IRB: Ethics & Human Research

Understanding, Therapeutic Misconceptions and Perceptions, and Enrollment Decision-Making: A Pediatric Preventive Malaria Trial in Rural Tanzania

This study entailed conducting extensive qualitative interviews of mothers who had been invited to have their infants participate in the Kilimanjaro Intermittent Preventive Treatment of Malaria in Infants (Kili IPTi) trial, designed to assess the efficacy of intermittent preventive malaria chemoprophylaxis for infants. Our study sought to explore whether there was a relationship between the mothers’ understanding of the research and the decision to enroll their infants or to decline. Such empirical data is necessary to address widely voiced concerns that limited understanding of research or exploitive inducement may undermine informed consent among clinical trial recruits in developing countries. The most striking result of the study was that, in general, those who declined to have their children participate in the trial had a better understanding of the trial than those who agreed to have their children participate. Decliners were more likely to understand the experimental nature of the trial and were less likely to confuse research with treatment, the conceptual error often labeled “therapeutic misconception.” Notably, in the context of the Kili IPTi trial, perception of clinical benefit may have been an accurate perception rather than therapeutic misconception because ancillary clinical benefits were provided to research participants. Neither participants nor decliners cited “altruistic” reasons for their decisions, despite presumptions among some scholars that such reasons are ethically the most appropriate motivations for participation in clinical research.

Key words: pediatric clinical trials, informed consent, therapeutic misconception, decision-making, ancillary clinical benefits, preventive malaria chemoprophylaxis

______________________________________________________________________________

Commentators have expressed concern that in resource-poor settings, many individuals decline to participate in research and participants may have low understanding about the study, be under a therapeutic misconception, and be unduly induced to participate.1 However, little empirical investigation has been conducted in these settings about individuals’ understanding of research and decision-making, especially among those who decline to participate. For example, some commentators assume that individuals in resource-poor settings who decline to participate do so because they do not understand the research study. Another assumption is that individuals are mistaken if they believe that a nontherapeutic trial offers them a therapeutic benefit and that the informed-consent process is a mechanism to counter this mistake.2 However, whether expectation of a health benefit is a mistake may depend on the actual benefits obtained, whether intrinsic to the trial design or not. Assumptions are also made about motivations. Current international research ethics norms that originated in affluent countries reflect the view that altruistic, community-oriented motivations should be the basis for enrolling in nontherapeutic research because such research is not designed to help individual participants.3 

Efforts in resource-poor settings to improve informed-consent processes and facilitate understanding about clinical trials should be supported by knowledge of the experiences of individuals recruited to participate, not by outsider presumptions about those experiences. The objective of this study was to explore the relationship between understanding of research and the decision whether to participate in a pediatric malaria chemoprophylaxis clinical trial in a resource-poor setting. Our study adds to a small set of others that empirically investigated research participants’ experiences in research-poor settings.4 We used qualitative methods to explore the understanding, motivations, and experiences of women recruited to enroll their infants in the trial. Our findings highlight challenges to constructing robust informed-consent processes for clinical trials in resource-poor settings and underscores challenges of fairness to research decliners in such a setting. Study results challenge pervasive nonempirical assumptions in the literature.

Context of Our Study: The Kili IPTi Clinical Trial

Our qualitative study explored the understanding and reasons of women who enrolled or declined to enroll their newborns in the Kilimanjaro Intermittent Preventive Treatment of Malaria in Infants (Kili IPTi) trial. We first explain the basic aims and structure of the clinical trial and then evaluate its formal consent process before describing the methods for our study. To avoid confusion, we refer to the Kili IPTi clinical trial as the “trial,” and to our exploration of the experience of the women recruited to enroll in that trial as the “study.”

Kili IPTi trial design. Conducted in association with the Tanzania Government Extended Immunization Program, the Kili IPTi trial was a phase III randomized, double-blind, placebo-controlled clinical trial testing a malaria chemoprophylaxis. It examined the efficacy of mefloquine, sulphadoxine-pyremethamine (SP), and chlorproguanil-dapsone for intermittent preventive treatment of malaria in infants (IPTi). Intermittent preventive treatment refers to a vaccine-like chemoprophylaxis with repeated inoculations of children between the ages of one month and two years. IPTi is vaccine-like in that its aim is preventive, it is given via injection, and it differs from conventional vaccines in two ways: it is not developed from a biological component or analogue of the infectious organism, and it does not aim primarily to prevent infection by the malaria parasite but, rather, to prevent development of clinical malarial symptoms.5 In public health terminology, IPTi aims at secondary prevention (prevention of the development of symptoms) rather than primary prevention (prevention of the establishment of a parasite). It thus may be referred to as a quasi vaccine.

