Secondary subjects still need IRB consideration
Are you asking PIs the right questions?
It has been five years since a complaint from the father of a study subject was lodged about a longitudinal study at Virginia Commonwealth University (VCU) in Richmond — a complaint that precipitated a shutdown of all research involving human subjects. But five years has not been long enough to answer all the questions about secondary subjects the case raised.
It was the summer of 1998 when the father of a girl in a long-running twin study opened a thick envelope from the university and found a questionnaire that included queries about the girl’s parents and twin. The father complained, saying the rest of the family was essentially participating in research without having consented to it, and the complaint was upheld. After the Office of Human Research Protections investigated the complaint and found the university’s response lacking, the Food and Drug Administration suspended all human research. More than a thousand studies were impacted, and it wasn’t until the spring of 2001 that the last of the restrictions on human research at VCU were removed.
"This was a real wake-up call to for IRBs," says Dale Hammerschmidt, MD, associate professor in the department of medicine and director of education in human subjects’ protection at the University of Minnesota in Minneapolis. "A lot of people disagree that the man was a research subject because the information the questionnaire asked for was hearsay."
Regulators felt differently, though, and Hammerschmidt still wonders how much of a damper that could put on "any study based on medical records where family medical history enables you to identify close relatives. I think a lot of people are still running scared on this topic."
The problem grows perhaps more convoluted in an era where a lot of research involves genetic information, he adds. "You are studying Timmy because he has some bizarre disease, but you have DNA from all the first-order relatives. Now you know that Ralph and Fred and Joan are all carriers. With this decision, they all become study subjects, and you have to get informed consent from them."
The pragmatic response is that IRBs must ask every single time whether the issue of secondary subjects — and potential secondary subjects — is being addressed, and being addressed adequately, Hammerschmidt says.
Get consent if required; otherwise, work on a waiver, and do so explicitly. "If the research contains no greater than minimal risk, state it," he says. "If consent for all potential secondary subjects is impracticable, state it. If the rights and welfare of the potential secondary subject won’t be compromised if you don’t get consent, state it."
In the past, the University of Minnesota never used to do a formal finding of a waiver. It was implied. "Since the VCU case, we always do," says Hammerschmidt.
"If you look hard enough, you’re going to find a secondary subject," says Cynthia Dunn, MD, formerly with Western Institutional Review Board (WIRB) and now owner and president of a consultancy, Clinical Monitoring and Development Advisor (CMD), in Pittsford, NY. "If you take a medical history, you are disclosing information that is identifiable, private health information. The issue then becomes context. If it is for patient care, people look the other way, but in research, it becomes more of an issue."
Dunn says it’s a "judgment call for an IRB because there are no hard and fast rules. The same IRB will come up with different determinations on the same issue on two different days. It depends on how they figure out what the risk is to the secondary subject. Can it cause psychological distress? Affect insurability? Do damage to social standing?"
If the information is potentially damaging, be cautious, she says. Red flags include genetic studies, and those looking at genetic markers. "That’s an area where you would be well advised to look into the potential consequences of the information being disclosed if it gets into the wrong hands," she notes.
One problem, says Dunn, is that there is so much emphasis on benefits of research, and not enough on the risk. "That part of the equation has to come more to the fore. We have to think about this more, watch for the red flags; and if you have an issue with risk, get informed consent."
Maybe this is more about common courtesy than anything else, says Marianne M. Elliott, MS, CIP, the director of the office of research administration at the Naval Medical Research Center in Silver Spring, MD.
"Before you do something to or about someone, maybe you want to ask, May I please,’" says Elliott, who used to work with the Office of Human Research Protections. "Even if it’s something awful, if you ask them, they may say yes. I think the father in the VCU case was more incensed that he wasn’t personally asked his permission than anything else."
Still, Elliott says she understands that you can’t ask everyone everything all the time. "The IRB’s responsibility is to ask whether a reasonable person would want to be asked if information could be collected about them," she says. "Most people want to help and won’t say no most of the time, provided you keep the information confidential. Then you have to look at the potential harm if there was inadvertent disclosure. Would it be stigmatizing or embarrassing? Could it cost someone his or her job?"
IRBs should then look at what measures or strategies for minimizing harm are included in the study proposal, says Elliott. And if an investigator doesn’t address the issue — either because he or she forgot or because it isn’t an issue — send the proposal back. "They have to be explicit in the potential for harm and how such harm, should it occur, would be remedied."
IRBs also should make sure researchers aren’t going fishing for information in their questionnaires. "Look at what the investigator is collecting and for what purpose," she suggests. "If you are doing a study about ice cream eating habits, why do you need a question about bicycle riding in there?"
Another kind of secondary subject
Along with potential secondary subjects due to the use of identifiable information, there is another group of people that can become secondary subjects: those whose health can be impacted by studies not being done on them directly.
Hammerschmidt has a couple of examples. Two decades ago, a study was done on whether those getting bone marrow transplants would have improved outcomes if they got blood transfusions from people who had never had cytomegalovirus (CMV). "In order to do the study, one had to randomize whether study subjects would get blood according to ordinary screening or get blood exclusively from those with no CMV antibodies," he explains. "The problem is that it would use almost all of the CMV-negative blood in a community. It was altering the blood supply in that area. Every transfusion recipient in the area would have had a research-related risk, even if they weren’t participants."
Either the study would have to meet the strict criteria for a waiver of consent, or the investigators would have to get consent from every transfusion recipient in the community. In the end, they were able to meet waiver criteria by finding communities that had very low incidence of CMV in the population.
Another example is a study done on cocaine metabolism in addicts vs. nonaddicts. "The research was done at a medical research center, and the issues of having cocaine addicts in a medical care setting led to concerns about safety for other people in the area," he says. These were addressed by having extra guards on site during the study, thus minimizing the risk to potential secondary subjects.
A more current example involves new vaccines being tested with live viruses, like for HIV and smallpox. "There is a chance with these that contacts of the patients can acquire the virus," says Hammerschmidt.
Dunn says that with potential risk to physical harm, there has been a problem with regulation not progressing as fast as science. Consider the ongoing discussions about using diluted live virus vaccines of smallpox on children. "In this, there is not only a potential risk to the subject, but to family members, other close contacts, health care providers, and the community at large," says Dunn. "That whole area hasn’t received a whole lot of thought and consideration yet."
Despite the real concern for secondary subjects, Elliott warns against getting stuck on the issue. "You have to consider it. Don’t stop when you realize there is no potential physical harm. Think of psychological or dignitary harm. But don’t brood on it. A proposal from an investigator should include statements on the risks, how big they are, and whether the potential risk or harm is so small that the study can be done without consent from secondary subjects," she says. "Understand that there are extremes in any research. A death to a subject is bad. A headache for a subject isn’t bad. A headache for every single subject is something you might want to tell people about and get their consent for."