Savitz, David A.a; Ness, Roberta B.b
The National Children's Study is a longitudinal cohort study initiated by the US National Institute of Child Health and Human Development (NICHD), in collaboration with the Centers for Disease Control and Prevention, US Environmental Protection Agency, and National Institute of Environmental Health Sciences, to examine a wide range of potential influences on the health of children. The study is designed to enroll over 100,000 children from across the US and to monitor their health and exposure from before birth through age 21 years.1,2 The scope of potential health influences includes environmental contaminants, social and behavioral factors, and genetics; health outcomes of interest include obesity, asthma, and neurodevelopmental disorders. Enrollment has begun at the 7 Vanguard Centers, where the study has been reframed to focus on feasibility, acceptability, and cost, and thus serve as a pilot for expansion to the additional 29 sites selected for initiation of the main study.
There are compelling arguments for initiating a very large, multipurpose, broadly representative prospective study of pregnancy and child health—all key features of the National Children's Study. However, despite some recent changes in the study leadership and recognition of the need to reconsider key methods, we maintain that this study is still in trouble. The problems call for fundamental change in the study methods and the decision-making process.
VALUE OF THE NATIONAL CHILDREN'S STUDY
The predecessor of the National Children's Study is the Collaborative Perinatal Project—with enrollment from 1959 to 1966 and follow-up of the children up to age 7 years. The Collaborative Perinatal Project was tremendously valuable as a platform from which groundbreaking scientific discoveries have emerged. Results include the recognition that neurodevelopmental disorders originate in utero rather than in labor and delivery, the importance of inflammation as a cause of adverse pregnancy outcome, and the nonrandom nature of sudden-infant-death syndrome and its association with maternal smoking.3 However, there have been large shifts in exposures, lifestyles, and medical practice in the last 50 years, as well as new opportunities for biomonitoring, measurement of the chemical, physical, and social environment, and behavioral assessment; these changes argue for an updated data-collection effort. Moreover, we now know that in utero and childhood experiences are profoundly important in determining the course of health throughout life. The opportunity to generate a forward-looking treasure trove of information for the next generation of scientific discoveries and public health advances has engendered strong public, political, and scientific support.
REASONS FOR CONCERN ABOUT THE FUTURE OF THE NATIONAL CHILDREN'S STUDY
With its vision established nearly a decade ago, and planning and preparation having consumed many years and millions of dollars, there are several indications that the path forward for the National Children's Study remains murky.
Concerns With Household Recruitment
The ambitious plan adopted by the National Children's Study to enroll women who are likely to become pregnant (as opposed to those already pregnant) offered the exciting potential to study children from conception forward, regardless of whether or when women sought prenatal care. This novel approach competed with a more pedestrian design to recruit from prenatal care sites, where pregnant women gravitate (over 95% obtain such care before the third trimester) and where most comparable studies are based. Valid reasons to consider household recruitment are (1) the appeal of including pregnancies among those who are uninsured, do not attend prenatal care, or otherwise would be missed and (2) the ability to study events of very early pregnancy. Initial estimates for the household recruitment strategy were that 10-40 households4 would need to be approached for every enrolled birth. Five years elapsed (2004-2009) before the National Children's Study attempted to implement this untested strategy and found that response and yield of pregnancies was markedly lower than expected: 510 enrolled pregnancies thus far from 83,152 households contacted (163 households per pregnancy).5 This raises concern about efficiency as well as selection bias from nonresponse. The study is now considering the more traditional approach of recruitment from prenatal care sites, but has not resolved the primary mode of recruitment.
Concerns With Field Methods for Probability-based Sampling
A related goal was to use probability-based sampling, a highly desirable study feature for ensuring generalizability and applicability of results to national policy. Recruitment of nationally representative probability samples is often accomplished by a central contractor or agency (eg, the National Center for Health Statistics, which conducts national surveys) that is able to efficiently conduct geographically dispersed, population-based enrollment. The extent of centralization (with 1 or a few primary data collectors, typically a contract research organization) or decentralization (among some number of academic centers or other local entities) is fundamental to the study's execution. There are successful models of both, with the Women's Health Initiative employing a widely distributed (albeit tightly standardized) recruitment network, and the Danish6 and Norwegian7 birth cohorts conducted centrally, with onsite recruitment in clinics but only one base of operations. To engender grassroots involvement in the study, the National Children's Study selected a model in which local academic institutions competed for the work of enrolling households from preselected county subareas. This resulted in contracts awarded to investigators with widely disparate levels of experience, who are proposing disparate field methods and requesting highly disparate budgets. It is not clear that the cost efficiency or quality control of alternative methods of data collection received the consideration they deserved.