The trial was conducted from 2005 to 2007 at two sites in parts of rural northern Tanzania experiencing high malaria transmission. Women were recruited during late pregnancy or at the first newborn visit for polio vaccinations and a combination of vaccines, referred to as DPT2, against three infectious diseases (diphtheria, pertussis, and tetanus), and they were invited to return one month later to enroll their child in the trial. Research assistants who facilitated the recruitment and consent procedures for the clinical trial were health workers with a medical background: nurses or clinical officers who also worked in clinical settings with the women they recruited. The health workers worked full time, were paid by the IPTi project, and spoke the local language. They underwent the same consent-process training and were assigned the same script to guide both recruitment and informed-consent processes. The women recruited who agreed to participate signed a consent form or, alternatively, took the consent form home to consider enrolling later. Infants enrolled in the trial received their first experimental IPTi at two months of age. The trial quasi vaccination was then repeated every three weeks until the infants were two years old. Seventy percent of the women recruited to participate agreed to join the trial, which resulted in approximately 1600 infant participants.

The trial was part of a wider pan-Africa cluster of IPTi trials. In our view, IPTi may have been wrongly interpreted by some major policy-makers primarily through lenses of therapeutic research, rather than of nontherapeutic research, because of high malaria prevalence in the study recruitment areas (as high as 60%). That interpretation may have been one factor underlying initial exuberance about IPTi. For several years, ongoing trials of IPTi in several African countries were coincident with large-scale efforts to widen access to IPTi. Although high prevalence of a disease may blur the distinction between therapeutic and nontherapeutic research, it neither completely erases the distinction nor renders ethically irrelevant the features more akin to nontherapeutic research. Significant numbers of those who did not contract malaria in the Tanzanian trial would be subject to trial risks without direct individual potential for unproven benefit. Moreover, subsequent study results demonstrably challenged researchers’ initial exuberance. Some studies indicated increased mortality despite reduced morbidity at the population level.6 In addition, concerns have been raised that IPTi may contribute to or be hampered by drug resistance.7 Based on some research findings, recommendations for when to use IPTi were revised recently, restricting use to high-prevalence areas with less than 50% permeation of an emerging malaria mutation conferring resistance to sulfa-based drugs.8 While these results were not available at the time our qualitative study was conducted, they retrospectively confirm our use of ethical guidelines for nontherapeutic research to guide analysis. Long-term trials and meta-analyses of IPTi continue to be evaluated.

The Kili IPTi informed-consent process. From our point of view, the consent process for the clinical trial had both significant strengths and weaknesses. On the positive side, the consent documents and scripts explicitly labeled the trial “research,” explained that participation was voluntary, noted the possibility of being randomly assigned a placebo (sugar pill), and listed potential side effects of the trial drugs. Most impressively, it is clear that the designed informed-consent process was indeed carried out. All (100%) of the mothers in our study recalled an informed-consent process in which they were asked if they wished to enroll their child, could say where the trial occurred and who conducted it, recalled their decision, and recalled signing an informed-consent document or declining to participate. On the negative side, the consent process missed opportunities to highlight three distinctions relevant to understanding the unique clinical trial: 1) the distinction between getting vaccines and participating in a vaccine (or quasi vaccine) clinical trial, 2) the distinction between using a drug for treatment and for prevention, and 3) the distinction between primary and secondary prevention.

The clinical trial recruiting was overlaid on a vaccination campaign for proven standard vaccines for children. Yet the consent process neither explicitly cautioned that this timing was a matter of recruiting convenience nor explicitly contrasted the proven vaccines covered by the vaccination campaign with the experimental quasi vaccine under trial. The drugs being combined and tested for potential chemoprophylactic effect were already widely used for malaria treatment. This was noted in the consent script only to discuss why risks were expected to be “rare”— misleadingly in our view, since drugs used in treatment are not inherently low risk. Significantly, the timing and dosages in the quasi-vaccine trial were not consistent with treatment protocols and were not intended as postsymptomatic treatment. Yet the consent process did not explicitly contrast the common use of these drugs in malaria treatment with the solely preventive strategy that was being assessed in the quasi-vaccine trial.

Though the consent process did identify malaria prevention as the main goal of the trial, it failed to differentiate goals of primary versus secondary prevention. Without such an explanation, posed in lay terms, discussions in the consent process of potential symptom reduction—which was a hope for the quasi vaccine’s effect at the population level—could be misinterpreted as discussions of conventional treatment for individuals. The initial description of purpose employed ambiguous language that could apply to both prevention and treatment (“protecting children from malaria,” “will be able to reduce malaria-related fever”). Indeed, a common name of the intervention, intermittent malaria prevention treatment, is paradoxically confusing in this regard. Ominously, at no place did the consent process exhort parents to seek evaluation and treatment immediately if their participating child showed symptoms consistent with malaria. Failure to stress consistently the distinction between treatment and an experimental effort at secondary prevention may have increased the actual risk of trial participation.

There were some other weaknesses in the consent script. Only one sentence noted that a sugar pill was one possible pill that would be “randomly allocated” to their child. This was not only a brief mention but also used a problematic analogy: a pill analogized as the placebo mechanism for a shot. The fact that research follow-up visits would check children for malaria and anemia was highlighted before the script noted the other research-related purpose of the visits: to ensure that the drugs were being taken according to protocol. Most significantly, in the discussions both of purpose and of risks and benefits, researchers’ optimistic expectations about trial results infiltrated the language of the consent script. The script describes researchers as wanting “to make sure the different medications in the study are safe,” as “believ[ing they] . . . will be effective,” and as “anticipating” success. In addition, the script claimed that one benefit of participating was that if a child was in the drug-receiving group, he or she might be protected from a life-threatening illness, which is a claim that could lead a consenting parent to assume the desired result.

