Assessing a clinical trial’s risks and benefits requires closer IRB attention
Research proposals often need work in this area
Here’s a scenario that’s all too common: A federally funded research proposal that would involve a medically vulnerable patient population in a multicenter trial includes a discussion of risks and benefits that is summarized as "Informed consent will be required." Instead of a thorough and ethical discussion of a trial’s risks of harm and potential for benefit, the research proposal dismisses the issue in a stock statement about informed consent.
"The thing that gives me the most concern is that IRBs are often not presented with enough information to make a realistic risk or benefit assessment," says Dale Hammerschmidt, MD, FACP, associate professor of medicine at the University of Minnesota Medical School in Minneapolis. He is a member of one of the university’s IRBs and has chaired IRBs for a decade.
Hammerschmidt recently reviewed a research proposal that included the above statement about informed consent in lieu of offering a risk-benefit assessment. The patient population of this project would include intensive care unit patients, some of whom would not be able to give their own informed consent and, therefore, would require surrogates to decide whether they would participate in the trial. "The protocol had no discussion of vulnerability or the implications of using a surrogate or why there was a benefit in the protocol to using people who couldn’t give consent," he says. "This information was only available if members of an IRB had the expertise to bring it to bear."
And yet, IRBs reviewing this protocol were expected to form some sort of rational judgment on whether the risks and benefits were in balance and would justify using subjects who were unable to make their own informed consent, Hammerschmidt adds. "The generic underlying problem is when people are crafting protocols they are spending a lot of time and attention on designing the protocol for the highest resolution of scientific question, but the ethical implications of the study seem to be delegated to someone else and are an afterthought in the protocol, as well," he says.
"There are all types of risks in research such as physical, economic, legal, social, confidentiality and psychological," says John Isidor, JD, CEO of Schulman Associates IRB of Cincinnati and IRB Advisor editorial board member. "Examples of physical risk are the risks of the study product or comparison product. Also, there may be risks from the study design such as a washout risk or placebo. There are also risks from the study procedures such as surgery, scopes, biopsies, X-rays, etc. Economic risks may include the costs of the research, particularly in device research where the subject pays the cost of the product and lost compensation for the time it takes to participate in research," he explains.
"Social and psychological risks could emanate from the social or psychological harm of the research. Illustrations could include research into activities such as sexual behavior, alcohol or drug abuse or psychological types of research. Legal risks include research involving illegal activities such as drug abuse, violence, etc., which could expose the participant to arrest if their identity is disclosed," Isidor concludes. "I think IRBs should establish categories of risk so they can review the risks involved in the research. I think a review outline or checklist identifying categories of risks is helpful for reviewers. Our IRB utilizes this approach."
All types of risks exist
Researchers and IRBs often do not devote enough attention to assessing risks and benefits in clinical trials, says Paul B. Gold, PhD, an assistant professor of psychiatry at the Medical University of South Carolina in Charleston. Gold also is a member and an associate vice chair for an IRB at the university. "There are many different dimensions of risk and benefit, and these are often dimensions that don’t get much consideration," he says. "There’s an overall global risk and benefit that people tend to view for their study. Unfortunately, that oversimplifies it."
Too often researchers and IRBs do not consider the various kinds of risks of harm and the probability of these occurring, he says. Likewise, many researchers fail to address the concept of benefit and do not provide the IRB with an assessment of the probability and magnitude of a benefit that’s being considered, Gold adds.
Hammerschmidt, Gold, and other experts say that the entire issue of assessing risks and benefits is poorly addressed and misunderstood by many researchers, study subjects, and even IRBs. This misunderstanding sometimes leads to the unrealistic expectations on the part of trial participants that they will personally benefit from their participation in a study, even in cases where this potential direct benefit is very unlikely.
"I believe this is the case, particularly in the context of phase 1 clinical trials," says Howard Mann, MD, program associate in the Division of Medical Ethics at the University of Utah School of Medicine in Salt Lake City. Mann, a radiologist with the university hospital and clinics in Salt Lake City, developed the ASSERT statement that represents an explanation of the requirements for the ethical conduct of human subjects research in the form of randomized controlled clinical trials.
What about benefits?
"IRBs, in my view, do not pay sufficient attention to this when comparing the trial’s protocol with what is stated in the consent document," Mann says. "In contradistinction, investigators do not mention the benefits associated with trial participation, per se."
