The quality of informed consent: mapping the landscape. A review of empirical data from developing and developed countries

Abstract

Objective

Some researchers claim that the quality of informed consent of clinical research participants in developing countries is worse than in developed countries. To evaluate this assumption, we reviewed the available data on the quality of consent in both settings.

Methods

We conducted a comprehensive PubMed search, examined bibliographies and literature reviews, and consulted with international experts on informed consent in order to identify studies published from 1966 to 2010 that used quantitative methods, surveyed participants or parents of paediatric participants in actual trials, assessed comprehension and/or voluntariness, and did not involve testing particular consent interventions. Forty-seven studies met these criteria. We compared data about participant comprehension and voluntariness. The paucity of data and variation in study methodology limit comparison and preclude statistical aggregation of the data.

Results and Discussion

This review shows that the assertion that informed consent is worse in developing countries than in developed countries is a simplification of a complex picture. Despite the limitations of comparison, the data suggest that: (1) comprehension of study information varies among participants in both developed and developing countries, and comprehension of randomisation and placebo controlled designs is poorer than comprehension of other aspects of trials in both settings; and (2) participants in developing countries appear to be less likely than those in developed countries to say they can refuse participation in or withdraw from a trial, and are more likely to worry about the consequences of refusal or withdrawal.

Introduction

Many prospective research participants in developing countries have little formal education, lack familiarity with biomedical research and consent procedures, and have limited access to healthcare services. Consequently, it is widely believed that they have more difficulty comprehending study information and providing voluntary consent than do their counterparts in developed countries.^1–11 Such views are echoed in ethics guidelines such as those of the Council for International Organizations of Medical Sciences (CIOMS),¹² in a report by the National Bioethics Advisory Commission,¹³ and in the popular press. For instance, a front-page New York Times article framed the problems with comprehension in a trial in the Ivory Coast as a matter of an impenetrable wall between scientific complexity and the ability of locals to understand it—one participant was described as “still not grasp [ing]—even after repeated questioning—what a placebo is or why she might have been given that instead of a real medicine”.¹⁴

But what do we know about the quality of informed consent in developing country research? Does available evidence demonstrate that the quality of informed consent from developing country participants is worse than the quality of informed consent from participants in developed nations?

To begin addressing these questions, we reviewed and compared available data on the quality of informed consent from research in both developing and developed countries. We identify similarities and differences between studies of consent in developed and developing countries, highlight gaps in the available data, and make recommendations for future research on the quality of informed consent.

Methods: search strategy and selection criteria

We conducted a comprehensive PubMed search using the Medical Subject Headings (MeSH) terms informed consent, comprehension and decision making in combination with clinical trials or randomized controlled trials (box 1). In addition, we examined bibliographies,¹⁵ literature reviews^16,17 and reference lists from relevant papers, and consulted with international experts on informed consent to clinical research.

Box 1. MeSH terms strategy.

(informed consent[mh] AND (Comprehension[mh] OR decision-making[mh])

AND (randomized controlled trials as topic[mh] OR clinical trial as topic[mh])

AND (Humans[Mesh] AND English[lang]))

We included studies that met four criteria: (1) used quantitative methods to study informed consent (to allow for comparison of relatively similar data sets); (2) surveyed participants or paediatric participants' parents in actual clinical trials rather than hypothetical scenarios (as we are concerned with what participants understand and how they make decisions in real trials); (3) did not test informed consent interventions aimed at improving its quality (to avoid confounding results); and (4) assessed at least one of two domains critical to measuring the quality of informed consent: comprehension of study information and voluntariness of consent. While some published data on disclosure exist, there are little to no comparable data from non-intervention utilising trials that evaluate understanding and comprehension relative to the quality of disclosure. A total of 427 studies were identified through the PubMed search (figure 1) and 79 from bibliographies, literature reviews, reference lists and consultations with experts. Of those 506 studies, 47 met all four criteria: 18 studies evaluated the quality of informed consent in trials in developing countries and 32 studies evaluated the quality of informed consent in trials in developed countriesⁱ (tables 1 and 2). Identified studies were reviewed by the authors and information extracted regarding the type and location of the clinical trial, the sample size, and the method and timing of assessing informed consent. Data about participants' comprehension of trial information and voluntariness were extracted, including understanding of the purpose and nature of the research, the risks and side effects, and randomisation and placebo controlled design (tables 3–5), as well as perceived pressure and participant knowledge of the right to refuse to enrol or withdraw from a trial (table 6). Direct comparison or meta-analysis of study data was not feasible, as the relevant studies did not employ a uniform methodology or study design.

Figure 1.