Study Methods

We recruited participants for our study from four villages a year after the Kili IPTi trial was initiated.9 A simple screening questionnaire was used with a door-to-door approach to identify women who had been asked to consider enrolling their newborns in the clinical trial, those who had agreed to participate, and those who had declined to enroll their children. We used modified “snowball” recruiting techniques to identify homes with infants who would be the appropriate age one year out from trial enrollment. Asking informants with infants whether they knew others in the area with infants of similar age was considered an ethically benign question (as opposed to asking if they knew anyone else who had been approached for, participated in, or declined to participate in the clinical trial, which was not done). An informed-consent process for the interview study was conducted orally with appropriate potential interviewees, with consent obtained in writing. As part of that informed-consent process, it was made clear that our study was independent of the clinical trial and not aimed either at convincing decliners to reconsider their decision or at convincing participants to remain in the study. Those who agreed to participate could either be interviewed on the spot or make an appointment for a more convenient time.

Between March and May of 2008, in-depth interviews were conducted with 35 women. Of these, 20 had agreed to enroll their infants in the trial, whereas 15 had declined. Respondents ranged in age from 17 to 48 years, they came from different ethnic backgrounds, and all were from a rural setting. A large majority had zero to seven years of primary education. The mode level of education was seven years, equivalent to basic primary school in Tanzania. Most of the women worked on family farms, in small businesses, or as housewives. The principal investigator and two college-educated research assistants conducted the interviews in Kiswahili, the local language. All interviewers were trained in both sociology and qualitative research methods, and the research assistants were given additional training to familiarize them with the questions and objectives of the study.

We asked participants about their understanding of the goals of the trial and of its risks and benefits and also about their reasons for agreeing or declining to enroll their infants in it. Interviews followed a semi-structured interview guide. All respondents were asked the same questions but were encouraged to answer in their own words, as expansively as they liked. The interviews allowed follow-up of respondents’ concerns unanticipated by the interviewer. Whenever possible, questions raised were answered during the interviews, or else appropriate referral was made. In most cases, interviews were audio recorded while notes were taken about observation of gestures and informal discussions. In three cases, the respondents declined to be tape recorded, and notes of the interview were taken as nearly verbatim as possible. The transcripts were transcribed, translated into English, and typed into a Microsoft Word file. Using a color-schemed hand-coding system, responses were coded for preidentified themes consonant with the interview guide. A consensus process among the research team was used to identify criteria for code matching and for addressing cases challenging to code. Additional themes were identified in the data and new codes developed to track them. A map of data themes was manually generated and then refined. It was oriented around two families of questions asked in the interviews, about knowledge of trial purpose and methods and about reasons for agreeing or declining to participate. Linkages between those families of questions could then be explored to evaluate levels of understanding about research and likely cases of therapeutic misconception.

Study Results

We identified six study findings that are particularly striking given their ethical import: 1) Respondents varied in their understanding about the Kili IPTi trial. 2) There was better understanding among those who declined to participant than among those who enrolled their infants; in particular, decliners were more likely to understand the experimental nature of the trial. 3) There was a tendency for all respondents to confuse the trial as treatment rather than research (the conceptual error often labeled “therapeutic misconception”). 4) Respondents who were trial participants were more likely to confuse research with treatment than were trial decliners. 5) Perception of clinical benefit may have been accurate, rather than an expression of therapeutic misconception, because ancillary clinical benefits were provided to trial participants. 6) And, finally, neither trial participants nor decliners cited concern for the common good as reasons for their decisions, despite presumptions in bioethics in the developed world that such “altruistic” reasons are ethically the most appropriate motivations for participation in nontherapeutic research.

Understanding about the trial. We devised data-based protocols to group respondents into categories of understanding about different points. For example, one tripartite grouping included those participants who clearly noted that the clinical trial was experimental, those who clearly said something suggesting that they did not understand the experimental nature of the trial, and those whose understanding of that point was unclear in the conversational data. Results of these “macro groupings” are summarized in Table 1. The suggestive patterns of the macro groupings reciprocally informed our analysis of the qualitative data which generated them.

Overall, respondents understood that the trial targeted malaria, and they all remembered the source of their information—either the clinic or friends or relatives who had attended the clinic and referred them there. They remembered their decision to either enroll their newborn in the trial or not, including either signing or declining to sign an informed-consent document. Two additional findings undergird all the others: understanding about the trial was related to whether respondents participated in the trial or declined to do so, and decliners, in general, understood the nature of the trial better than the trial participants. Eleven out of 15 decliners, compared to 2 out of 20 participants, expressed some understanding that the trial involved the experimental testing of drugs. The fact that the drugs being used were “on trial,” a frequent expression of decliners, was also the most frequently cited reason for declining participation (by 73% of decliners). Said one decliner, “I decided not to take part because they are giving trial drugs to the children. I didn’t like my child to be tested because he can be affected by the drugs” (Decliner D). Decliner G explained, “[T]here was a certain part that was written in that paper that stated the drugs are on trial, so after reading it I said [to myself], ‘Something which is under trial means it is not complete.’” In contrast, only 2 out of 20 participants noted the experimental nature of the trial when they recalled the trial or their decision-making.

