliberations, 15 of the 16 participating American institutions approved the deferred consent mechanisms. Because of cultural and philosophical differences from the United States, approval of the deferred consent mechanism was more easily obtained in the eight participating European hospitals.

“obtaining consent to continue with an experimental therapy after administration of that therapy has already begun”


Waiver of consentUSA and Europe

‣ US Federal Code of Regulation accepted emergency research to be performed without informed consent in 1996

• necessity to explain the protocol to the relevant community

‣ European Council Directive in 2001 required written informed consent and a waiver of consent was not possible

• some countries did not comply (Belgium, France, The Netherlands, Spain and Germany) where others did (Denmark, Ireland, Portugal and UK)

✓ surrogate (spouse, legal representative, (independent-dependent) physician) may provide consent


Waiver of (prior) consentPublic view

‣ Public views supports exceptions on informed consent in emergency situations

• survey in 530 people: 88% prior information is required § 49% enrollment in emergency situation is acceptable § 70% would not object to be enrolled without prior consent


Ethics of Deferred (Proxy) Consent

1. Can proxies give a valid judgement for consent or refusal on immediately on admission?


Proxy consent‣ Proxies are not always available in first hours of admission

‣ Emotional nature of the emergency situation limits the reliability of proxy consent

• Acad Emerg Med 2001;8:851 § Lancet 2000;356:2045 § Crit Care Med 2003;31:S178

‣ Comprehension of provided information is limited in emergency situations

• Lancet 2003;361:918 § Soc Sci Med 1995;40:1573 § Am Hear J 2000;140:2

‣ Families are in psychological stress and have limited effective coping mechanisms

• J Adv Nurs 1990;15:1402 § Intensive Crit Care Nurs 2007;23:43

‣ Families want to know about diagnosis, prognosis and uncertainty influences their reactions, actions and strategies

• J Clin Nurs 2005;14:501 § Intensive Crit Care Nurs 2007;23:170


Proxy consent

‣ Decision making competence

• factual understanding

• evidencing a choice (consent or refusal)

• reasoning and appreciation of the situation

Family members are probably (temporarily) incompetent in these 3 points and are not likely to provide valid proxy consent during the acute phase


Intensive Care Med (2006) 32:2020–2025DOI 10.1007/s00134-006-0358-4 O R I G I N A L

Sheila E. HarveyDiana ElbourneJoanne AshcroftCarys M. JonesKathryn Rowan

Informed consent in clinical trials in criticalcare: experience from the PAC-Man Study

Received: 19 January 2006Accepted: 26 July 2006Published online: 21 September 2006© Springer-Verlag 2006

Electronic supplementary materialSupplementary material is available in theonline version of this article at http://dx.doi.org/10.1007/s00134-006-0358-4 andis accessible for authorized users.

The authors wrote this article on behalf ofthe PAC-Man Study Collaboration

This article is discussed in the editorialavailable at: http://dx.doi.org/10.1007/s00134-006-0359-3

Ethics approval: The PAC-Man trial(Application reference numberMREC/00/2/83) was approved by the NorthThames Multi-centre Research EthicsCommittee on 12 March 2001, as well as allrelevant Local Research Ethics Committees,and Research & Development Departmentsat collaborating hospitals.

Funding: The PAC-Man trial (HTA ProjectNo. 97/08/03) was funded by the UKNational Health Service Research &Development Health TechnologyAssessment Programme. (Disclaimer: Theviews and opinions expressed in this articledo not necessarily reflect those of theNational Co-ordinating Centre for HealthTechnology Assessment.) The NHSResearch & Development HealthTechnology Assessment Programme had noinvolvement in the design, conduct andanalysis of the study.

Competing interest statement: All authorsdeclare no conflict of interest.