PRISMA 2009 flow diagram (adapted from Moher D, Liberati A, Tetziaff J, et al; The PRISMA Group. Preferred reporting items for systemic reviews and meta analyses: the PRISMA statement. PLoS Med 2009;6: e1000097. For more information, visit http://www.prisma-statement.org). This figure is produced in colour in the online journal—please visit the website to view the colour figure.

Table 1. Developing country consent studies.

Authors	Country	Sample	Type of clinical research	Method of evaluation
Ellis et al, 201018	Mali	89 M and F	Malaria vaccine phase 1 trial	Questionnaire administered after IC document reviewed but before consent
Vallely et al, 201020	Tanzania	99 F	Placebo controlled trial of HIV vaginal microbicide	Interviews at 4, 24 and 52 weeks
Sarkar et al, 200921	India	368 Parents	Birth cohort study of diarrhoeal disease	Structured interviews 3–7 months post-trial
Oduro et al, 200822	Ghana	270 Mothers	Paediatric trials evaluating immune correlates of protection against malaria	Questionnaire administered at end of study
Hill et al, 200823	Ghana	60 F	Vitamin A supplementation trial	Semi-structured interviews after consent
Minnies et al, 200824	South Africa	192 Mothers	Paediatric case–control trial of immune correlates against severe childhood TB	Self-administered questionnaire with staff help if necessary, within 1 h of consent
Kaewpoonsri et al, 200625	Thailand	84 M and F	Malaria drug trials	Interview at third follow-up visit
Marshall et al, 200619	Nigeria	307 M and F	Genetic studies of hypertension	Interviews administered at variable times usually long after consent
Krosin et al, 200626	Mali	163 Parents	Paediatric malaria vaccine prevention trial	Questionnaire within 48 h after consent
Moodley et al, 200527	South Africa	334 M and F	Influenza vaccine trial	Interviews 4–12 months post-trial
Pace et al, 200528	Thailand	141 M and F	HIV study of IL-2 effectiveness	Interviewers administered survey immediately after consent
Pace et al, 200529	Uganda	347 Parents	Paediatric malaria treatment study	Interviews immediately after consent
Ekouevi et al, 200430	Côte d'Ivoire	55 F	HIV mother-to-child transmission prevention trial	Interviews a median of 136 days after consent
Joubert et al, 200331	South Africa	92 F	Trial of vitamin A for prevention of mother-to-child HIV transmission	Interviews a median of 14 months after consent
Lynöe et al, 200132	Bangladesh	105 F	Nutritional trial of iron supplements for pregnant women	Interviews after consent
Leach et al, 199933	The Gambia	137 Mothers	Paediatric trial of Haemophilus influenzae type B conjugate vaccine	Interviews within a week of consent
Karim et al, 199834	South Africa	Evaluation study group: 56 F	Perinatal HIV transmission trial	Questionnaires administered before and after counselling and consent
		Sensitisation control group: 56 F*
Pitisuttithum et al, 199735	Thailand	33 M and F	HIV vaccine trial with drug users	Questionnaire before signing consent

^*To evaluate the informed consent obtained for the HIV testing that preceded induction into the perinatal transmission trial, researchers administered both pre- and post-counselling questionnaires to an evaluation study group (n=56). A sensitisation control group (n=56) received only post-counselling questionnaires, so as to measure the sensitising effect of the pre-counselling questionnaire given to the evaluation study group.

F, female; IC, informed consent; M, male.

Table 2. Developed country consent studies.