While study protocols often do discuss risks of harm, there is very little discussion about potential for benefit, notes Nancy M.P. King, JD, professor of social medicine at the University of North Carolina-Chapel Hill. "IRBs are used to looking at consent forms with a great deal of detail about the risk, including likelihood and magnitude," she says. "But that doesn’t happen very much with benefit." Even the most conscientious investigators may simply state that a participant may or may not benefit, a vague statement that leaves its meaning up to the subject’s imagination, King says.
The subject is left to guess the answers to these questions: "Is it a cure? Is it a temporary change in some kind of measure of illness? How long will it last? How likely is it? What are your chances?" she asserts. "There is an assumption that taken to an extreme is a therapeutic misconception and it is that there is a potential for benefit because the person has a disease and the potential, down-the-road treatment is being tested," King says.
Despite the fact that investigators are not accustomed to thinking in terms of describing potential benefit, this can be and should be done, she states. "It’s possible to be somewhat specific because we’re somewhat specific about the risks of harm all the time, even from animal studies," King says. "So one of the things I think is it’s important for investigators to become accustomed to providing potential subjects with more information about how If everything goes the way we envision it, here’s what we expect, and if not, here’s what is likely to happen.’"
One of the more controversial issues has to do with the IRB’s role in conducting or ensuring an adequate scientific review of a protocol. "IRBs for a variety of reasons really do have to do a scientific review of a protocol," King says. "They have to be able to make their own judgment about whether the risk for harm and potential for benefit calculus has been done with enough information."
Potential social and scientific value should be considered in relation to what is already known about the medical issue, Mann says. "This should be assessed in relation to a systematic review of the relevant medical literature."
It’s a matter of resources
At major medical academic centers IRBs see many protocols that have thorough scientific reviews because the studies have been through a National Institutes of Health review or were reviewed as part of a pharmaceutical industry sponsorship. But the problem may lie with small IRBs at community hospitals, start-up companies, and homegrown studies, where the IRB may be the only body that is capable of doing a scientific review of the protocol, King explains. "I think ultimately what IRBs may not do enough of, especially small IRBs without resources, is to say, We cannot evaluate the science of this and are worried about it,’ and then leave it at that and make the investigator figure out what to do to make a persuasive case," she says.
Even when a protocol contains a scientific review, it should be assessed for bias. Research applications typically contain an investigator’s or sponsor’s narrative review, but the relevance of these may be biased by selective literature citations, Mann says. One example of an instance where reference to a systematic review of the literature may result in the decision that the proposed study does not meet an IRB’s requirements pertains to a placebo-controlled trial of an intervention, he says. "A literature review may show that many previous placebo-controlled trials have been completed, and that the research question has already been answered: The intervention is efficacious," Mann says. "Randomization of participants to a placebo arm is unethical when effective treatment exists."
Although the idea hasn’t caught on universally, there is at least one strategy IRBs might use to make sure risks and benefits have been adequately addressed. This would be to require investigators to include a risk-benefit ratio tool with their protocol proposals. A number of these tools exist, and one example is a risk-benefit matrix that Gold developed from the Belmont Report for his own research. The tool is a simple chart that includes columns on ethical principles and practical applications. (See risk-benefit chart.)
"What I do in my work is I lay out the possible harms, and I always try to think in terms of parallel construction of use of terms because one of the things that’s devastating to people in the world of medicine and science and life is that we use words like risk’ and benefit’ and interpret the meaning of those words differently," Gold explains. The term risk equals a probability, and the term benefit equals an outcome, Gold says. "Thus, a risk/benefit assessment is a bit of a misnomer because a probability is compared with an outcome."
A better strategy is to frame risks as probability and magnitude of potential harms and, in parallel, to frame benefits as probability and magnitude of gain, Gold adds. "Then it may be possible to array harms and benefits to make a determination about whether benefits exceed harms and thus argue that a study ought to go forward," Gold says. "This approach assumes that other criteria necessary to justify a study are clearly met."
This type of thoughtful assessment of risks and benefits could be a gold standard that IRBs expect from investigators, but this often isn’t the case. "IRBs are put into the awkward position of being perceived as a real barrier to research if they say, You haven’t given us enough information.’ They’re seen as obstructionist," Hammerschmidt says. "But if researchers have included the ethical questions in the planning of their research, then they probably would have known what to say in those parts of the protocol, and it would have been very clear," he adds. "The fact that it’s not there makes me think it hasn’t been thought about in a clear way."