Contributors: S.E. Harvey conceived anddesigned the study and participated in the

analysis and interpretation of data and inwriting the paper. D. Elbourne participatedin the analysis and interpretation of data andin writing the paper. J. Ashcroft and C.M.Jones collected the data and participated inthe interpretation of data and in writing thepaper. K. Rowan conceived and designedthe study and participated in theinterpretation of data and in writing thepaper. S.E. Harvey is the guarantor of thepaper and accepts full responsibility for theconduct of the study, had access to the data,and controlled the decision to publish.

S. E. Harvey · K. RowanIntensive Care National Audit & ResearchCentre, Tavistock House,Tavistock Square, WC1H 9HR, London,UK

D. Elbourne (!)London School of Hygiene & TropicalMedicine, Medical Statistics Unit,Keppel Street, WC1E 7HT, London, UKe-mail: [email protected].: +44-20-79272629Fax: +44-20-76372853

J. AshcroftUniversity of Manchester, UK MentalHealth Research Network,PR2 8DY, Preston, UK

C. M. JonesRoyal Berkshire Hospital, Intensive CareUnit,RG1 5AN, Reading, Berkshire, UK

Abstract Objectives: To identifythe proportion of critically ill patientsable to consent to participation ina randomised controlled trial (RCT)

and to assess to what extent patientconsent and relative assent processescould be conducted according toethics committee permissions. De-sign: Descriptive study nested in anRCT. Setting: Fifty-six UK intensivecare units participating in the PAC-Man trial. Patients and participants:First 500 patients consecutively en-rolled into PAC-Man. Measurementand results: The outcome measureswere patient consent and/or relativeassent. Of the 498 patients included,13 (2.6%) provided consent beforerandomisation. Of the remaining 485patients, relative assent was obtainedfor 394 patients (81.2%), and refusedpost-randomisation for 3 patients(0.6%). No relatives were availablefor 15 patients (3.1%), and it wasunclear from documentation whetherrelative assent had been obtainedfor 73 patients (15.1%). Of the 482patients who did not provide consentprior to randomisation, 188 (39%)survived. Of these, 175 (93.1%) gaveretrospective informed consent, six(3.2%) refused, and seven (3.7%) didnot regain mental competency. Con-clusions: A very small proportionof patients were able to give consentbefore randomisation. Due to the highin-hospital mortality (60.6%), onlyaround one third of the remainingpatients could provide consent retro-spectively. This study demonstratesdifficulties experienced in obtainingconsent from critically ill patients

Intensive Care Med (2006) 32:2020–2025DOI 10.1007/s00134-006-0358-4 O R I G I N A L

Sheila E. HarveyDiana ElbourneJoanne AshcroftCarys M. JonesKathryn Rowan

Informed consent in clinical trials in criticalcare: experience from the PAC-Man Study

Received: 19 January 2006Accepted: 26 July 2006Published online: 21 September 2006© Springer-Verlag 2006

Electronic supplementary materialSupplementary material is available in theonline version of this article at http://dx.doi.org/10.1007/s00134-006-0358-4 andis accessible for authorized users.

The authors wrote this article on behalf ofthe PAC-Man Study Collaboration

This article is discussed in the editorialavailable at: http://dx.doi.org/10.1007/s00134-006-0359-3

Ethics approval: The PAC-Man trial(Application reference numberMREC/00/2/83) was approved by the NorthThames Multi-centre Research EthicsCommittee on 12 March 2001, as well as allrelevant Local Research Ethics Committees,and Research & Development Departmentsat collaborating hospitals.

Funding: The PAC-Man trial (HTA ProjectNo. 97/08/03) was funded by the UKNational Health Service Research &Development Health TechnologyAssessment Programme. (Disclaimer: Theviews and opinions expressed in this articledo not necessarily reflect those of theNational Co-ordinating Centre for HealthTechnology Assessment.) The NHSResearch & Development HealthTechnology Assessment Programme had noinvolvement in the design, conduct andanalysis of the study.

Competing interest statement: All authorsdeclare no conflict of interest.