Authors	Country	Sample	Type of clinical research	Method of evaluation
Ellis et al, 201018	USA	171 M and F	Malaria vaccine phase I trial	Questionnaire administered after IC document reviewed but before consent
Ravina et al, 201036	USA	149 M and F	Phase II Parkinson's trial	Self-administered questionnaire at final clinical trial visit
Bergenmar et al, 200837	Sweden	282 M and F	Phase II and phase III oncology trials	Mail surveys sent within 3 days–2 weeks of consent
Knifed et al, 200838	Canada	21 M and F	Neuro-oncology trial	Interviews within 1 month of IC
Agrawal et al, 200639	USA	163 M and F	Phase I oncology trials	Interview immediately after consent
Franck et al, 200740	UK	109 Parents	25 Different paediatric studies	Questionnaire taken immediately after and 3 months after consent
Marshall et al, 200619	USA	348 M and F	Genetic studies of hypertension	Interviews long and variably after consent
Gammelgaard et al, 200441	Denmark	103 M and F	Acute myocardial infarction trials	Mail survey sent to participants in the study 3 weeks after IC
Kodish et al, 200442	USA	137 Parents	Paediatric leukaemia trial	Parent pairs interviewed within 48 h of consent
Lynöe et al, 200443	Sweden	44 M and F	Chronic haemodialysis trials	Mail survey about 1 week after disclosure of information
Criscione et al, 200344	USA	30 M and F	Rheumatoid arthritis trial	Questionnaire 1–4 weeks after consent
Kupst et al, 200345	USA	20 Parents	Paediatric oncology trials	Interviews 1 month after IC
Pope et al, 200346	Canada	190 M and F	Cardiology, ophthalmology and rheumatology trials	Mail survey 2–5 months after consent
Schats et al, 200347	The Netherlands	37 M and F	Subarachnoid haemorrhage emergency management trials	Interviews 7–31 months after IC (median of 20 months)
Simon et al, 200348	USA	Majority English speakers: 60 parents	Paediatric oncology trials	Parents interviewed 48 h after consent
		Minority English Speakers: 27 parents
		Minority non-English speakers: 21 parents
Joffe et al, 200149	USA	207 M and F	Oncology trials, phase I, II and III	Mail survey 1–2 weeks after consent
Daugherty et al, 200050	USA	144 M and F	Phase I oncology trials	Interviews within 1 week of first administration of investigational treatment
Hietanen et al, 200051	Finland	261 F	Oncology trial of tamoxifen	Mail survey 5–17 months after consent
Montgomery et al, 199852	UK	158 M and F	3 In-house and 3 multi-centre anaesthesia trials	Mail survey up to 24 months after consent
Van Stuijvenberg et al, 199853	The Netherlands	181 Parents	Paediatric trial of ibuprofen for febrile convulsions	Mail survey up to 2–3 years after consent
ACHRE, 199654	USA	570 M and F	Oncology and cardiology trials	Brief interviews followed by in-depth interviews
Harrison et al, 199555	USA	71 M and F	HIV vaccine trial	Self-administered questionnaire after disclosure and before consent
Harth et al, 199556	Australia	62 Parents	Paediatric trial of oral asthma drug	Self-administered questionnaire 6–9 months after entered trial
Estey et al, 199457	Canada	29 M and F	Not specified	Interviews 1–6 weeks after consent
Miller et al, 199458	USA	168 M and F	Trial of analgesic drugs	Interviews 30–90 days after entered trial
Lynöe et al, 199159	Sweden	43 F	Gynaecology trial of antiphlogistic drugs for fallopian tube inflammation	Mail survey 18 months after study
Benson et al, 198560	USA	Depression study: 24 M and F	Antidepressant trial and antipsychotic trial	Interviews immediately following IC
		Schizophrenia study: 24 M
Penman et al, 198461	USA	144 M and F	Oncology trials, phase II and III	Interviews 1–3 weeks after consent
Riecken et al, 198262	USA	112 M*	50 Different trials	Interviews within 10 weeks of consent
Howard et al, 198163	USA	64 M and F	Cardiology trial of β-blockers (BHAT) for acute myocardial infarction	Interviews 2 weeks–15 months after consent
Bergler et al, 198064	USA	39 M	Hypertension trial of hydrochlorothiazide versus propranolol	Interviews and quizzes just after consent; repeated 3 months later

^*The trial involved 156 participants, but only 112 indicated that they were aware that they were participating in a trial, and therefore only 112 were asked questions about voluntariness. ACHRE, Advisory Committee on Human Radiation Experiments; F, female; IC, informed consent; M, male.

Table 3. Understanding of research nature and purpose^*.

Developed country studies			Developing country studies
Author	Country	Understood purpose	Author	Country	Understood purpose
Knifed et al, 200838	Canada	100%	Leach et al, 199933	The Gambia	90%
Ravina et al, 201036	USA	92.6%	Pace et al, 200528	Thailand	88%
Franck et al, 200740	UK	85%	Minnies et al, 200824	South Africa	80.6%
Howard et al, 198163	USA	80%	Pace et al, 200529	Uganda	80%
Miller et al, 199458	USA	73%	Kaewpoonsri et al, 200625	Thailand	50%
Van Stuijvenberg et al, 199853	The Netherlands	53%	Sarkar et al, 200921	India	43%
Marshall et al, 200619	USA	41%	Marshall et al, 200619	Nigeria	39%
Benson et al, 198560	USA	37% each in depression and schizophrenia studies	Krosin et al, 200626	Mali	26%
Daugherty et al, 200050	USA	31%	Joubert et al, 200331	South Africa	28%, but 40% knew the substance being tested was vitamin A
Harth et al, 199556	Australia	13%
Riecken et al, 198262	USA	10%
Developed country studies			Developing country studies
Author	Country	Understood nature	Author	Country	Understood nature
Knifed et al, 200838	Canada	100%	Ekouevi et al, 200430	Côte d'Ivoire	95%e100%
Criscione et al, 200344	USA	100%	Moodley et al, 200527	South Africa	95%
Hietanen et al, 200051	Finland	100%	Minnies et al, 200824	South Africa	85.4% knew was research; 36.7% knew there were no immediate benefits
Lynöe et al, 199159	Sweden	98%	Vallely et al, 201020	Tanzania	77% knew gel may not prevent HIV
Howard et al, 198163	USA	92%	Hill et al, 200823	Ghana	75% knew was research, but 93% thought trial capsules were a ‘medicine or vitamin’
Ravina et al, 201036	USA	89% understood drugs were experimental†	Lynöe et al, 200132	Bangladesh	47%
Ellis 201018	USA	85%	Krosin et al, 200626		26%
Penman et al, 198461	USA	78%
Gammelgaard et al, 200441	Denmark	72%
Riecken et al, 198262	USA	72%
Bergenmar et al, 200837	Sweden	∼70%
Kupst et al, 200345	USA	55%
Schats et al, 200347	The Netherlands	38%
Joffe et al, 200149	USA	30%

