Has NIH Finalized Design for Children's Study?

Jocelyn is a staff writer for Science magazine.

Controversy continues to dog the National Children's Study (NCS), an ambitious federal plan to follow the health of 100,000 children from before birth to age 21. Next Tuesday, an advisory board will consider cost-cutting alternatives to the original study design. But some researchers involved with the study think the National Institutes of Health (NIH) has already made a decision.

Study planners at NIH originally wanted to recruit pregnant women and their babies for the planned $3 billion NCS by knocking on household doors in a random sample of 105 U.S. counties. When pilot studies begun in 2009 found this to be too expensive, NIH began testing alternatives, such as finding women through health care providers' offices in the counties. Then, in February, the NIH announced a 15% cut in the NCS's $193 million annual budget that could mean replacing the county sampling with pregnant women recruited through providers who are part of health care organizations.

This caused an uproar (and two adviser resignations): Many researchers involved with NCS said that only a geographically based sample would yield findings that apply to all U.S. children, including the poor. And abandoning the 105 counties would waste years of work on building community support, they said. National Institute of Child Health and Development Director Alan Guttmacher, whose institute runs NCS, responded that no final decision had been made.

Indeed, next week, NCS's advisory committee will consider new sampling designs described in this white paper. Some would still sample providers geographically; others would recruit women through interested health care providers; a third set would use hybrid designs.

According to Nigel Paneth of Michigan State University in East Lansing, who runs one of 40 NCS "vanguard" pilot sites, he and others have figured out ways to cut costs in the 105 counties, for example, by collecting placentas but not cord blood at birth. Paneth expects many vanguard site investigators will sign on to a proposal that they will present to the advisory board next week.

But Paneth also worries that NIH's mind is already made up. In a document dated 13 April that Paneth says NIH prepared to brief the staff members of Senator Thad Cochran (R-MS), the agency writes: "Thus, the Main Study will use ... HMOs and other health care provider networks as the primary source for recruitment." It goes on to say, "The Main Study will no longer construct a 'national probability sample.' "

"This is a rather damning illustration of how they're [NCS leaders] not being honest with us," Paneth says.

NCS leaders deny that any decision has been made. In a statement, NCS Director Steven Hirschfeld said that the three-page Senate briefing document was "only a part of the longer verbal briefing, which ... provided more detailed information." He says that "the Main Study design is still being finalized" with input from the advisory committee and other stakeholders. Hirschfeld added that NCS has already cut some costs by making changes such as using non-proprietary software and a central ethics board.

Community members in the 105 counties are also upset about the NCS changes. Last month, the community advisory board for a University of Mississippi-run NCS pilot site in Hinds County wrote Hirschfeld "to express our confusion, deep exasperation, and substantial concern."

"We feel misled," the letter says. The possibility that the 105 study sites will be dropped from the main study "is a travesty" and is especially troubling for a state that has the nation's worst pregnancy and birth outcomes and significant health disparities, the letter says. "The Mississippi Delta is the closest thing our nation has to a third world country," the letter says, and abandoning the state's NCS sites "seriously undermines the validity of the NCS."

Posted in Health