Any research involving people requires that scientists go through a lengthy process of getting informed consent and approval from ethics panels. But does testing techniques on dead people constitute research? That question has stumped scientists who have tried to use new methods to identify victims of disasters.

After the events of 11 September 2001, scientists went through the proper channels to study the day's impact on survivors. But it never occurred to those leading the identification of the victims that their project required approval.

New York City's office of the chief medical examiner led the identification project, the largest ever to use DNA profiling. Robert Shaler, then the agency's director of forensic biology, says the project shouldn't have needed approval because investigating deaths is the office's mission.

This is the clearest example I've heard of in a long time of what is not research. Arthur Caplan, University of Pennsylvania

The US National Institutes of Health (NIH) thought otherwise.

Shaler had suggested using single-nucleotide polymorphisms (SNP), or minor variations in genes, to analyze extremely degraded DNA fragments found at the World Trade Center site. Because the use of SNPs was still novel at the time, members of NIH's advisory board thought the project required approval from an institutional review board (IRB).

“They said we needed IRB approval because we were using new technology,” says Shaler, now director of forensic science at Penn State University. But “these [were] not human subjects, it's not doing research on human subjects.”

In the end, the disagreement didn't slow identifications. Shaler applied for approval and the city health department's IRB sided with him, waiving approval in August 2003. The first SNP identification was made that December, and SNPs produced 20 of the 850 identifications made based solely on DNA results. Scientists eventually identified 1,594 people using DNA-based techniques and more traditional forms of identification, such as dental records.

The experience, says Shaler, is a clear illustration that research agencies must clarify what constitutes human subject research—particularly during disasters.

The NIH declined repeated requests for comment on the topic. But the Department of Health and Human Services, the NIH's parent agency, doesn't view the use of new technology to identify disaster victims as research, says Christina Pearson, a spokeswoman. The agency only requires IRB approval for work on living individuals, she says.

The NIH scientists clearly misunderstood US research regulations, says Arthur Caplan, director of the University of Pennsylvania's Center for Bioethics. “This is the clearest example I've heard of in a long time of what is not research.”

Approval from an IRB wasn't an issue after Hurricane Katrina, because scientists were able to use standard forensic methods.

Ideally, scientists would use new technologies before a disaster occurs, Shaler says, but the federal government must prepare for other possibilities. “These things should be automatic,” he says. “We shouldn't have to jump through hoops.”