^*Arranged from highest to lowest %.

^†Yet only 57% knew that participation in the study was not part of usual Parkinson's disease treatment.

ACHRE, Advisory Committee on Human Radiation Experiments.

Table 5. Understanding of study design and randomisation^*.

Developed country studies			Developing country studies
Author	Country	Understood study design	Author	Country	Understood study design
Van Stuijvenberg et al, 199853	The Netherlands	88% placebo design	Moodley et al, 200527	South Africa	49% knew they had a 50% chance of receiving placebo; 19% understood placebo
Criscione et al, 200344	USA	87% placebo design	Hill et al, 200823	Ghana	13% understood ‘not all trial capsules were the same’
Howard et al, 198163	USA	86% double blind design	Leach et al, 199933	The Gambia	10% placebo design†
Harrison et al, 199555	USA	79%
Pope et al, 200346	Canada	76% placebo design
Bergler et al, 198064	USA	64% (at start), 28% (3 months later)
Developed country studies			Developing country studies
Author	Country	Understood randomisation	Author	Country	Understood randomisation
Ravina et al, 201036	USA	90%, yet only 67% understood there was a 1 in 3 chance of receiving placebo	Ellis et al, 201018	Mali	80% of adults 90% of parents
Simon et al, 200348	USA	68% of majority English speakers	Krosin et al, 200626	Mali	68%
		26% of minority English speakers
		14% of minority non-English speakers
Bergenmar et al, 200837	Sweden	85%	Pace et al, 200528	Thailand	31%
Gammelgaard et al, 200441	Denmark	79%	Moodley et al, 200527	South Africa	21%
Criscione et al, 200344	USA	50%	Pace et al, 200529	Uganda	19%
Van Stuijvenberg et al, 199853	The Netherlands	50%
Kodish et al, 200442	USA	50%
Howard et al, 198163	USA	43%
Pope et al, 200346	Canada	39%
Benson et al, 198560	USA	33% of depression study, 16% of schizophrenia study
Hietanen et al, 200051	Finland	23%
Schats et al, 200347	The Netherlands	22%

^*Arranged from highest to lowest %.

^†When broken down into those who had received a written information sheet at least a week before consent, 15% of those who had received the sheet understood that there was a placebo group versus 4% of those who had not received prior written information.

Table 6. Voluntariness: knew could withdraw or refuse^*.

Developed country studies			Developing country studies
Author	Country	Knew could withdraw	Author	Country	Knew could withdraw
Joffe et al, 200149	USA	90%; 99% knew could refuse	Ellis et al, 201018	Mali	96% of adults 93% of parents
Ellis et al, 201018	USA	98%	Karim et al, 199834	South Africa	93% of evaluation study group, 88% of sensitisation control group†
Ravina et al, 201036	USA	98% felt ‘free to refuse to participate’	Pitisuttithum et al, 199735	Thailand	88% knew could refuse
Marshall et al, 200619	USA	97%	Moodley et al, 200527	South Africa	87%
Criscione et al, 200344	USA	96%	Pace et al, 200528	Thailand	71%
Benson et al, 198560	USA	Depression study: 95%; 75% knew could refuse.	Marshall et al, 200619	Nigeria	67%
		Schizophrenia study: 83%; 75% knew could refuse	Minnies et al, 200824	South Africa	65%
Bergenmar et al, 200837	Sweden	93%; 100% knew could refuse	Pace et al, 200529	Uganda	65%; 41% knew could refuse
Franck et al, 200740	UK	91%	Kaewpoonsri et al, 200625	Thailand	53.1%
Van Stuijvenberg et al, 199853	The Netherlands	91%	Lynöe et al, 200132	Bangladesh	48%: 65% knew could refuse
Lynöe et al, 200443	Sweden	90%	Sarkar et al, 200921	India	50%
Simon et al, 200348	USA	90% of majority English speakers	Ekouevi et al, 200430	Côte d'Ivoire	27%
		78% of minority English speakers	Joubert et al, 200331	South Africa	24% (but 92% said care would no longer be good if they quit)
		60% of minority non-English speakers	Oduro et al, 200822	Ghana	21%
Montgomery et al, 199852	UK	83%	Krosin et al, 200626	Mali	10%
Penman et al, 198461	USA	80%
Riecken et al, 198262	USA	80%; 95% knew could refuse
ACHRE, 199654	USA	78%
Bergler et al, 198064	USA	77% (at start), 61% (3 months later)
Harth et al, 199556	Australia	45% (but 32% said would not be allowed)
Schats et al, 200347	The Netherlands	25%; 59% knew could refuse‡

