How much is too much? Many IRBs lack guidelines for paying research subjects
Studies show wide variance in how much participants are paid
When an IRB reviews a proposal to pay research participants, members often have little more guidance to go on than their own gut feeling of what is appropriate and what is too much.
That’s the assessment of two researchers who have studied the payment practices of research institutions.1,2 They say the amounts paid to study participants vary tremendously — from site to site, even for the same multisite studies, and from study to study, even at the same institution for similar tasks.
They argue that more concrete guidelines would help IRBs make consistent decisions about payment and more fairly compensate volunteers who sacrifice time, comfort, and more to help advance science.
"My view is, and has been for a while, that the best thing that IRBs or institutions could do is to put together guidelines that help people find out what’s an acceptable way to calculate [compensation]," says Christine Grady, RN, PhD, head of NIH’s section on Human Subjects Research in the Department of Clinical Bioethics.
"Just to say to investigators, Whatever you do, don’t give me an amount [that could create] undue influence,’ gives them absolutely no guidance in terms of how to figure out how much to offer," says Grady, who emphasizes that she is speaking for herself only and not for the NIH.
Grady and Neal Dickert, a doctoral student at the Phoebe R. Berman Bioethics Institute in Baltimore, surveyed 32 U.S. research institutions to discover their policies regarding payment of research subjects.
They found that only 12 had any written policies or guidelines regarding payment — all but one reported using unwritten rules of thumb to determine compensation. Few could even provide an estimate of the proportion of their studies that paid subjects.
A second study, analyzing 467 protocols approved by 11 IRBs, found wide variation in the amounts of money paid for relatively similar procedures, both across institutions and within individual institutions. In one case, one offered $1,000 to participants at one site and $2,000 at another for the same study.
Dickert describes one study, on treatment of attention deficit hyperactivity disorder in children, where there were variations not only in how much was paid, but in who was paid.
"Payment went anywhere from zero to $275 to the teachers who were filling out the assessments for the kids," he says. "And some sites that didn’t pay teachers paid children and parents a fair amount of money — from nothing to over $300."
The amount offered for a subject to undergo an MRI varied among 12 studies from $25 to $120; in fact, the dollar amount for an MRI sometimes differed within one institution.
"It just shows that for whatever reason, investigators are coming up with different ways to pay people," Dickert says. "And all these different strategies were submitted to the IRB and the IRB probably looked at them and decided on a case-by-case basis whether each was reasonable rather than looking at whether they were all the same."
Some argue that IRBs need flexibility in assigning payment for studies, based on the population they’re working with, the specifics of the study involved and other factors that can’t be foreseen by a set of guidelines.
"I would be concerned if one tried to adopt a very formulaic approach to it," says Richard Mattes, MPH, PhD, RD, chair of the biomedical IRB at Purdue University in West Lafayette, IN. "Ethical questions are not black and white; they’re not scientific questions where you have a hypothesis, you do a test, and it’s yes or no. It’s all interpretation and it’s subtlety and it’s sensitivity. That would be very difficult to capture in a rigid system."
But developing more concrete guidelines would create advantages, including keeping researchers at an institution from competing for volunteers on the basis of payment, Dickert says. He says it also would promote simple fairness, the idea that two people doing the same amount of work — for instance , submitting to a fingerstick or visiting a clinic four times — would be compensated equally.
Dickert and Grady say an institution should start with the most basic question: Whom do we pay, and what are we paying for? Does the institution believe in paying only to cover a subject’s expenses, such as child care or taxi fare? Does the IRB believe in providing money as an incentive, or to compensate people for the time and effort it takes to participate in the study?
They argue for a sort of wage payment system — calculating the amount of time the subject must spend, and then paying something close to the local rate for unskilled labor.
Grady notes that IRBs often don’t see the situation from the point of view of a study volunteer, who must give up time from work and deal with other problems to participate.
"I’ve heard from research participants, They offered me $25 for each visit. But it cost me $15 to get here; by the time I added up all the buses and subways, and now I’ve got to take two or three hours off work. I’m losing money on this deal. To offer me $25 feels disrespectful to me.’"
Other issues that institutions should consider when developing guidelines for compensating subjects:
• Risk — Should riskier studies pay more? In Grady and Dickert’s survey of IRBs, nearly a third paid subjects for incurring risk. But paying for risk raises ethical issues as well as issues of science, Dickert says. Because people’s risk tolerance varies, paying extra for riskier studies may result in a study population that is riskier than the norm, he says.
"From a more systemic point of view, I think we spend a lot of time trying to minimize risk in studies, and if you can just raise payment for riskier studies, then there’s a sense that it removes an incentive on investigators to minimize risk," he says.