Contributors: S.E. Harvey conceived anddesigned the study and participated in the

analysis and interpretation of data and inwriting the paper. D. Elbourne participatedin the analysis and interpretation of data andin writing the paper. J. Ashcroft and C.M.Jones collected the data and participated inthe interpretation of data and in writing thepaper. K. Rowan conceived and designedthe study and participated in theinterpretation of data and in writing thepaper. S.E. Harvey is the guarantor of thepaper and accepts full responsibility for theconduct of the study, had access to the data,and controlled the decision to publish.

S. E. Harvey · K. RowanIntensive Care National Audit & ResearchCentre, Tavistock House,Tavistock Square, WC1H 9HR, London,UK

D. Elbourne (!)London School of Hygiene & TropicalMedicine, Medical Statistics Unit,Keppel Street, WC1E 7HT, London, UKe-mail: [email protected].: +44-20-79272629Fax: +44-20-76372853

J. AshcroftUniversity of Manchester, UK MentalHealth Research Network,PR2 8DY, Preston, UK

C. M. JonesRoyal Berkshire Hospital, Intensive CareUnit,RG1 5AN, Reading, Berkshire, UK

Abstract Objectives: To identifythe proportion of critically ill patientsable to consent to participation ina randomised controlled trial (RCT)

and to assess to what extent patientconsent and relative assent processescould be conducted according toethics committee permissions. De-sign: Descriptive study nested in anRCT. Setting: Fifty-six UK intensivecare units participating in the PAC-Man trial. Patients and participants:First 500 patients consecutively en-rolled into PAC-Man. Measurementand results: The outcome measureswere patient consent and/or relativeassent. Of the 498 patients included,13 (2.6%) provided consent beforerandomisation. Of the remaining 485patients, relative assent was obtainedfor 394 patients (81.2%), and refusedpost-randomisation for 3 patients(0.6%). No relatives were availablefor 15 patients (3.1%), and it wasunclear from documentation whetherrelative assent had been obtainedfor 73 patients (15.1%). Of the 482patients who did not provide consentprior to randomisation, 188 (39%)survived. Of these, 175 (93.1%) gaveretrospective informed consent, six(3.2%) refused, and seven (3.7%) didnot regain mental competency. Con-clusions: A very small proportionof patients were able to give consentbefore randomisation. Due to the highin-hospital mortality (60.6%), onlyaround one third of the remainingpatients could provide consent retro-spectively. This study demonstratesdifficulties experienced in obtainingconsent from critically ill patients

Intensive Care Med (2006) 32:2020–2025

2022

Patient consent and relative assent

Two patients (0.4%) were randomised in error. Of theremaining 498 patients, only 13 (2.6%) were able togive informed consent prior to randomisation (Fig. 1).Of the remaining 485 patients, agreement from a relativewas obtained in writing for 312 (64.3%) and orally for82 (16.9%) either prior to randomisation or as soon aspossible afterwards. Most frequently, relative assent wasprovided by the patient’s partner (44.4%), or by a sonor daughter (27.9%). In only three cases (0.6%), did therelatives refuse assent following randomisation. Thesepatients were withdrawn. In 15 cases (3.1%), the patienthad no relatives or next-of-kin who could be asked forassent. For 73 patients (15.1%) documented evidence wasunavailable in the ICU study files (Fig. 1). The mediantime between admission to the ICU and randomisation did

Fig. 1 Patient consent and/orrelative agreement processaIn 12 cases a form had beencompleted but no copy wasavailablebIn 1 case a form had beencompleted but no copy wasavailable; in 13 cases oralagreement was obtained but noform completed; in 4 cases oralagreement documented in thehospital notescIn 1 case a form had beencompleted but no copy wasavailable; in 1 case patientconsent obtained orally anddocumented in the hospital notes

not differ significantly (Wilcoxon rank-sum test: p = 0.43)between three groups of patients: the 13 patients who gaveconsent pre-randomisation (14.5 h, interquartile range0.6–22.7); the 393 with relative assent (14.7, 4.9–33.3);and the 85 without relative assent (12.8, 5.2–40.4).