^*Arranged from highest to lowest %.

^†However, 98% of the evaluation study group and 100% of the sensitisation control group said hospital would not allow them.

^‡Knew participation was not obligatory.

ACHRE, Advisory Committee on Human Radiation Experiments.

Results

Study characteristics

Eighteen studies conducted in 11 different developing countries examined the consent of participants in clinical research on vaccines, nutritional supplements, HIV treatments, immune correlates in children, diarrhoeal disease in children, anti-malarial drugs and genetics (table 1). Sample sizes ranged from 33 to 700 research participants. Seven studies interviewed a parent of a participating child,^{18,21,22,24,26,29,33} and of those seven, three interviewed only the mothers.^22,24,33 Thirteen studies^{19–21,23,26–33,35} used structured or semi-structured interviews, while five used questionnaires.^{18,22,24,25,34} In nine studies, participants were interviewed close to the time of consent^{18,23,24,26,28,29,33–35} and in eight others interviews were conducted 1–14 months or longer after the participant gave consent.^{19–22,25,27,30,31} In one study, the timing was not specified.³²

Thirty-one studies, conducted in eight developed countries, examined the consent of participants involved in oncology, cardiology, gynaecology, HIV, analgesics/anaesthesia, neurological, antidepressant, antipsychotic, emergency management, arthritis, paediatric asthma, paediatric febrile convulsion, diabetes, malaria and genetics research (table 2). Sample sizes ranged from 21 to 570 research participants. Six studies surveyed the parents of children in paediatric trials.^{40,42,45,48,54,57} Sixteen studies used structured interviews,^{19,38,39,42,45,47,48,50,54,57,58,60–64} nine used mailed surveys,^{37,41,43,46,49,51–53,59} and six used questionnaires.^{18,36,40,44,55,56} In eight studies, questions were asked close in time to when consent was given, in each case within 48 h of consent^{18,39,40,42,48,55,60,64}; the remaining 23 studies surveyed participants weeks to months after consent.^{19,36–38,41,43–47,49–54,56–59,61–63}

Comprehension and recall of trial information

Participant understanding of research purpose, risks/side effects and design varied substantially across informed consent studies from both developing and developed countries. Across studies, comprehension of trial purpose or nature appeared to be better than comprehension of trial design and randomisation.

Trial purpose and nature

Available data show no substantial difference between participants in developing countries and those in developed countries with respect to their understanding of trial purpose, defined as the goal of a given clinical trial (table 3). In the developed country studies that measured it, understanding of trial purpose ranged from 10% of US males who understood the purpose of a variety of trials they were participating in⁶² to 100% of Canadian participants who understood the purpose of a neurooncology trial.⁴⁵ Understanding of trial purpose in developing country studies also varied, ranging from 26% of Malian parents who understood the purpose of a malaria trial for their children²⁶ to 90% of mothers with children in a paediatric influenza trial in The Gambia.³³ Similarly, reported understanding of trial nature, assessed by participants' understanding that they were participating in research and of the investigational and experimental nature of research interventions, varied from 31% of participants in a US phase 1 oncology trial⁵⁰ to almost 100% of participants in both a Swedish and a Finnish trial,^51,59 and from 47% of women in a Bangladeshi nutritional trial for iron supplements³² to 100% of women in an HIV trial in Côte d'Ivoire.³⁰

Risks/side effects

The percentage of participants who could recognise or name trial side effects and risks also ranged widely among the studies reviewed (table 4). Reported understanding of side effects varied depending on how the questions were framed—more participants were able to recognise side effects from a list than were able to name or explain them in response to open-ended questions. For example, 86% of participants in a US analgesic trial recognised at least one side effect from a list, but only 48% were able to name at least one without the help of a list.⁵⁸ In a US rheumatoid arthritis trial, 30% responded that they knew the trial drugs were not completely safe, but were not asked to recognise or name the specific risks of the drugs.⁴⁴

Table 4. Understanding of risks and side effects^*.

