Leaders of a multi-billion-dollar study that aims to track the health of 100,0000 US children told senators unequivocally this week that their abandonment of a nationally representative sampling strategy is a done deal, a leaked document shows.
This document, provided on 16 April to a private US Senate briefing by Alan Guttmacher and Stephen Hirschfeld, leaders of the National Children’s Study (NCS), states that the study “will no longer construct a ‘national probability sample’ or be ‘generalizable’ ” to the entire US population, as originally planned.
The study has been the subject of perennial controversy since it was authorized by Congress in the Children’s Health Act of 2000. The administration of former US president George W. Bush repeatedly tried to zero out its funding, which Congress never failed to restore. In 2009, its top leadership was removed after it emerged that the 25-year study might cost nearly twice the US$3 billion originally slated for it. Last month, two advisers on an external panel resigned in protest, after the National Institutes of Health (NIH) declared to Congress in February, without the panel’s knowledge or input, that it would be abandoning the nationally representative study design in favour of a “well-described cohort followed longitudinally”.
The leaked document suggests that this change is now in effect despite online declarations that the study’s sampling strategy is still open for discussion. According to the leaked document, using the geographic sample of 105 counties initially included in the study with the intent of representing all US children “would [not] yield sufficient families within either a scientifically sound timeframe or a fiscally sound budget”.
Instead, it says, the NCS will employ a “hybrid” design, recruiting subjects from health-maintenance organizations and patching holes by enrolling additional subjects at government clinics serving poor and vulnerable populations.
The Obama administration’s proposed budget figure for the NCS in 2013, the Senate briefing document notes, contains a 15% reduction in study funding, to $165 million, that is predicated on implementing these changes to the study design.
The declarations made in the document handed to the Senate are in line with what leaders of the nascent study of environmental and biological influences on children’s health from before birth to age 21 told Congress in this budget document in February. However, on the same day that they delivered the bluntly worded document privately to the closed-door Senate meeting, Guttmacher and Hirschfeld published a much more undecided-sounding paper on the public website of the panel of external advisers that is supposed to advise the study’s leaders.
The paper, posted Monday in preparation for the NCS Federal Advisory Committee‘s next meeting, on Tuesday 24 April, is entitled “National Children’s Study: Potential Sampling Strategies”.
Its first words declare:
“This document is intended to structure discussions about the future NCS Main Study sampling design. The proposed sampling strategies presented are examples and not an inclusive list of all options. Alternative proposals for sampling strategies are encouraged during this period of design evaluation.”
Asked why they seem to be telling their advisers and the public that all study-design options are still on the table, while telling senators that a national probability sample has been definitively ruled out, Guttmacher, who, as director of the National Institute of Child Health and Human Development is the senior NIH official overseeing the study, responded in this statement today. It says, in part:
“The National Children’s Study is in a period of design evaluation. The Senate briefing document and the white paper on sampling are consistent in the central message that the national probability sample with door to door household recruitment is not a feasible model for a study with the goals and the scale and scope of the National Children’s Study.”
Some of the principal investigators at the study’s 37 pilot sites were nonetheless angered by the appearance of duplicity. “They can’t lie to the Senate but they can certainly lie to us. Most of us are pretty disgusted,” says Nigel Paneth, a professor of epidemiology and paediatrics at Michigan State University in East Lansing and principal investigator at the NCS site in Wayne County.
“How do you tell one group one thing and everybody else the other? And why are we wasting time if the decision has been made?” says another disgruntled principal investigator who asked not to be named.
Those who, like Paneth, oppose the NIH’s move to abandon a national probability sample argue that the study’s value will be substantially diminished if this happens. Studying a cohort of children that is not representative of the country as whole will restrict researchers’ ability to test a full and representative range of both rare and common conditions and exposures in American children, they argue. What’s more, they say, the findings that do result will not be generalizeable to all children in the country.
The briefing with Guttmacher and Hirschfeld in the Senate Dirksen Office Building on 16 April was attended by about a score of aides to senators who are particularly concerned about the proposed changes in the study. It was arranged by staff of Senator Thad Cochran of Mississippi, the senior Republican on the Senate Appropriations Committee, which, with its House counterpart, is the final overseer of spending on the NCS.
Cochran received a copy of this letter to Hirschfeld, the NCS director, from a local board of advisers at a Children’s Study site. The board at the University of Mississippi Medical Center in Jackson, after they learned of the proposed changes to the study design, wrote that, although they had entered the study with great excitement and in good faith, “we are now questioning all that we have been told” by NCS leaders. It continues: “The Mississippi Delta is the closest thing our nation has to a third world country; abandoning [the two Mississippi NCS sites] seriously undermines the validity of the NCS.” (There are no health-maintenance organizations in Mississippi, although these would be the primary recruitment sites under the NIH’s new plan.)
The letter writers add that they are hopeful that the NIH’s announced changes will be followed by “careful examination of the evidence and conversation with key constituencies. Should no such intentions be forthcoming, we strongly suggest that the title National Children’s Study be abandoned and a new, more honest title reflecting the limited sampling design be adopted.”