Retrospective patient consent

Of the 482 patients unable to give informed consent priorto randomisation, 188 (39.0%) survived; of these, 181(96.3%) regained mental competency. Of the 181, six(3.3%) refused consent retrospectively and were with-drawn. Retrospective patient consent varied by relativeagreement (0.7% where relatives agreed versus 12% withno relative agreement; p = 0.003). The seven patients whodid not regain mental competency were included in the

2022







2022







3,3% refused consent retrospectively0,7% refused consent despite prior consent of relatives


Waiver of (prior) consentEffect on recruitment

Annane et al. Intensive Care Med 2004;30:321-324 § JAMA 2002;288:862-871

‣Randomization within 3h from start of septic shock

‣7 months: inclusion rate 4 /mo (expected 10p /mo)

‣300 patients randomized

• 10 patients gave informed consent (10%)

• 70 relatives gave informed consent (23%)

• 220 no informed consent (74%) before start of the study (following amendment for waiver of consent)

• Survivors asked to sign informed consent afterwards


Waiver of (prior) consentEffect on recruitment

Emerg Med J 2004;21:703

‣ CRASH trial: effect of steroids in head injuries § 4000 patients randomized

‣ 78 hospitals: waived consent - 38 hospitals: consent required

***

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

1

Time (h±SE)Recruitment /mo

***

Time to randomisation No randomisations /mo

waived

waived


Ethics of Deferred (Proxy) Consent

2. What to do with the data when the patient dies before consent


Intensive Care Med (2007) 33:894–900DOI 10.1007/s00134-007-0580-8 LEGAL AND ETHICAL ISSUES IN CLINICAL RESEARCH

T. C. JansenE. J. O. KompanjeC. DrumlD. K. MenonC. J. WiedermannJ. Bakker

Deferred consent in emergency intensive careresearch: what if the patient dies early?Use the data or not?

Received: 18 January 2007Accepted: 5 February 2007Published online: 7 March 2007© Springer-Verlag 2007

T. C. Jansen · E. J. O. Kompanje (!) ·J. BakkerErasmus MC University Medical Center,Department of Intensive Care,P.O. Box 2040, 3000 CA Rotterdam,The Netherlandse-mail: [email protected]

C. DrumlMedical University of Vienna, EthicsCommittee of the Medical University ofVienna and the Vienna General Hospital,Borschkegasse 8b, 1090 Vienna, Austria

D. K. MenonUniversity of Cambridge, Division ofAnaesthesia, Addenbrooke’s Hospital,CB2 2QQ Cambridge, UK

C. J. WiedermannCentral Hospital of Bolzano, Department ofMedicine l,Böhler Street 5, 39100 Bolzano, Italy

Introduction

Respect for individual autonomy, expressed in the conceptof informed consent, is the basic ethical principle in re-search with humans. Many ICU patients are unable to giveconsent as a consequence of mental incapacity, and this canbe further complicated in emergency situations, in whichtreatment needs to be initiated without delay. Various ap-proaches are used as surrogate to subject consent: waiverof consent, consent by an independent physician and de-ferred consent. Deferred consent involves randomizationat the investigator’s discretion according to criteria thathave been explicit during ethical review of the protocol,followed by the request for patient’s (deferred subject con-sent) or representative’s (deferred proxy consent) informedconsent in a later phase. Several emergency trials have useddeferred consent [1, 2, 3, 4].

During the enrolment process in an ongoing Dutchmulti-centre randomized controlled trial using deferredconsent the situation arose that no deferred (subject orproxy) consent was obtained from patients who died earlyafter start of the study. Should data of these patients beused or not? In this article we analyse this practical andethical problem.