Developed country studies			Developing country studies
Author	Country	Understood risks or side effects	Author	Country	Understood risks or side effects
Daugherty et al, 200050	USA	100% named >1 side effect	Pace et al, 200528	Thailand	98% recognised side effects
Harrison et al, 199555	USA	89% recognised side effects	Pitisuttithum et al, 199735	Thailand	97% recognised side effects
Knifed et al, 200838	Canada	71% knew at least one general risk†	Minnies et al, 200824	South Africa	79.2% knew risks
Benson et al, 198560	USA	Depression study: 62% knew risks	Leach et al, 199933	The Gambia	53% named ≥1 side effect
		Schizophrenia study: 42% knew risks
Howard et al, 198163	USA	61% could name 1 side effect	Oduro et al, 200822	Ghana	20% knew direct risks
Miller et al, 199458	USA	48% named and 86% recognised >1 risk	Pace et al, 200529	Uganda	18% named 1 or more side effects
Ravina et al, 201036	USA	47% knew which drugs had highest risks, 93% knew PD could get better, worse or not change	Krosin et al, 200626	Mali	7% said there were side effects‡
Estey et al, 199457	Canada	41% named >1 risk	Kaewpoonsri et al, 200625	Thailand	6.6% recalled being told of risks
Van Stuijvenberg et al, 199853	The Netherlands	40% knew side effects
Joffe et al, 200149	USA	37% knew research risks
Penman et al, 198461	USA	31% named >3 of 11 risks
Criscione et al, 200344	USA	30% knew there were risks§
Bergler et al, 198064	USA	28% (at start), 3% (3 months later)
Bergenmar et al, 200837	Sweden	18% knew research risks
Schats et al, 200347	The Netherlands	6% knew side effects

^*Arranged from highest to lowest.

^†29% did not recall ANY risks of the trial drug, the rest of the participants could name general risks, and at most participants could name up to four specific risks or side effects.

^‡The question was complex and multi-choice. The correct answer was the only one that included a mention of side effects. However, it also included information about the potential benefits of the medicine (eg, that it could prevent malaria and correct other health problems).

^§But were not asked to name or identify them.

PD, Parkinson's disease.

In consent studies of developing country trials, 79% of participants in a South African vaccine trial knew the risks involved²² and 97% of Thai participants recognised possible side effects of an experimental HIV vaccine,³⁵ yet only 7% of Malian parents recognised that the investigational vaccine being given to their child might have side effects.²⁶

Randomisation and placebo trial design

Understanding of randomisation also varied among participants in both developing and developed country trials, but across all studies, understanding of randomisation was low compared to understanding of other aspects of a trial (table 5). In developed country studies, understanding of randomisation appeared to vary according to how close to actual consent it was measured. For example, 68% of parents understood randomisation when asked within 48 h of consent in US paediatric oncology trials,⁴⁸ and as many as 79% understood randomisation in an HIV vaccine trial when assessed immediately after disclosure.⁵⁵ Yet fewer than half of the participants were reported to comprehend randomisation in six developed country studies in which understanding was assessed months or years after consent.^{44,46,47,51,53,63}

Five developing country studies measured understanding of randomisation (table 5); four of those five measured it within 1 week of consent.^{18,20,26,28,29} Comprehension of randomisation ranged from as high as 90% of parents whose children were enrolled in a malaria vaccine trial in Mali¹⁸ to as low as 19% of parents whose children were enrolled in a malaria treatment trial in Uganda.²⁹

Between 64% and 88% of participants understood the study design in six developed country trials,^{41,44,46,53,63,64} yet only 39% of the participants in a set of Canadian trials recalled their own chance of receiving placebo, and 29% of them “thought that the doctor [had known] what kind of medication they were taking”.⁴⁶ Knowledge of placebo was measured in three developing country studies: 10% of mothers enrolling children in a Gambian trial understood the placebo control design,³³ 13% of Ghanaian trial participants knew that not all trial capsules were the same,²³ and 49% of South African participants knew they had a 50% chance of receiving placebo.²⁷

Although measured infrequently, individuals' understanding of research design diverges from their understanding of how it specifically applies to them. In one Thai HIV treatment trial, 31% correctly responded that half the participants would get the investigational drug, yet 48% said they had a 50/50 chance of receiving it.²⁸ In a Ugandan malaria trial, 19% of parents knew that not all children would receive the same treatment, even though 84% recalled being told about treatment assignment.²⁹ Similarly, in a US rheumatoid arthritis trial, 87% of participants said that some people in the trial would get placebo, but only 50% thought they personally could receive placebo.⁴⁴

Voluntariness

Data on voluntariness is organised into two categories: (1) participants' perceptions of pressure (not reported in a table); and (2) participants' knowledge of the right to refuse or withdraw from participation (table 6).