A majority of participants stated that researchers were giving their children malaria treatment (13 of 20). Confusion about treatment versus preventive goals, as well as ambiguities between primary and secondary prevention, enabled some participants to interpret experienced cases of clinical malaria as evidence of trial-drug success. For example, one participant noted that “the drugs have helped a lot as he was not getting malaria so often” (Participant N). Other participants correctly discussed prevention as the focal question of the trial but assumed that their child was given an effective malaria vaccine. Strikingly, 8 of the 20 participants said yes to the question, “Do you think the project helped to reduce [risk of] malaria [for your child] by 100%?” Some participants whose children did not get malaria assumed the trial drug helped, despite the actual equal odds of having been given a placebo. As Participant F explained, “I thank their drugs, as they have helped my child a lot. Since she was born, she has never suffered from malaria.” And one of the two participants who expressed formal understanding that the trial was an experiment intimated a belief that it was the last confirmatory step of a new successful vaccine from which her child would be among the first to benefit.

Nearly one-third of participants (8 of 20) expressed optimism that the drug their child was being given could both prevent and treat malaria. This confusion may relate to the fact that the drugs in the quasi-vaccine trial were a combination of drugs currently used for malaria treatment. For example, when asked about which drugs were used, Participant F realized they were in use for treatment: “I don’t remember the names of those drugs. I think it was sulphadoxine-pyremethamine [SP].” Asked by the interviewer why she thought the drug was SP, the respondent said, “When I looked at it, I thought it was SP even though it was made in a different form. I thought it was SP because by that time SP was used most.”

Although decliners generally understood the experimental nature of the trial better than participants, neither group demonstrated understanding that the trial used randomization to assign children to either a group that received a placebo or a group that received a study drug. The sole decliner who alluded to this was the exception that proved the rule.10 Generally, decliners disagreed with participants about the risks and benefits of receiving the trial drugs, but both participants and decliners assumed everyone in the trial was given the same drugs. When asked what treatment her child received, Participant G responded, “Everyone was getting the same treatment.” The use of the term “sugar pill” in the formal consent process may have contributed to a lack of understanding that through randomization only some infants got the study drug.

Assessing risks and benefits. Both participants and decliners assessed the trial’s risks and benefits in reference to their child, not to the community. While they disagreed on probability or magnitude, both participants and decliners assessed the same benefits: potentials for immunity, therapeutic benefit, or ancillary clinical services. Both participants and decliners considered ancillary clinical benefits a primary potential benefit, with participants more likely to expect immunity or therapeutic benefit and decliners more likely to consider risks as well as benefits. Trial participants expected malaria prevention or malaria treatment as benefits; 8 of the 20 participants expected total immunity as a result of participation. Decliners were more likely to understand that prevention was the goal of the trial; indeed, no decliners expected that trial participation could decrease the chance their child could get malaria “by 100%.” Strikingly, 13 of the 20 participants expected that the trial’s quasi vaccine would provide malaria treatment if their child got malaria, a hope unrelated to trial design that only one decliner intimated. Decliners were much more likely than participants to note that they had considered potential risks as well as benefits. While only one participant explicitly raised the topic of risk analysis without prompting, half of the decliners did. As Decliner E explained, “I did not know what would be the benefits and the risks of it [emphasis added]. I decided not to join. . . . He [her husband] also didn’t want it; he also said since the drugs are on trial, we should not join it.” Decliner E’s comments reflect how uncertain or negative perceptions of the risk-to-benefit ratio, generally intertwined with concerns that the drugs were experimental, were the major reasons that some women declined to participate: “For me, I said since they are trial drugs, they may affect my twins [badly] since they were not proven to work.” Notably, one decliner specifically cited rumors of side effects as having a major effect on assessment of risk: “I heard rumors from people who were saying there are children who were given these drugs and they affected them. Some got rashes; others, the skin looks like it has been burned. So that made me worry” (Decliner G).

With one exception, participants seemed not to have considered potential risk at all. In other words, their expectation of greater benefit was not the result of a calculation of a risk-to-benefit ratio. “Mainly there are benefits” was a marker of trust: “I was not thinking of risks, as they told me in case of anything the doctors are there and . . . . so I was not worried” (Participant G). Significantly, some participants noted either discussion of purported side effects among participants or even having noticed potential side effects in their children—without revising their assessment of minimal risk at the time of consent. They either questioned whether the noticed effect was related to the trial drug, or they weighed the benefit as overwhelmingly greater. As Participant J explained, “No [I was not worried about risk], but there were some people who said that the drugs have side effects on the children . . . . Later on, my child developed spots on his head, and some people started to point fingers at him, saying it is the malaria [trial] drugs infection. I told them, ‘Why only me, while there are many people who participated?’ The problem didn’t persist long, and my child was fine.”

Among participants who reported potential side effects at the time of injection (diarrhea, nausea, lethargy), either responsive clinical care or potential therapeutic misconception seemed to diminish any retrospective impetus to reconsider the risk-benefit ratio of the trial. Participant J said, “My child vomited a lot and became weak, but she was given other drugs, and everything was O.K.” When asked if she was afraid of potential risk, Participant R responded, “When I was given the medicine, my child had severe diarrhea.” The interviewer asked, “How has the project helped your child? Do you think it helped to reduce malaria by 100%?” and Participant R replied, “Yes!”

