Improving Informed Consent and Enhancing Recruitment for Research by Understanding Economic Behavior

The use of financial compensation as a recruitment tool in medical research continues to be debated on ethical grounds. Critics are against economic or market models, in which market forces determine payment practices, primarily because of the perceived undue inducement potential^{1 - 11} and on other grounds,^{1 ,12} including that participants from lower socioeconomic levels bear a disproportionate share of the research burden.³ Many of these arguments apply not only to the purest market model but to any model in which compensation to participants exceeds some threshold. We argue that economic forces operate regardless of how investigators choose to compensate participants. Thus, it is imperative that investigators acknowledge these forces and design compensation schemes that explicitly take economic forces into account.

Economic principles predict that for otherwise undecided individuals, incentives serve as inducements, regardless of the compensation model used. If they did not, why would researchers pay research participants? In place of a model in which incentives do not substantially influence behavior, we describe a general positive model (describing how payments actually are determined) of costs and benefits factoring in the decision. This model is essentially the same as the market model rejected by Dickert and Grady,¹ and the freedom of contract model proposed by Wilkinson and Moore,^{6 - 7} with an emphasis on nonpecuniary aspects of costs and benefits. Quite simply, we argue that individuals will participate in research if they think the benefit (including but not limited to monetary compensation) of doing so is greater than the cost. These costs and benefits vary across research projects, as well as across potential participants for any given project, as individuals perceive and value the costs and benefits of participation differently. Because monetary compensation is just one of many determinants of participation, regulating it alone will not prevent undue inducement.

Several important implications flow from this model. First, research participants who fail to grasp the facts or implications of participation could be induced by low payment amounts: truly informed consent is thus integral to preventing undue inducement.^{13 - 14} Second, external validity, an important ethical dimension¹⁵ that depends on adequate and representative samples, will be limited unless compensation—monetary or otherwise—is recognized for what it is: a factor in real-world decision making. Third, financial gain is one of several factors in the research participation decision. By focusing on other factors, investigators may be able to enhance recruitment through other channels.

Economic principles predict that researchers will be willing to compensate research participants up to the amount they expect the individuals’ participation to benefit the research. An individual will be willing to participate only if the expected benefits of doing so equal or exceed his or her expected costs. Only if both sides (the researcher and the participant) find the arrangement worthwhile will participation take place.

Economic theory further predicts that allowing unregulated exchanges to determine the level of compensation will yield the optimal quantity and composition of participants, under several quite restrictive assumptions. We do not argue that these assumptions hold and that the unregulated market will determine the best outcome. For example, one flagrant violation of the necessary conditions for an efficient market is the obvious lack of symmetric information: researchers know more about what experiments entail than do those who participate in them despite the requirement of informed consent.^{16 - 17} We review the major determinants of research participant supply and how these mold our expectations of who will choose to participate in research. Because most ethical concerns are centered on the supply side of this market, we do not discuss the determinants of the demand (by investigators) for research participants. Regulatory bodies can (and sometimes do) influence this demand by requiring representative samples and by examining pressures on investigators that may lead to unethical recruitment practices.⁹

Rational potential research participants make decisions based on expected costs and benefits of participation. Rational is defined as those participants with intact decision-making capacity; vulnerable populations, such as children or individuals with cognitive impairments, require special protections. We emphasize expected costs and benefits because potential research participants do not know the outcomes of participation with certainty but rather weight possible outcomes according to perceived probabilities and their own risk preferences.¹⁸ The model allows individual preferences to determine the relative importance of various costs and benefits.

Costs of participation include the opportunity cost of time spent participating, transaction costs, and potential discomfort or adverse responses. The opportunity cost is the value of the next best option faced by the potential research participant, including benefits derived from working or leisure activities. Opportunity costs are higher for higher-wage research participants. Among nonworking individuals, the opportunity cost is higher for those valuing their leisure time more. When all else is equal, lower reimbursement levels will result in recruitment of research participants with lower opportunity costs. These participants, however, may not accurately represent the target population. Transaction costs include transportation, child care, and other relevant study-specific costs. These costs are relatively free of controversy and data support the relevance to participants.^{19 - 20} In addition, participation may entail other barriers to participation that are not acknowledged generally such as greater expenditures of physical or psychological resources.^{20 - 21} One review²¹ found that eligible older adults who refused geriatric depression studies had poorer health and were more socially isolated, indicating the relevance of these other costs to participation decisions, and suggesting implications for generalizability.

The cost of greatest interest in the ethical debate has been the possibility of adverse responses, reasoning that payment would induce some individuals to agree to risky research.²² Research participants vary from one another and, importantly, from investigators in their appreciation of the potential for and likelihood of adverse events. These individual-specific participant-calculated costs (rather than the costs calculated by an investigator) are used by the participant to determine participation.²⁰

Benefits of research to participants include any perceived or expected health benefits (including the possibility of receiving active medication in placebo-controlled trials), altruistic or intellectual gratification, and cash or in-kind compensation. Access to medical care may be construed as a benefit, which in some cases may raise ethical concerns.^{5 ,23 - 25} Expected health benefits vary with the specifics of each trial, the probability of different treatment assignments (and the individual’s ability to assess these probabilities), as well as the individual’s own medical condition.²⁰ Many participants report psychological benefits, such as hope and altruistic gratification.^{19 - 20 ,26 - 32} Social interactions stemming from research visits also provide psychological benefits.³³ Moreover, any cash or cash-equivalent compensation would be valued differently by each participant.

