Does your IRB have a plan for handling incidental findings?

If not, here's what you need to know

How IRBs should handle incidental findings is becoming such a notable issue among IRB professionals that there was a recent conference devoted to the topic.1

"I think it's getting increasing scrutiny," says Elizabeth Hohmann, MD, an associate professor of medicine at Harvard Medical School, and a chair and physician director at Partners Human Research Committee of Massachusetts General Hospital and Brigham & Women's Hospital in Boston, MA.

"The most common incidental findings in clinical medicine are the newer imaging studies," Hohmann says. "You tend to find a lot of things, and it's unclear how important they are."

For example, both MRIs and CT scanning of the whole body easily can turn up incidental findings, Hohmann says.

The big question is: What should be done with the incidental findings?

For instance, suppose an investigator finds an unknown bright object in the brain or a small polyp or cyst in the kidney or liver?

"These are very common findings on scannings," Hohmann says. "In both incidents, you have something new, so what do you do with it?"

IRBs have to grapple with these questions because the answer isn't simple, particularly when something appears in the head scan of a health volunteer.

"Typically, I would say the head ones are a little problematic," Hohmann says. "One approach you could take is to say that everyone who gets a head scan needs a clinical scan read by a radiologist, and if you find something bad then you have to deal with it."

The other approach is to tell the research participant that the scan won't be done for clinical purposes and that if something is found that needs follow-up, the participant will be referred to a physician, Hohmann says.

For studies funded by the National Institutes of Health (NIH), the first approach is the one to use. NIH requires studies funded by NIH to make a reasonable scan and have it read by clinicians. Then if something is found it will be necessary for the clinical trial team to deal with these incidental findings, Hohmann says.

"I think either of those approaches is defensible," Hohmann says. "It depends a little bit on practical issues, such as what you can actually accomplish and whether you have enough radiologists to have the scan done and to have it examined."

This can become expensive and in a worst case scenario could grind research to a halt, she notes.

"We've sought input from a radiologist group, and they each have different approaches," Hohmann says. "We don't yet have a consensus."

Here are the major categories of incidental findings:

Routine lab tests and screening of health volunteers. In this category, a research study might involve a new vaccine. As part of the screening of health volunteers, investigators will determine whether the participant has anemia or syphilis, for example, Hohmann says.

Those are incidental findings with a straightforward answer: Investigators know how to handle these findings and can make appropriate referrals.

"They need to be clear with volunteers up front and say, 'If you don't want this kind of information, you shouldn't be in the study,'" Hohmann says.

Also, investigators should tell participants that the research study will not pay for their medical care once the diagnosis is made, and obtaining treatment will be the volunteer's responsibility, she adds.

Radiologic findings. These include head scans and total body scans. Handling these findings is trickier because the findings could be serious or insignificant, and researchers will have to make decisions about how to handle these on a study-by-study basis, she says.

Genetic incidental findings. Another area in which incidental findings become an issue is in genomic research.

"With genomics our approach frequently is to tell people that we don't know what these things are going to mean," Hohmann says. "But if they're worried about breast cancer susceptibility, for example, then they should go to their own doctor and get a test done formally through the clinical laboratory as part of their own health care."

"Our IRB has thought about that issue up front and wants investigators to tell people, 'We may get information that's important to your clinical care, and this is how we'll handle it,'" Hohmann explains.

The other nuance of reporting incidental findings to participants involves determining whether the finding is of a well-understood medical problem.

For example, if genomic research uncovers an alpha1-antitrypsin deficiency in a participant, this is a finding for which there is a great deal of information, Hohmann says.

"You would probably want to tell the person that this could be very serious for his health and here is what he needs to do to carry on," she says.

With alpha1-antitrypsin deficiency there is a diagnostic test that other labs can duplicate, and there is a course of clinical care that can be prescribed once the diagnosis is made.

This isn't true of many incidental findings.

"For many we'll be in early stages of scientific development, and maybe this is the first study looking at this marker," Hohmann says. "So you don't want to panic because you have found something you think might be of concern because maybe it won't be."

IRBs also might request that investigators de-identify the genetic information they collect before testing.

"But a lot of researchers don't want to do that because they want to be able to go back and make clinical correlations," Hohmann says.

Mental health and sociobehavioral findings. Examples of this category are findings that research participants are clinically depressed or that pregnant participants are drinking heavily, Hohmann says.

The key is for the IRB to determine the researcher's responsibilities in these cases.

"Maybe this wasn't the focus of research, and maybe the researcher was studying something else among pregnant women," Hohmann says. "You have to think about these things up front and how you're going to approach them."

These types of findings also pose legal concerns.

"Illegal behaviors, in general, are problematic for IRBs," Hohmann says. "A large part of IRB's work is to anticipate these in advance."

With preparation they will have a set course of action to take, and they can advise investigators to design studies in a way that would be less likely to uncover these types of incidental findings, she adds.

"You have to look at mental health findings carefully with respect to your state's regulations," Hohmann says. "There's a legal issue and an ethical issue."

These are hard lines to draw, but IRBs could at least support investigators in doing what's right clinically, but not reporting these findings to legal authorities unless they are required by law to do so.

"You'd never get anybody to enroll in your study if you used draconian approaches to reporting and told people about it," Hohmann notes.

Researcher have preconceived notions. IRB members also need to be aware of their own preconceived notions about various types of incidental findings.

For instance, some research professionals assume that when a complete scan is conducted it is a good thing to find a problem and send the participant into medical care, Hohmann says.

"They think, 'We'll find the brain tumor early,'" and that's the natural assumption related to that approach," she says.

However, in medicine it rarely works out so perfectly.

"I think people forget about the findings of things of no importance which require more scans and extreme amounts of anxiety and invasive procedures to diagnose," Hohmann explains. "Then we find out it was nothing, and we didn't need to do anything about it."

Worse, a patient could have complications related to the procedures.

"It becomes very complex, and you can argue either side of the equation," Hohmann says. "Everybody on IRBs is aware of some apocryphal case where a person had a brain tumor, but those incidences are unusual."

It's challenging for IRBs to make decisions about this issue because they rarely have access to outcomes for decisions made in either direction.

"I go to IRB meetings every week, and we discuss these things a fair amount," Hohmann says. "We don't often hear the outcome of these things, in part because once you find something then the person by definition has a medical problem and is treated outside the research context at that point."

The chief information IRBs receive is of this nature: "We may hear of adverse event reporting," Hohmann says. "'The college student had a lesion head scan; was referred to clinical care.'"

So IRBs will think about incidental findings when they are reviewing a protocol and will offer investigators directions on how to proceed, she says.

"For NIH-funded studies we have a more complete approach," she says. "In other studies, we consider the issues and see what investigators propose."

IRB members also need to keep in mind that investigators want to give information back to participants, and they're often convinced their work will solve medical problems.

"But that's rarely the case for a single disease or research project," Hohmann says. "It takes many studies to produce the test that says, 'We can do this and we know what it means.'"

Researchers cannot take a short cut to scientific certainty just because they might have found a genome that is associated with sudden death according to a single study, she notes.

"These are complex [ethical] areas, and they depend on the state of science in a given field," Hohmann says.


  1. Managing incidental findings in human subjects research from imaging to genomics. University of Minnesota, Minneapolis; May 2007.