Therapeutic misconception or accurate perception of treatment? Notably, 80% of participants and 73% of decliners said the most widely expected trial benefit was ancillary clinical services. This may have been an accurate perception rather than therapeutic misconception, even though the clinic through which the research was being conducted offered free antenatal clinical care for area residents, trial participation notwithstanding. Participants ranked desire for ancillary clinical services as their biggest reason for participating more frequently than any other expected benefit. Like many participants, a majority of decliners presumed access to clinical resources was one of the biggest incentives to participate. Thus, their evaluation of benefit turned heavily on whether participants really did get clinical access that decliners did not (most thought yes) and, if so, whether it would be much harder to attain those benefits without participating (a question on which they evenly disagreed). Decliners who thought the advantage modest were the most likely to conclude that participation in the trial actually offered no, or very little, potential benefit.

The perception of some ancillary clinical benefit may have been an accurate perception rather than therapeutic misconception. Several participants expressed regret about coming to the end of the trial period because of concern that their child would lose clinical benefits. Although none of the participants expressed retrospective regret about their decision, one-half of the decliners did. Strikingly, all the regretful decliners regretted their decision because they believed that they lost clinical benefits by not participating. Clinical benefit may have been explicitly cited by at least one trial recruiter, though not as part of the formal informed-consent process. According to Participant B, “They [health researchers] said that the child will be given good treatment from the age of two months for up to two years. If the child will get any problem in case of illness, they will be responsible for that.”

Two potential clinical benefits appeared tangible to both participants and decliners: the periodic monitoring of the infants for malaria, which effectively provided well visits, and speedier access to clinical services with less waiting time. The monitoring was a necessary feature of study design. Privileged clinic access was not, but since both participants and decliners perceived such privileged access, it may have been a real “spin-off” effect of the monitoring. The fact that the researchers were the regular clinical providers in the antenatal clinic could also have contributed to a perception of enhanced access. Several participants discussed the benefit of the monitoring itself for their child. They hoped malaria or other health problems would be caught early because of it. As Participant N noted, “I was very happy to participate because there was no malaria test service in [town name] dispensary, and if you take the child there, he gets only medication without a test. I was happy that if I participated my child will get malaria tests frequently.” Participant D’s comments reflect the view that such testing could either help them [medical personnel] treat malaria early if her child contracted it or could prevent unnecessary malaria treatment by verifying that vague symptoms were not the onset of malaria (whereas otherwise the child might be given antimalarial drugs more readily without the time- and resource-intensive testing): “The reason for participating is that I want to know if my child has malaria or not because it is not easy to know if the infant is sick. So, if you joined the project, your child must do a checkup every clinic appointment, such as malaria, Hb, and if it is sick, you can know.” In a resource-poor setting, well visits might not be consistently available, or parents might feel pressure not to pursue them when the clinic is pressed to serve sick children. The pass card to proactive malaria testing and to what effectively amounted to regular well visits was both a significant incentive for research participation and a real benefit.

Although our study did not include independent efforts to assess actual access, it is noteworthy that both participants and decliners discussed faster clinical attention being given to participants. Participant F explained that “the main reason for joining was quick services provision. If you go there [to the project clinic], you get faster service, and they attend you very kindly,” and Participant C noted that “if you are in the project (clinical trial), you can go there and get the services soon and come back home earlier.” Decliner T expressed a similar view: “For me, I think the clinic where the project people were going was very good, because when you reach there you get quick and good services.”

While striking convergences between participants and decliners on ancillary clinical benefits suggest that some of those benefits were concrete, some participants clearly conflated the clinical trial with other available charitable medical resources in the area. For example, Participant J thought the trial’s aim was “to help children [through] health care provision. For example, if your child gets sick, they [medical personnel] were providing free treatment, and even if they didn’t have drugs in their stock, they give a prescription to buy drugs somewhere else, but they also give you money to buy.” One decliner also equated the trial with charitable care for people who could not afford drugs, implying that she did not need such assistance. In contrast, several decliners either specifically noted that they understood that everyone was eligible for free clinic services regardless of whether they participated in the trial, or they specifically distinguished the trial from available charitably offered resources. According to Decliner Y, “Even if you don’t understand a thing about the project, they [the researchers] will still treat you well and give you the good medicine.” Decliner W noted that “[t]his project has no benefit; we only benefitted from Hati Punguzo [a local charity program] during pregnancy because we were given [antimalarial] and insecticide-treated bed nets. And Decliner S reported, “If [the researchers] would give aid like Compassion [a philanthropic nongovernmental organization], participating could be better, but their service is no different because at the clinic the child will get the same free service for up to five years, so I knew it was not a must for me to join.”

