Separating public health from research a challenge
IRBs must walk a blurred line
The activities of state public health departments — including disease tracking, cancer registries and death statistics — can be a rich source of data for research.
But determining when the use of this data constitutes public health research that requires review by an IRB can be tricky, says David Perlman, PhD, former director of the Human Research Ethics Program for the New Jersey Department of Health & Senior Services.
Perlman, now an associate at the University of Pennsylvania Center for Bioethics in Philadelphia, also operates an independent ethics education company, Eclipse Ethics Education Enterprises or E4. He says public health department professionals often don't realize they need an IRB's review and waiver of informed consent for the research-related use of data they otherwise would have access to normally as part of their jobs.
Even for seasoned IRB members, the line between what is public health practice and what is public health research can be a bit fuzzy, he says. What starts as a normal public health activity may evolve into a research study.
"Even with all that experience and expertise, it was sometimes very difficult to make that determination," he says. "There was sometimes a very blurred line. You don't want to say, we'll punt on this one and go through the IRB review (if it is unnecessary) because you could actually stifle a really important public health project."
For example, Perlman says, someone collecting sensitive data from the HIV/AIDS division and using it to try to improve services being offered to that population may run across data that suggests another use for the information.
"They might say, 'This data seems to show that we ought to tell everyone in the country that the way we do it is beneficial,'" he says. "They actually then develop a hypothesis on the basis of just doing a program evaluation. Once they do that, and once they're thinking they might be able to generalize the results of their particular data to a population beyond that, that sort of blurs the line."
This issue doesn't only affect state health department IRBs, he says. Up to 15 states currently don't have specific IRBs within their state health departments, instead outsourcing IRB review to outside institutions.
Dealing with outside IRB
That's the situation Perlman found himself in when he took charge of the human research protection program for the New Jersey health department. There was no departmental IRB, and reviews were conducted by a nearby medical university IRB. This sometimes caused problems between that IRB and health department researchers, he says.
"Reports were coming up from the researchers that the IRB didn't understand what we were trying to do," he says. "They were sort of on the overprotection end of things and it was stifling a bunch of important projects we had at the public health department."
As a result, Perlman says, he worked to take control of which projects were sent to the outside IRB, and eventually helped establish a state health department IRB.
But then he was faced with the same dilemma: How would his office determine which projects were truly research that required IRB review and which were the regular functions of a public health agency? To help draw up a policy, he looked at guidance from the Council of State and Territorial Epidemiologists (CSTE), a national professional association of public health epidemiologists.
Perlman says the CSTE guidance stood out as having highly nuanced and well articulated advice about determining whether something is research. In the council's view:
An activity is a public health practice when it involves: applying proven methods to monitor a community's health status, investigate occurrences of disease, and to implement preventive measures based on existing understanding in public health sciences.
An activity is public health research when it involves: testing unproven treatments or strategies that are not known to be effective. In these cases, the collection and use of identifiable private health information requires either informed consent or a waiver of that consent.
'Just tell us'
In those latter cases, Perlman says, he often had to educate investigators that they needed to submit their projects to the IRB for review.
"People on the inside of the public health department aren't typically aware of when their projects cross the line," he says. "They'd say, 'Just tell us when we have to get this reviewed.' So we'd say 'Come to us with your project, we'll try to give you an initial determination, and when you get farther along, get back to us and we'll look at it again.'
"We tried to be as helpful as possible in doing that, and for the most part we were successful, because most people had the protection of subjects in mind."
Once a project is determined to be research, the IRB then must decide whether to grant a waiver of informed consent for the use of identifiable health information. Often the most difficult part of the decision, Perlman says, is determining whether the research could practicably be carried out without a waiver, one of the criteria the IRB must consider under the federal regulations.
"What does that mean? Does it mean it has to be one small shade away from being impossible?" Perlman says. "In the end, we sort of know impracticability when we see it."
For example, asking a researcher to get consent from 200,000 people in a database he wants to use could be described as impracticable, he says.
"However, there was a project we had that involved the census, and the researchers argued it was impracticable to get informed consent," Perlman says. "And we argued, 'No it's not.' The census worker is going to be right in front of the people who are going to be the research subjects. Why not just ask them?
"We got the sense that in a lot of public health projects, people just didn't want to do it. They thought they'd slide it in under the category of impracticability."
Perlman says IRBs that work with a school of public health should be sure that they understand the distinction between public health research and public health activities. He says an IRB should be ready to seek assistance from organizations such as the CTSE, the Centers for Disease Control or the American Public Health Association.
"I guess the best thing to do is to be humble," he says. "IRBs should be humble and say we don't have the expertise to do this, we're going to call in a consultant. Or we're going to try to educate ourselves so that we can make this determination."