The “Early Lactate-Directed Therapy on the ICU”study as an example

To evaluate the efficacy of early lactate-directed therapytwo of the authors (T.C.J., J.B.) are currently conductinga multi-centre trial. Patients eligible for inclusion are ran-domly allocated to either 8 h of early lactate-directed ther-apy or control group therapy. Since early timing of goal-directed therapy is essential [5, 6], patients are random-ized immediately after the first available lactate level, re-sulting in a very short inclusion time window. The EthicsCommittee approved the use of deferred consent, referringto the Dutch revised “Medical Research in Human Sub-jects Act” [7]. Study procedures are temporarily under-taken without consent, and, as soon as possible, writtenconsent from the patient or legal representative is sought.

Until December 2006 we had collected data from115 patients (Fig. 1). In 11.7% of cases (13/111) consentcould not be obtained due to early death (< 72 h, beforerelatives could be approached). Given the predicted sam-ple size of 350 patients, not using data of these patientswould result in an additional requirement of 41 patients.In 2.7% of cases (3/111) relatives could not be identified,or contact was lost. In 2.7% (3/111) relatives refused

1. Introduces selection bias

2. Violates the intention-to-treat principle of the primary analysis

3. Validity of proxy consent in these situations is ethically questionable

4. Very few patients/relatives object to the use of already obtained data in emergency situations

5. Using the data will not “harm” the patient or relatives if confidentiality and privacy are assured

6. Jeopardizing results of the study might harm future patients and society

7. Devalues the “contribution” made by the subjects entered with consent

8. Confronting bereaved relatives with consent procedure represents additional burden

9. The individual’s decision about the privacy of their medical information is not absolutely binding if the processing is necessary and proportionate for ‘the protection of health’, if sufficient safeguards apply or if it is necessary and proportionate for the goals of medical research


Study completed before consent obtained

‣ Study on the effect of early goal (lactate) directed therapy following ICU admission

• therapeutic protocol lasted for 8h

• study procedures for 72h following start of the study

‣ Deferred consent accepted by the Ethical Committee

• 10 % of the patients died before consent could be obtained

• In another 5% of the patients consent could not be obtained within 72 hours

Jansen et al. Intensive Care Med 2007;33:894


Cox’s proportional hazards analysis Stratified by: center and sepsis group Adjusted for co-variables: age, sex, APACHE II and SOFA

Table 1. Mortality with and without patients with missing deferred consent

Control

group

Lactate

group

Relative risk

(95%CI)

P

Value

In-hospital mortality: unadjusted analysis - % (n)

All patients (n=348) 43.5% (77/177) 33.9% (58/171) 0.78 (0.60-1.02) 0.067

Excluding early death

(n=309)

33.8% (51/151) 28.5% (45/158) 0.84 (0.60-1.18) 0.32

Excluding all missing

deferred consent (n=289)

33.1% (46/139) 28.7% (43/150) 0.87 (0.61-1.22) 0.42

In-hospital mortality: adjusted analysis - hazard ratio (95% CI)

All patients (n=348) 0.61 (0.43-0.87) 0.006

Excluding early death

(n=309)

0.80 (0.53-1.21) 0.28

Excluding all missing

deferred consent (n=289)

0.81 (0.53-1.25) 0.35

chi-square test

Tim C. JansenJan BakkerErwin J. O. Kompanje

Inability to obtain deferred consent dueto early death in emergency research:effect on validity of clinical trial results

Received: 3 February 2010Accepted: 14 June 2010Published online: 6 August 2010! The Author(s) 2010. This article ispublished with open access atSpringerlink.com

T. C. Jansen ! J. Bakker !E. J. O. Kompanje ())Department of Intensive Care, Erasmus MCUniversity Medical Center, PO Box 2040,3000 CA Rotterdam, The Netherlandse-mail: [email protected]