Perception of having lost an opportunity for greater clinical attention was the reason cited for regret among all seven decliners who would have changed their decision if they could have gone back in time. Even several decliners who did not regret their decision concurred with the perception of easier access for participants but explained that they preferred to pursue clinical access that they understood as their right in other ways. No participants said they regretted their decision, and several had asked whether they could extend their participation (which would not make sense in terms of trial design) to continue receiving the ancillary clinical benefits. In several important ways, then, perception of clinical benefit may have been an accurate perception, not therapeutic misconception, although we cannot verify the actual extent of benefit as compared to perception.

A limitation of our study is that respondents were interviewed a year after they were recruited to participate in the Kili IPTi trial. Thus, it’s possible that after a year of reflection and discussions with others about the trial, they responded to the interview questions differently than they would have if interviewed soon after agreeing or declining to enroll in the trial.

Discussion

Participants’ self-understandings of the trial were closely linked to the decision made to participate or decline. Women were more likely to participate if they thought their child was likely to benefit from prevention or treatment or if they expected participants to receive clinic benefits such as quick access to care or subsidized medicine. Overall, decliners had a more accurate understanding of the clinical trial than did participants in that they understood its experimental nature. However, decliners shared with participants a lack of understanding of randomization and of the fact that infants randomized to the placebo arm did not get the experimental intervention. Decliners were more likely to consider potential risks as well as potential benefits. They were also more likely to distinguish the clinical trial from public or charitable clinical health services in the area, and they placed the burden of proof on why they should participate, rather than why not. These results are consistent with those of other studies that suggest that in resource-poor settings, perceived financial or clinical benefit is a key motivator to enroll in clinical trials and that many research participants do not understand what it means to be randomized to a placebo arm.11 Notably, not one participant or decliner mentioned the value to the community of increased scientific knowledge about malaria prevention as a factor in her decision-making.

Our findings from this qualitative study challenge presumptions that in resource-poor settings, what drives many decliners not to participate in trials is lack of understanding about the trial for which they are recruited to participate, fears of malicious disregard on the part of researchers (though some pointed to the location of the study in one area and not others as proof that it was an experiment), or “conspiracy theory” about why they were recruited to participate. Rather, those in our study who declined to participate in the trial went through steps of risk-benefit analysis that prospective research participants with understanding about a trial are expected to undertake and decided that the risk-benefit ratio did not warrant their child’s participation.

Notwithstanding our critiques of the clinical trial’s informed-consent process, it clearly accomplished some successes. In any cultural or socioeconomic setting, it would be laudable to attain these metrics: 100% of surrogate decision-makers reporting both remembering an informed-consent process and knowing they had a choice of whether to participate. However, many of the most common, and potentially most harmful, misunderstandings resulted from participants understanding some aspects of the study but not others. Partial understanding, even amidst laudable efforts to inform, can be as dangerous as ignorance. Some of the partial misunderstandings might have been addressed by an improved informed-consent process. Other authors point to endemic challenges of doing research in research-poor populations with high disease burdens. Partial misunderstandings among participants included 1) realizing that the trial’s quasi vaccine would be given in conjunction with other vaccines, without realizing that, unlike the others, this one was experimental; 2) realizing that the particular drugs being used in the intervention are already commonly in use for malaria treatment, without understanding the difference between secondary prevention and treatment or that treatment was not a goal of the trial; and 3) understanding what was given in the intervention drug cocktail without realizing the equal chance of being randomized to receive a placebo. Misunderstanding the purpose of a clinical trial and its elements might have increased the chance of harm from participating. For instance, trial participants might be less likely to see early intervention for malaria symptoms, either assuming that their children are protected or that the trial’s quasi vaccine will mitigate symptoms in case of malaria.

All of the misunderstandings may contribute to the therapeutic misconception. Yet Appelbaum et al.’s typology of therapeutic misconception is simultaneously helpful and unhelpful in identifying whether any of the Kili IPTi trial participants were under the therapeutic misconception.12 The authors articulate two types of errors: research participants’ failure to recognize that randomized interventions will not be individualized to personal needs and participants’ unreasonable appraisals of the nature or likelihood of medical benefit from trial participation. Lack of evidence of any understanding among participants in our study of randomization to a placebo arm suggests the first type of misconception. Yet every one of the participants understood that the intervention would not be personalized to her child—that all children given the intervention would be given the same thing. In this case, however, that knowledge may have contributed to misconception rather than being protective. Given the piggybacking of IPTi with an established vaccine program, that realization seems to have encouraged some participants to conceptualize the intervention as being more like a standard vaccine and less like research.

The second type of misconception occurred in two ways—either overestimating the chance of malaria prevention (by misunderstanding secondary prevention or by assuming the result researchers hoped for) or believing that the intervention would treat acute malaria. Participants with these misconceptions failed to realize that use of a placebo in the trial meant that their child might not receive a product with active ingredients. Both of these “type two” misconceptions could increase actual risk of trial participation, potentially resulting in parental delay in seeking medical attention for malaria symptoms. Not all degrees of therapeutic misconception among participants can be attributed to irrationality, being uninformed, or even partial understandings. Some “wishful thinking” might indeed be rational. Given high malaria burden in the area, it might be rational to think that the trial vaccine would prove effective and that the trial their child was participating in would prove to be malaria’s analogue to the Salk polio vaccine trials. Indeed, several participants responded in a way that reflects that rational hope. However, because there may be rational as well as irrational therapeutic misconception, researchers have special obligations to explain risks and random assignment when conducting research in areas of high disease prevalence.