Pressure

Questions assessing perceptions of pressure differed across informed consent studies—some focused on whether or not participants knew or felt that participation was voluntary, while others asked more specific questions about the source and amount of pressure felt by participants.

Most (90%–99%) participants in a US hypertension trial, a Canadian neuro-oncology trial and UK paediatric trials reported no pressure to participate^19,38,40 or reported that participation was voluntary.⁵² At the same time, 31% of US oncology and cardiology trial participants said that they felt that they had little other choice than to participate,⁵⁴ 25% of parents in a Netherlands paediatric oncology trial indicated that they felt obliged to participate⁵³ and 18% of Danish participants in an acute myocardial infarction trial reported feeling ‘under pressure’, although 70% said the decision was ‘fully theirs’.⁴¹

Five developing country informed consent studies measured general perceptions of pressure and voluntariness. Most mothers (95%) in a Ghanaian paediatric trial²¹ and most participants (99%) in a South African influenza vaccine trial²⁷ said participation was voluntary. Similarly, most parents in an Indian paediatric trial (98%) reported that they joined the study freely without any pressure or compulsion.²⁴ In contrast, in another South African trial, 84% of the evaluation group and 93% of the sensitisation group reported feeling that participation was compulsory.³⁴

In consent studies that distinguished sources of pressure, more trial participants reported feeling pressure from their disease or circumstances than from other people. Although 29% of US phase I and phase II oncology trial participants said that their physician did not actively want them to make their own decision,⁶¹ only 14% in a Swedish gynaecology trial,⁵⁹ 7% in another US oncology trial⁶² and 6% in a set of varied US trials reported feeling pressure from a clinician.³⁹ In the same oncology study in which 7% reported pressure from a clinician and 9% from their families, a full 75% reported pressure due to their progressive cancer.³⁹ In another US paediatric oncology trial, 70% of the parents cited high levels of distress and ‘feeling overwhelmed’ during the consent process.⁴⁵ Few participants in developed country trials reported pressure from anticipated consequences of withdrawing: 98% of UK anaesthesia trials participants,⁵² 86% of Canadian neuro-oncology trial participants³⁸ and 85% of Danish cardiology trial participants⁴¹ knew that refusal to participate would not compromise their care.

In developing country studies, reported pressure from others was also generally low, ranging from 6% of participants reporting pressure from spouses, family or the research team in a Ugandan paediatric malaria treatment trial²⁹ to 26% reporting pressure from village elders in a Malian paediatric vaccine trial.²⁶ Reported pressure came from various sources, for example, from village elders (26%), the research team (12%) and a spouse (7%) in the aforementioned Malian study,²⁶ and from a close friend (15%), a family member (7%) or their doctor (2%) in an HIV treatment trial in Thailand.²⁸ Similarly, in a Ugandan paediatric trial, 15% of parents reported feeling pressure from others, including spouses (6%), family or friends (6%) or the research team (6%), but 58% reported pressure because of their child's illness.²⁹ However, in one Gambian trial, 9% of mothers offered spontaneously and 36% agreed when directly questioned that it would have been hard to refuse participation–some reported feeling group pressure after watching other mothers agree to participate.³³

Participants in developing countries reported pressure from fear of the consequences of withdrawing. Although in one South African trial, 88% said their usual care would not be affected if they refused,²⁷ 87% of participants in a Bangladeshi trial felt that the trial offered such advantages that they couldn't refuse.³² Similarly, 32% of the evaluation study group and 23% of the sensitisation group in a South African perinatal HIV transmission trial thought that care would be compromised if they did not participate,³⁴ and 44% of parents in a paediatric malaria vaccine trial in Mali said they would lose healthcare access if they withdrew.²⁶

Knew they could refuse or withdraw

The clearest differences between respondents in developed and developing country informed consent studies were related to knowledge of the right to refuse to participate in research or to withdraw (table 6). In 15 of 18 developed country studies that measured this, more than 75% of trial participants knew they could withdraw or refuse,^{18,19,36,37,40,43,44,49,52–54,60–62,64} and in 10 of these studies, 90% or more said they could withdraw from research.^{18,19,36,37,40,43,44,49,53,60} In one US paediatric oncology trial, 90% of the majority race English speaking parents, 78% of the minority race English speaking parents and 60% of the minority race non-English speaking parents knew they had a right to withdraw their children from the trial.⁴⁸