Some have argued^{3 ,8 ,34} against higher payment levels, worrying that they will disproportionately induce the poor. We argue that relatively low compensation levels would often do the same because wealthier research participants have higher opportunity costs and value their last dollar earned less than poorer participants. In such cases, higher payment levels may result in more representative research participant pools by inducing wealthier research participants to participate. In some instances (eg, Alzheimer disease clinical trials), disproportionate enrollment of wealthier individuals tends to occur despite typically minimal financial compensation, indicating that other perceived benefits of participation are important factors and that perception of such benefits varies within the population.^{35 - 36} Regardless of which group disproportionately enrolls, it is undesirable from the standpoint of generalizability.

Decisions are also influenced by factors such as trust and perceived fairness. Beneficence, justice, and fairness are established requisites for ethical research.^{1 ,15 ,37} Differences in willingness to participate among individuals of different ethnicities suggest that trust in research and awareness of past ethical violations affect participation decisions and that establishing and maintaining trust is central to ethical and effective recruitment.^{38 - 41} Evidence from behavioral economics suggests that economically rational choices are often rejected when perceived as unfair,⁴² arguing against setting artificially low compensation levels or providing different levels of compensation for different research participants.

Available evidence on how individuals decide whether to participate supports the model we describe.^{18 ,20 ,43} For example, in one study, hypertensive patients presented with hypertension-related protocols reported less willingness to participate as study risks increased, probability of placebo assignment increased, and payment level decreased.¹⁸ In another study, healthy volunteers considering hypothetical protocols of varying risk and payment levels reported a greater willingness to participate as payment levels rose. The magnitude of this increase, however, was independent of the risk level of the study, and the reported willingness to participate was responsive to risk within each payment level.⁴⁴ Although data on actual payment policies and practices are sparse, what is known suggests that compensating research participants is common and is affected by the forces we and others have described.^{45 - 46}

If a potential research participant perceives the medical risks of participation to be lower than the actual risks, but understands the financial gain from participation, he or she may decide to participate when he or she would not have participated if paid less. Undue inducement arguments against market pricing emphasize the role of compensation rather than incorrectly assessed risk in this scenario. We argue that this decision is not ethically desirable because the costs are not known to the participant and are greater than the benefits, and not because the benefits are “too high.”

Arguments to dispel fears regarding undue inducement^{6 - 8 ,22} depend heavily on the role of independent review as Emanuel recently pointed out.²² However, we disagree with the conclusion that research participants should be paid an hourly wage (equivalent to what an unskilled worker would earn). Emanuel does acknowledge the concern that lower levels of compensation could lead to a disproportionate recruitment of the poor. This is precisely our concern because disproportionate recruitment of any group diminishes generalizability, threatening external validity.^{15 ,47}

To promote recruitment conforming to ethical principles of voluntarism and validity, several strategies should supplement current practices and oversight. First, information asymmetries must be reduced by promoting meaningful informed consent.² Despite the ideal of informed consent,^{48 - 49} investigators often produce long, legalistic consent documents that exceed most individuals’ reading abilities.^{50 - 53} Decades of research show that poor understanding of consent documents is widespread,^{2 ,16 ,51 ,54 - 58} and that many individuals are not able to discern important differences between research and routine care.^{27 ,59} This failure, termed the therapeutic misconception,⁶⁰ may result in overestimation of benefits, underestimation of risks, or both.^{13 ,61 - 62} Many participants, moreover, believe that consent forms are primarily designed to protect investigators rather than research participants.^{63 - 64} In addition, most participants have already decided to participate prior to the formal consent process.^{20 ,65}

These findings demand a major overhaul of federal priorities and regulations regarding consent, with less rather than more regulation needed. Overall priorities should include implementing a process model of informed consent^{64 ,66} ; attending to individual preferences and motivations regarding participation, with opportunities to address misconceptions^{20 ,43} ; reframing consent as an educational process and studying innovative teaching methods^{2 ,16 ,51 ,54 ,57 - 58} ; and combating therapeutic misconceptions.¹³ Payments may mitigate therapeutic misconceptions by highlighting distinctions between treatment and research.¹

Second, institutional review boards can address fears that payments could drive some to falsify information to be eligible,⁶⁷ potentially placing the research participant at risk. While such instances have been rare,²² institutional review boards must consider requiring more rigorous screening (especially for higher-risk studies)¹⁵ regardless of the compensation offered. Other issues that institutional review boards and investigators may need to consider include individuals’ personal financial interest in the outcome of a trial.⁶⁸ Compensation of investigators is a powerful yet seldom-discussed factor in recruitment practices.⁹ The overall message is that institutional review boards must muster in DeRenzo’s words “moral courage” to protect research participants.⁵

Third, institutional review boards (or appropriate federal bodies) could collect data on levels of financial and other compensation for different protocols and participant groups, and on the demographic makeup of research participants. These data could be used to identify unusually high compensation levels or unrepresentative samples to identify studies warranting further investigation with regard to risks, recruitment, and consent while acknowledging that such identification would be neither necessary nor sufficient indicators of ethical violations. In addition, creative recruitment strategies and incentives should be encouraged and studied⁶⁹ ; alternative protocol designs may also be needed to enhance recruitment.⁴³

Fourth, research reports should include more information on recruitment and compensation of research participants.^{47 ,70} Studies examining real responses to pecuniary and nonpecuniary aspects of research participation are needed.^{1 - 2 ,43} Differences in outcomes may be affected by compensation practices, but this question requires empirical attention. Research is also needed regarding ethics review practices⁶⁴ as well as the influences on voluntarism.^{25 ,71}

Such information could promote recruitment based on relevant dimensions of diversity, identify ethical lapses, and enhance generalizability—goals consistent with value, validity, favorable risk-to-benefit ratio, and fair participant selection.¹⁵ Whatever compensation model appeals to investigators, market forces exist and have implications for recruitment. Acknowledging and understanding these forces is a necessary precursor to regulation and is needed to ensure ethical research practices.