Abstract Purpose: To illustratethe impact on the validity of trialresults due to excluding patients froma randomized controlled trial forwhom no deferred consent could beobtained after randomization becausestudy procedures had already beenfinished. Methods: The unadjustedand adjusted primary outcome mea-sures of a recent randomizedcontrolled multicentre study in thefield of intensive care medicine werecompared, including (n = 348) orexcluding (n = 289) patients withmissing deferred consent. Results:Thirty-nine patients (11%) died early,before the patient or his/her proxycould be approached and consent beobtained. In another 20 patients (6%),it was not possible to inform proxies

and ask consent within the period ofstudy procedures. A significant treat-ment effect (p = 0.006) in theadjusted analysis became non-signif-icant (p = 0.35) when the patientswith missing deferred consent wereexcluded. Conclusions: Exclusionof patients without obtained deferredconsent can reduce statistical power,introduce selection bias, make ran-domization asymmetrical, decreaseexternal validity and thereby jeopar-dize study results. This may haveimplications for emergency researchin various disciplines.

Keywords Deferred consent !Ethics ! Clinical trial !Informed consent ! Emergency

Introduction

Respect for individual autonomy, expressed in the con-cept of informed consent, is a basic ethical principle inresearch with humans. Many critically ill patients areunable to give consent as a consequence of mentalincapacity, and this can be further complicated inemergency situations, in which treatment needs to beinitiated without delay. Proxies are not always availablein the first hours of hospital or intensive care admissionor are too overwhelmed to understand the providedinformation to give valid consent. Deferred consent is anacceptable substitute in these emergency circumstances[1]. However, clinicians and investigators may encoun-ter an important practical and ethical problem afterenrolling patients under deferred consent: should the

researcher use the study data if study procedures havealready been finished before it was possible to informthe patient or his/her proxy and ask consent? Thisincludes the situation in which the patient has died early.

The intention-to-treat principle implies that the pri-mary analysis should include all subjects. Compliancewith this principle would necessitate complete follow-upof all subjects for study outcomes. As patients who dieearly are the most severely ill, excluding them couldreduce external validity, jeopardize the balance betweenstudy arms and influence the effect of the intervention[1, 2].

In an earlier discussion paper, we argued that deferredproxy consent is the preferable substitute for informedconsent in emergency critical care research. In case of thesituation when study procedures are finished or the patient

Intensive Care Med (2010) 36:1962–1965DOI 10.1007/s00134-010-1988-0 LEGAL AND ETHICAL ISSUES

Tim C. JansenJan BakkerErwin J. O. Kompanje

Inability to obtain deferred consent dueto early death in emergency research:effect on validity of clinical trial results

Received: 3 February 2010Accepted: 14 June 2010Published online: 6 August 2010! The Author(s) 2010. This article ispublished with open access atSpringerlink.com

T. C. Jansen ! J. Bakker !E. J. O. Kompanje ())Department of Intensive Care, Erasmus MCUniversity Medical Center, PO Box 2040,3000 CA Rotterdam, The Netherlandse-mail: [email protected]

Abstract Purpose: To illustratethe impact on the validity of trialresults due to excluding patients froma randomized controlled trial forwhom no deferred consent could beobtained after randomization becausestudy procedures had already beenfinished. Methods: The unadjustedand adjusted primary outcome mea-sures of a recent randomizedcontrolled multicentre study in thefield of intensive care medicine werecompared, including (n = 348) orexcluding (n = 289) patients withmissing deferred consent. Results:Thirty-nine patients (11%) died early,before the patient or his/her proxycould be approached and consent beobtained. In another 20 patients (6%),it was not possible to inform proxies

and ask consent within the period ofstudy procedures. A significant treat-ment effect (p = 0.006) in theadjusted analysis became non-signif-icant (p = 0.35) when the patientswith missing deferred consent wereexcluded. Conclusions: Exclusionof patients without obtained deferredconsent can reduce statistical power,introduce selection bias, make ran-domization asymmetrical, decreaseexternal validity and thereby jeopar-dize study results. This may haveimplications for emergency researchin various disciplines.