In contrast, in five of 15 developing country studies that measured it, less than half of respondents knew they could withdraw from research.^{22,26,30–32} As few as 10% of mothers in Mali knew they could withdraw their child from a malaria vaccine trial at any time,²⁶ and 27% of participants in an HIV trial in Côte d'Ivoire knew they could withdraw at any time.³⁰ However, in some developing country trials a higher percentage of participants knew they could withdraw or refuse, for example 50% of parents in a paediatric diarrhoeal trial²¹ knew they could leave the trial at any time, >90% of adults and parents of children in a Malian malaria vaccine trial¹⁸ knew they could withdraw from the trial and 88% of Thai vaccine participants knew they could ‘refuse to participate at any time’.³⁵ One study of a South African HIV trial¹⁷ reported that 93% of the women knew they had the right to quit, but 98% said they believed the hospital would not allow them to quit.³⁴

Discussion

This is the first comparison of quantitative studies of the quality of informed consent from individuals participating in clinical trials in both developed and developing countries. Our review shows that the assertion that research informed consent is worse in developing countries than in developed countries is an oversimplification of a complex picture of the quality of consent. The quality of informed consent depends on the type and amount of information disclosed, adequate comprehension of trial information, and a voluntary decision to enrol. The existing data, which use comprehension and voluntary decision-making as measures of the quality of consent, do not support a categorical difference between the quality of consent from individuals in developed countries and the quality of consent from individuals in developing countries.

A paucity of data, especially from participants from developing countries, as well as variations in trial type, study methodology, sample size, measures used and timing of data collection relative to obtaining consent, limits comparison and statistical aggregation. Nonetheless, these data suggest certain important trends and point to the need for further research.

Our review highlights the following: (1) comprehension of study information varies among trial participants in both developed and developing countries, and comprehension of randomisation and placebo controlled designs is generally lower than comprehension of other aspects of a trial; (2) research participants report different sources of pressure to enrol, and those in developing countries are less likely than those in developed countries to say they can refuse or withdraw from participation, and more likely to worry about the consequences of refusal or withdrawal.

Data show a range of understanding of trial information in both developed and developing country trials. Individuals across studies tended to know that they were involved in research and often responded correctly to questions about the nature and purpose of the research, yet participants everywhere had more difficulty understanding information about trial design, randomisation and placebo controls. Not only are these methods and concepts unfamiliar to many people, but such methods may be contrary to their expectations or hope for therapeutic benefit, making them more difficult to comprehend. Notably, some studies reveal discrepancies between participants' understanding of what will happen in a trial and how this information will affect them directly. Knowledge of facts and appreciation of those facts are different aspects of understanding, both of which are important to informed consent.⁶⁵ This discrepancy is a challenge for informed consent everywhere, and although few studies attempted to measure it, the present data do not suggest a difference in appreciation between developed and developing country participants.

Second, the data on refusal and withdrawal indicate a troubling trend. Finding it difficult to refuse participation in or withdraw from a trial, feeling pressure to join or stay enrolled in a trial, or worrying about the consequences of withdrawing all relate to the voluntariness of an enrolment decision. Studies which used these measures of voluntariness show that a disquieting number of participants, and more in developing country trials than developed country trials, do not know or do not believe that they can refuse to participate or can withdraw from research. Few studies probed these responses further to explain why participants felt they could not refuse or withdraw. Possible explanations include deference to authority, cultural norms, or a founded or unfounded fear of not being able to access needed care.

Lastly, while investigations of the impact of pressure on voluntariness were limited, overall few research participants report feeling pressured to participate in research, and those that did often felt pressure from their circumstances–such as worsening illness or fear that care would be withdrawn—more than from other people. Participants in developing countries were more likely to report pressure from fear of the consequences of withdrawing, including decreased access to healthcare. These issues merit further study.

Recommendations for future research

These data reveal that there is much to be done to improve the quality of informed consent in both developed and developing countries and that additional research would facilitate definitive conclusions about the quality of informed consent around the world. Currently available evidence regarding the effectiveness of strategies to improve consent is limited.^66–68 Variation in methodology, trial types and populations across studies reviewed raised challenges about how to accurately understand and measure the quality of informed consent. Design and implementation of improvement measures depends on careful attention to, and rigorous delineation of, what the quality of consent entails.

Studies of the quality of informed consent would be greatly enhanced by a core set of validated questions that measure the comprehension and voluntariness of participants at the time of decision-making, and by comparison of participants from similar medically defined groups participating in similar types of research. Studying the quality of consent in multi-national trials, such as was done in one multi-site hypertension study we reviewed,¹⁹ would allow for useful comparisons between developed and developing countries. Additionally, more detailed and comprehensive studies of voluntariness are needed, including investigation of sources of pressure to participate and fears about withdrawal or refusal. Future studies should include detailed investigation of associations between cultural norms and attitudes, and socio-demographic characteristics such as education, literacy and socioeconomic status to better understand the impact of these factors on informed consent in both developed and developing countries. Innovative strategies and rigorous studies are sorely needed to facilitate improvement in informed consent to better satisfy one of the fundamental requirements of ethical research.