Responses of both participants and decliners suggest that the perception of potential ancillary benefits may have been at least in part an accurate perception rather than therapeutic misconception. Under conditions of relative clinical scarcity, study features that the researchers did not consider to be clinical benefits may be just that. Routine malaria screenings were identified as a benefit by participants who assumed they could not get such care outside the trial. Similarly, in an environment in which even overtly ill patients may face long clinic lines, the perceived guarantee of quick access to the de facto well care provided by trial follow-up was seen as a benefit by decliners as well as participants. It is sobering that desire to receive such ancillary benefits, rather than any reassessment of the risks and benefits of research, drove the thinking of roughly half of the decliners, those who, in retrospect, wished to participate. The incentive to participate for ancillary clinical benefits was a distraction from primary assessment of risks and benefits and was unrelated to any diagnostic features of the participants. For these reasons, it is important to consider whether the ancillary benefits were an undue incentive to participate.13

The strong desire for such ancillary benefits and what appears to be a reasonable expectation of receiving them call into question the research-ethics discourse developed in affluent countries that suggests that altruistic motivation should drive willingness to participate in a clinical trial. This discourse also tends to applaud altruistic motivation as a good in itself as well as an indicator of understanding because participants who cite community benefit understand that the trial is not intended to benefit them personally. Our study highlights that such accounts of “ideal” altruistic motivation may lack adequate attention to other ethical factors that legitimately influence actual research decisions, such as overall access to medical care or the primacy of love for one’s child as motivation to obtain such access. Thus, nonaltruistic motivation to participate in a trial may make ethical as well as practical sense in some contexts, rather than being viewed as a de facto indicator of misunderstanding or the therapeutic misconception.

Even in resource-rich settings, conducting informed-consent processes that result in truly informed consent is challenging—and practice often falls short of the ideal.14 In resource-poor settings, related challenges are even greater.15 Our study results suggest the need for straightforward strategies to improve informed-consent processes in these settings. Any aspects of trial delivery that are planned for study feasibility and convenience should be acknowledged as such and distinguished from the scientific aspects of the trial’s methodology. Explaining to prospective trial participants that the Kili IPTi trial was being conducted at the same time that regular vaccines were given to infants simply to save time for clinic staff would have helped emphasize that the trial intervention was very different from a proven vaccine. The difference between prevention and treatment should also be clarified to prospective participants, especially in areas where people’s most common interaction with the medical system may be for acute clinical care. In addition, researchers should state explicitly that while people may recognize the trial drugs as malaria treatment drugs, for trials like the Kili IPTi one, the interventions are given to participants in a way that would work for treatment. Thus, if parents detect symptoms of malaria in their child, they should seek standard medical care promptly.

More detailed information about randomization and placebos should be provided in the informed-consent process. Three elements of randomization should be conceptually separated: that participants randomized to the placebo group do not get the trial drug, that who gets into the trial’s drug and placebo groups is based on random assignment, and that the researchers themselves do not know who is in which group if the trial is double-blinded. Potential participants should be told in lay terms why a comparison group is needed for the research; why it is important that some participants do not receive the trial drug (but instead receive a placebo); and why, given the experimental nature of the trial vaccine, it is not clear in advance whether the placebo or intervention group is better off. A helpful open-ended validation question could be included as part of the informed-consent process: “If you participate, what do you think will be given to your child?” If a respondent fails to answer that his or her child might be given nothing (a placebo), there should be further discussion about randomization to a placebo arm.

Finally, informed-consent processes should state clearly that the goal of the research is to gain knowledge that will help the community in the future, not to help the trial participants. It may be more important for potential research participants or their surrogate decision-makers to understand that community benefit is the researchers’ goal rather than for them to rank community benefit highly among their own considerations. Our study results suggest a mismatch between the researchers’ goals and the motivations of the women who considered enrolling their child. Awareness of that mismatch could help potential research participants make decisions that best negotiate the ethical trade-offs in their own decision-making.

Rose Mwangi, MPhil in Bioethics, is the deputy chair of the College Research Ethics Committee at Kilimanjaro Christian Medical Center; Paul Ndebele, PhD, is the director of the Medical Research Council of Zimbabwe; and Ann Mongoven, PhD, is an assistant professor in the Department of Religious Studies at Michigan State University.

Acknowledgment

This study was approved by the Kilimanjaro Christian Medical Center institutional review board (IRB) and the Michigan State University IRB. It was financially supported by the National Institutes of Health Fogarty African Bioethics Training Program, through Michigan State University and the University of Malawi College of Medicine.

References

1. Council for International Organizations of Medical Sciences and World Health Organization. International Ethical Guidelines for Biomedical Research Involving Human Subjects. 3rd ed. Geneva: CIOMS, 2002; Krosin MT et al. Problems of comprehension in informed consent in rural and peri-urban Mali, West Africa. Clinical Trials 2006;3:306-313; Hill Z et al. Informed consent in Ghana: What do participants really understand? Journal of Medical Ethics 2008;34:48-53.