Keywords Deferred consent !Ethics ! Clinical trial !Informed consent ! Emergency

Introduction

Respect for individual autonomy, expressed in the con-cept of informed consent, is a basic ethical principle inresearch with humans. Many critically ill patients areunable to give consent as a consequence of mentalincapacity, and this can be further complicated inemergency situations, in which treatment needs to beinitiated without delay. Proxies are not always availablein the first hours of hospital or intensive care admissionor are too overwhelmed to understand the providedinformation to give valid consent. Deferred consent is anacceptable substitute in these emergency circumstances[1]. However, clinicians and investigators may encoun-ter an important practical and ethical problem afterenrolling patients under deferred consent: should the

researcher use the study data if study procedures havealready been finished before it was possible to informthe patient or his/her proxy and ask consent? Thisincludes the situation in which the patient has died early.

The intention-to-treat principle implies that the pri-mary analysis should include all subjects. Compliancewith this principle would necessitate complete follow-upof all subjects for study outcomes. As patients who dieearly are the most severely ill, excluding them couldreduce external validity, jeopardize the balance betweenstudy arms and influence the effect of the intervention[1, 2].

In an earlier discussion paper, we argued that deferredproxy consent is the preferable substitute for informedconsent in emergency critical care research. In case of thesituation when study procedures are finished or the patient

Intensive Care Med (2010) 36:1962–1965DOI 10.1007/s00134-010-1988-0 LEGAL AND ETHICAL ISSUES


Deferred consent

‣When the patient dies before consent, ask the family?

• proxy represents the patient during treatment however this ends when the patient dies and thus their autonomy is limited but not absent

✓ information of these proxies is usually indicated

✓ informing the proxies and asking consent may introduce selection bias

• the consent in research is focused on participation (risks, harm, benefit etc) not so much on use of data

• patients have a legal right to see their data however proxies do not have patients’ rights after the death of the patient and so no legal right to see research data

• proxies do not have to consent on the use of data at this point


Deferred Proxy Consentpractice

LactateInterventional

n=362

NGALObservational

n=550

Total

n=856

Deferred Proxy Consent

Deferred Proxy Refusal

No consent

273 (75.4%) 494 (89.8%) 767 (89.6%)

14 (3.9%) 4 (0.7%) 18 (2.1%)

75 (20.7%) 52 (9.4 %) 127 (14.8%)

Crit Care 2009;13(suppl 1):S196 - Poster 488


Recommendations

‣ Waiver and deferred (proxy) consent has an ethical basis in emergency research

‣ Approach the relatives with information about the research and ask for consent when this is ethically valid to do so

‣ Study data can be used before consent can be obtained (including the situation the patients has died) when sufficient privacy measures are applied

• time limit of 72h (or use of independent physician)

• inform general practitioner

• report to local ethical committee


Deferred proxy consent in emergency critical care research: Ethically valid and practically feasibleTim C. Jansen et al. Crit Care Med 2009;37:S65-S68

wait until present

stop studyno use of data

YESNO

Proxy present or available within appropriate time frame?

NOYES

ethically valid to inform and ask for consent?

start study procedures

NO

wait until valid to inform and ask for consent

continue studyuse data

study procedures already finished or patient died?

YESinform proxy and ask for consent a.s.a.p.

OBTAINED

don't ask for consentinform proxy

use data

REFUSED

> 72h


wait until present

stop studyno use of data

YESNO

Proxy present or available within appropriate time frame?

NOYES

ethically valid to inform and ask for consent?

start study procedures

NO

wait until valid to inform and ask for consent

continue studyuse data

study procedures already finished or patient died?

YESinform proxy and ask for consent a.s.a.p.

OBTAINED

don't ask for consentinform proxy

use data

REFUSED

> 72h


Documents

Deferred consent - Erasmus Critical Care · Deferred consent Prof. Jan Bakker MD, PhD Chair dept Intensive Care Adults [email protected] Thursday, March 22, 12