• Healthy subjects vs. patients — The vast majority of organizations surveyed (90%) reported paying healthy subjects and patient-subjects similarly when similar procedures were involved and there was no apparent difference in the likelihood of direct benefit.
Grady says paying patients could help address the problem of therapeutic misconception, the inability of patients to differentiate research from therapy.
"It’s at least possible that if you started to pay people for a certain study, it would help them to see the difference between what’s research and what’s not," she says, "because they don’t get paid for being taken care of."
Dickert says paying a patient might actually help avoid undue influence because it changes the dynamic between doctor and patient.
"If you’re offering money, you’re changing the nature of the relationship," he says. "It’s easier to say no — you’re not asking for a favor, it’s a different transaction entirely."
• Children and others who cannot consent — Dickert says special concerns accompany payments to children and others who can’t make the decision to participate in a study themselves.
"The biggest concern in my view with paying children is that you might create incentives for parents to make decisions based on what’s good for them rather than what’s good for the children," he says.
That problem can be mitigated by keeping payments reasonable and minimal.
Dickert notes that IRBs already strictly regulate children’s research, especially with an eye toward keeping risk levels lower.
• International research — Grady notes that there are tremendous differences from culture to culture regarding the role of money in a society. In some, paying research subjects would be unheard of, while it others, it would be expected. And of course, a sum of money that would be considered fairly minimal in the United States could be overwhelming, even an undue influence, in a poor nation.
Grady points out that these types of studies would have to be dealt with on a case-by-case basis. "My view has been that at least at this stage in our evolution of understanding this issue, the most appropriate locus for making those determinations is the local IRB," she says. "Maybe there should be at a central level for a multinational study — a decision about whether we offer money to anybody, and if we do, on what basis and what kind of amounts are we prepared to offer. Then, let the local IRBs determine whether that’s appropriate in their setting."
• Completion bonuses or escalating payments — Most organizations (84%) in the survey of IRBs reported requiring prorated payments to subjects, rather than requiring them to finish the entire study in order to be paid. More than half had rules on the use of completion bonuses, usually limiting the amount or advising against excessive completion bonuses as an undue influence to continue in a study.
"I think small completion bonuses to encourage people to complete the study are reasonable," Dickert says. "But I think it’s important that payment be prorated, if people are to withdraw."
He says completion bonuses or escalating incentives — increasing the size of the payments over the course of the study — can be useful, especially if there are important study endpoints, or if the study is a particularly long one.
"You just need to make sure that the amount is appropriate, where you’re not putting people in a position where they feel like they have to stay in the study and do things they really don’t want to do," he says.
There are other steps that IRBs can take to help minimize the possibility that compensation will become an undue influence on participants.
Rebecca Pentz, PhD, a clinical ethicist at Emory University’s Winship Cancer Institute who has served on a number of IRBs, says an IRB can require stricter screening to ensure that a person doesn’t lie about eligibility to participate in a trial.
"If this study has pretty stringent eligibility requirements and you’re relying on self-report, but you’re also paying the patient, then you probably want to require that there actually be a test to make sure volunteers are eligible," she says. "There are cases in the literature of people lying about eligibility for relatively small amounts of money and having bad outcomes."
Pentz says when she is serving on an IRB and sees a study that raises concerns, she suggests an ethics companion study to collect data on why people volunteered for it.
"How did they view the money? Would they have done it if there was another way to get the exact same money that didn’t involve any risk at all?" she asks. "Then, you can use that preliminary data, on that very study, if you’re finding out that you’re getting some really odd results."
The results also could be used to make determinations about future similar studies, she says.
Dickert says it’s important to focus first on the study itself, and then decide whether the payment is appropriate.
"If a study is ethically problematic to start with, no amount of payment or nonpayment is ever going to make it acceptable," he says. "Similarly for studies that pose no risk — a fingerstick blood glucose test, or listening to someone’s heart or asking what their favorite color is — I really don’t think most of us have a worry about how much you’re paying.
"I think that in the absence of strict policy guidance, it’s important to really contextualize payment and to worry about it in the cases where it’s important to worry about it, which in my view are studies that are particularly risky or studies where you have reason to believe a good number of people might have good strong reasons not to participate," Dickert says.
- Grady C, Dickert N, Jawetz T, et al. An analysis of U.S. practices of paying research participants. Contemp Clin Trials 2005; 26(3):365-375.
- Dickert N, Emanuel E, Grady C. Paying research subjects: Analysis of current policies. Ann Int Med 2002; 136(5): 368-373.