2. Donovan JL, Brindle L, Mills N. Capturing users’ experiences of participating in cancer trials. European Journal of Cancer Care 2002;11:210-214; Townsley CA et al., Understanding the attitudes of the elderly toward enrollment into cancer clinical trials. BMC Cancer 2006;6:34; Ellis PM. Attitudes towards and participation in randomised clinical trials in oncology: A review of the literature. Annals of Oncology 2000;11:939-945.

3. Canvin K, Jacoby A. Duty, desire or indifference? A qualitative study of patient decisions about recruitment to an epilepsy treatment trial. Trials 2006;7(1):32; Ross S et al. Barriers to participation in randomised controlled trials: A systematic review. Journal of Clinical Epidemiology 1999;52:1143-1156; Mueller MR. Clinical, technical, and social contingencies and the decisions of adults with HIV/AIDS to enroll in clinical trials. Qualitative Health Research 2004;14:704-713.

4. Tindana PO, Kass N, Akweongo P. IRB: Ethics and Human Research 2006;28(3):1-6; see ref. 1, Krosin et al. 2006; Molyneux S, Bull S. Consent and community engagement in diverse research contexts: Participants in the consent and community engagement workshop, Kilifi, Kenya, March 2011. Journal of Empirical Research on Human Research Ethics 2013;8(4):1-18; Ndebele P et al. Trial participants’ understanding of randomization, double-blinding, and placebo use in low-literacy populations: Findings from a study conducted within a microbicide trial in Malawi. Journal of Empirical Research on Human Subjects Research 2014;9(1):83-85.

5. Our general characterization, that IPT aims at secondary prevention, is an oversimplification. By combining drugs with both known prophylactic and treatment effects, IPT hopes to deliver a one-two punch, so to speak, at the population level, both reducing parasite infection and reducing progression from infestation to symptomatic malaria. Categorizing the strategy as secondary prevention is adequate for our purposes since the trial researchers had no expectation that IPT could achieve anything like the true herd immunity of a conventional vaccine. The hope was more modest, that a combined reduction in parasite infestation and clinical disease could be achieved at the population level.

6. World Health Organization, WHO policy brief for the implementation of intermittent preventive treatment of malaria in pregnancy using sulfadoxine-pyrimethamine (IPTp-SP), April 2013, revised January 2014, https://www.who.int/entity/malaria/publications/atoz/policy_brief_iptp_sp_policy_recommendation/en/.

7. McGready R. Intermittent preventive treatment of malaria in infants. Lancet 2009;374(1400):1478-1480; Gesase S et al. High resistance of Plasmodium falciparum to ulphadoxine/pyrimethamine in northern Tanzania and the emergence of dhps resistance mutation at Codon 581. PLoS One 2009;4(2):e4569.

8. World Health Organization. WHO Policy Brief for the Implementation of Intermittent Preventive Treatment of Malaria in Pregnancy Using Sulfadoxine-Pyrimethamine (IPTp-SP). April 2013. Revised January 2014. https://www.who.int/entity/malaria/publications/atoz/policy_brief_iptp_sp_policy_recommendation/en/.

9. Those who enrolled but later dropped out were excluded from the study. There was not a large enough group of dropouts among study recruits to constitute a third group for comparison to participants and decliners. However, we posit that efforts to recruit and interview trial dropouts could be another promising strategy to explore relationships between understanding and decision-making, as well as how clinical trial participants’ understandings and attitudes may change over time.

10. Directive wording in the interview prompts may have influenced the almost 100% failure of interviewees to articulate that trial participants might receive a placebo, since the tricky question “What drug do you think was given/would have been given to your child?” was used. In retrospect, a less specific open-ended question would have been preferable (for example, “What do you think was given/would have been given?”). In our discussion, we recommend the open wording as a helpful question to assess understanding during a consent process.

11. Doshi MS et al. Evaluation of factors that motivate participants to consent for nontherapeutic trials in India. Journal of Medical Ethics 2013;39(6):391-396; Nabulsi M, Khalil Y, Makhoul J. Parental attitudes towards and perceptions of their children’s participation in research: A developing country perspective. Journal of Medical Ethics 2011;37(7):420-423; Mooley K, Pather Myer L. Informed consent and participant perceptions of influenza vaccine trials in South Africa. Journal of Medical Ethics 2005;31(12):727-732; see ref. 4, Ndebele et al. 2014.

12. Appelbaum PS, Lidz CW, Grisso T. Therapeutic misconception in clinical research: Frequency and risk factors. IRB: Ethics & Human Research 2004;26(2):1-8.

13. Emanuel EJ, Currie KE, Herman A. Undue inducement in clinical research in developing countries: Is it a worry? Lancet 2005;366:336-340. 

14. Grady C. Enduring and emerging challenges of informed consent. NEJM 2015;372:855-862.

15. Mandava A et al. The quality of informed consent: Mapping the landscape—a review of empirical data from developing and developed countries. Journal of Medical Ethics 2012;38:356-365; Lema VM, Mbondo M, Kamau VN. Informed consent for clinical trials: A review. East African Medical Journal 2009;86(3):33-42; Hyder AA et al. Ethical review of health researchers: Perspective from developing country researchers. Journal of Medical Ethics 2004;30:68-72.