NIH Releases New Design for Children's Study

There's both good news and bad news for researchers who had worried about a new design released this week for the troubled National Children's Study, a $3 billion plan to follow 100,000 children from before birth to age 21.

The good news is that the study's overseers at the National Institutes of Health (NIH) want to recruit from a geographic sample of hospitals rather than from large health care providers. However, to reduce costs, NIH wants to recruit about half these women when they give birth, not before, which NIH staff admit will compromise studies of prenatal exposures. The new plan makes no mention of retaining the 105 randomly chosen counties where the study was originally to take place, including 40 "vanguard" sites where pilot studies have been done.

Requested by Congress in 2000, NCS aims to explore links between environmental exposures, genetics, and health. So far it has recruited about 3650 babies at the vanguard sites and spent nearly $1 billion. In February, NIH announced that it was trimming NCS's budget and could not afford to recruit women for the full study as first planned—by knocking on doors or using other methods in the 105 counties. Instead, the agency said it might recruit participants through large health care organizations. NIH also announced that the 40 vanguard sites run by academic researchers would be turned over to a new contractor.

Vanguard site leaders and some NCS advisers were outraged. They argued that only a probability-based sample would yield findings that apply to all U.S. children, regardless of ethnicity or family income.

After a meeting in April where vanguard investigators urged NIH to stick with the 105-county plan, NIH now says it will recruit through randomly chosen hospitals and birth centers in a sample of geographic areas that "have yet to be determined." Two "layered cohorts" of women and babies would be recruited: half at birth at the hospitals, and the other half pregnant women enrolled through providers who deliver at those same hospitals. While these cohorts would make up 90% of the study participants, two smaller cohorts might include women enrolled before conception.

Epidemiologist Nigel Paneth of Michigan State University in East Lansing, who leads a vanguard center, says the new plan could result in a nationally representative sample. "It's a far better design than we had been hearing about," he says. However, he wonders why NCS can't recruit only prenatally. This would allow for rigorous studies of prenatal influences on health and make it easier to arrange to collect placentas and cord blood, he notes. He also questions NIH's apparent decision to abandon the vanguard sites where researchers have built relationships with hospitals. Those running NCS "probably don't realize how difficult it is," he says.

NIH's new proposed design will be discussed on 24 July by the NCS advisory committee.

Even if advisers endorse the new plan, what happens next isn't clear. A draft House of Representatives bill directs the agency to hold off on any design changes until the Institute of Medicine has done a study. Paneth says investigators are worried that NIH will issue a request for proposals for the main study this fall to avoid the House directive, which likely won't become law until the end of the year.