Massive U.S. children's study needs major tweaks, report finds

U.S. Government

Massive U.S. children's study needs major tweaks, report finds

Jocelyn is a staff writer for Science magazine.

A controversial plan to study the health of 100,000 U.S. babies to age 21 has some strong points—but also a host of weaknesses that could further delay its launch, an outside review has concluded. The critique from an Institute of Medicine (IOM) and National Research Council (NRC) panel raises questions about whether the National Children’s Study (NCS) can sustain the political support needed to assure funding for the ambitious effort, which has already cost $1 billion and could require billions more in coming years.

NCS “offers enormous potential, but it also presents a large number of … challenges,” states the 16 June IOM/NRC report. To overcome them, IOM/NRC recommends that the study’s leader, the National Institute of Child Health and Human Development (NICHD), undertake some major changes, including fine-tuning the study’s guiding hypotheses and bolstering scientific input and oversight. And it should drop an existing plan to enroll nearly half of the children at birth, and instead enroll all of them earlier, during the mother’s pregnancy.

“It’s not like [NICHD needs] to start at square zero again,” because much of the groundwork for NCS has already been laid, says panel chair Greg Duncan, an economist at the University of California, Irvine. But the needed changes would likely delay initial recruitment, now planned for 2015.

NCS grew out of a request from Congress, made 14 years ago, that the National Institutes of Health (NIH) follow a large group of children from birth to adulthood in a bid to understand how environmental factors, including social settings and chemical pollutants, influence health. Planners decided to recruit 100,000 women and their unborn babies by knocking on household doors in a random sample of about 100 U.S. counties. But this approach proved too expensive, so NICHD turned to other designs. Members of Congress expressed concern in 2012, after NICHD, having spent $1 billion, decided to pull the plug on 40 NCS sites run by academic investigators and turn over to large contractors the job of tracking the 4000 families it had already enrolled in a pilot study. In March 2013, Congress called for IOM to review the study and suggest improvements.

In general, the IOM/NRC panel endorses the NCS concept. Similar studies are under way in Europe and Japan, it notes, but the U.S. version would be more comprehensive in part because researchers would collect extensive data on environmental exposures. The panel also agreed with NICHD’s decision to design the study as a “data collection platform” that will start out focused on testing just a few hypotheses—such as that exposure to kitchen dust exacerbates respiratory problems—and add more later. But these “exemplar hypotheses” need to be more scientifically robust, IOM/NRC found.

The panel found fault, however, with how NCS planned to enroll subjects. Instead of enrolling 45,000 babies at birth and the same number before birth, the panel calculated the study could enroll 95,000 mothers during pregnancy for the same cost—if it dropped plans for separate studies of 10,000 women. Those smaller studies were to focus on particular questions such as preconception exposures and the effects of natural disasters. Enrolling siblings, as is planned, could also provide useful data on preconception exposures, the panel noted.  

Such changes would enable NICHD to enroll participants for about $1.5 billion over 7 years, the panel estimates, with annual costs peaking at just over $300 million. (NIH is now allocating $165 million per year for the study.) The IOM/NRC total appears to put NCS on a lower cost path than earlier NIH estimates, which predicted the study could cost $3 billion. But the IOM/NRC estimate does not include costs such as archiving data and storing biological samples, the report notes. If NIH must further trim costs, the panels says it would be better to enroll fewer families than cut back on exposure measurements.

The panel could not determine whether NICHD’s plan to enroll pregnant women through a sample of U.S. hospitals would yield a group as representative of the population as recruiting women through prenatal providers within the original 100 NCS counties. The problem is that “no such list [of all hospitals] exists,” Duncan says.  

Scientific leadership of NCS is a major deficiency, the report finds. Although the study has various advisory committees, “the processes by which study decisions are made and vetted are opaque,” it says. NCS needs more expertise within its program office, the panel concludes, as well as a new outside scientific advisory group that has the authority to approve the study’s design.

Such conclusions are vindication for some NCS critics, including academic scientists who were pushed out of the study 2 years ago. “From my point of view, this is an excellent report,” says pediatrician and epidemiologist Nigel Paneth of Michigan State University in East Lansing, who led a former NCS site in Detroit. “Its conclusions are essentially the same as mine. They’re just putting it more nicely.”

Now that IOM/NRC has had its say, the question is whether NIH will embrace its recommendations—and whether Congress will be willing to keep funding the study.

*Correction, 16 June, 4:35 p.m.: The report is from the National Research Council as well as the Institute of Medicine, both part of the National Academies.

*Update, 16 June, 4:35 p.m.: NIH Director Francis Collins issued a statement in response to the report saying that it “raises significant concerns.” He is putting the main study on hold. A team of experts will meet in the coming weeks to advise Collins on whether the study is “actually feasible” given current budget constraints and if so, how to implement the recommendations. If not, the panel will look at “new methods to answer key research questions that are most important to pediatric health today.”

Posted in Health, People & Events, Policy