Concerns With Organizational Structure and Study Management
Organizational problems have hampered the National Children's Study from its inception, and continue to engender doubts about the future course of the study. First, decision-making authority and organizational structure were and remain vague. Dozens of working groups involving hundreds of scientists began deliberations in 2002, and made recommendations on the study design and protocol, yet there was no process for reconciling those priorities with one another or with practical aspects of feasibility. There are multiple advisory groups representing the lead federal agencies, the extramural research community; an executive committee consisting of selected Center investigators; and oversight by the National Children's Study Program Office, the Director of NICHD, and the Director of NIH—yet how the major decisions affecting the future of the National Children's Study will be made remains unclear. A review conducted by the National Research Council and Institute of Medicine generated an extensive list of specific suggestions,8 which were clearly considered carefully by the National Children's Study team.9 Even so, fundamental questions regarding the structure of the study and its operations (feasibility, decentralized data collection, adequacy of response) did not result in fundamental change. Who has the authority and responsibility to decide, for example, to revert to recruitment from prenatal-care sites or to centralize some aspects of data collection if such changes are found to be more favorable? What is the mechanism for center investigators—who are the most knowledgeable and experienced field researchers associated with the study—to reshape the protocol as needed?
Second, key aspects of the study methods remain unresolved, including methods of recruitment and the protocol for data collection. To decide where the study should be done and who is capable of doing it, one has to know what the study will entail. The lack of a clear protocol has led to an extended, expensive, and not-fully informative pilot phase. The relegation of the initial data-collection efforts to function as much-needed pilot studies, with methodologic work planned for the future, is a positive development. But this change has not been reconciled with the dozens of contracts in place for recruitment based on a protocol that is no longer operative. Thus, the study is being launched on a large scale despite the lack of a final protocol for recruitment, for follow-up, or for data collection, and thus with uncertainty about fundamental aspects of study design.
The scope and timeline have been delayed and modified repeatedly, inevitably resulting in variable and inefficient prerecruitment efforts (eg, community engagement to generate support for enrollment) across sites. Leadership changes at the lead funding agency, NICHD, resulted in part from a reprimand by the Senate Appropriations Committee (which has supported funding of the study) when the Committee perceived a lack of candor regarding costs.10 However, in the absence of a final study design and protocol, the funding needs remain speculative.
As exemplars, the Danish National Birth Cohort7 and the Mothers and Babies8 study in Norway (each characterizing over 100,000 children enrolled in utero) successfully implemented pregnancy cohort studies on a national scale and in a cost- and time-efficient manner. These studies, each led by a very small group of investigators, took an early and consistent approach to their specifications and made tough decisions about what aspects to promote and what to curtail, based on a shared, explicit vision for the study.
RECOMMENDATIONS FOR DESIGN AND MANAGEMENT OF THE NATIONAL CHILDREN'S STUDY
We argue that putting the study on track requires an organizational structure with clear lines of responsibility and authority that can promptly finalize and implement a plan for the National Children's Study. A broad range of experts is needed to make the key decisions, but there needs to be a mechanism for translating that collective wisdom into a feasible, efficient protocol that optimizes the scientific and public-health benefits. We offer our suggestions for a way to do so.
Feasibility must be reconciled with scientific aspirations in rethinking the study design. Lofty ambitions were an appropriate and laudatory starting point for the National Children's Study, but some of the ideas generated by innovative investigators were not feasible for large-scale studies, were mutually incompatible, or were simply too burdensome. Specifically, it is time to abandon household recruitment and the accompanying desire for preconceptional enrollment, and revert to enrollment through geographically distributed and at least broadly representative prenatal-care sites. The participating centers would need to prove their ability to implement this new protocol, with new contracts awarded corresponding to the new protocol. This approach to recruitment requires no pilot testing because it is known from many previous studies to be effective. Although there is a growing recognition on the part of the National Children's Study leadership of the need to reexamine household recruitment, there is no clear plan for implementing the cascade of needed changes in a timely manner.
A logical order for decisions and actions needs to be established. The protocol for recruitment and initial data collection needs to be finalized and the best method to implement that protocol determined; only then can contracts be implemented for data collection. Conflating pilot-testing, methodologic studies, initial data collection, and expansion to multiple sites is incompatible with study quality and efficiency.
Leadership by a multidisciplinary team of experienced field researchers is critical for success. An authoritative management team is needed, consisting predominantly of scientists who are recognized experts in the key disciplines and who have experience with analogous studies of pregnant women and children. This team would solicit expert advice from multiple sources, but have the latitude and obligation to make the decisions needed to get the National Children's Study back on track, decisively and with transparency, to avoid further delays. A small group of carefully chosen experts, including epidemiologists, clinicians, social scientists, statisticians, geneticists, and federal research managers, needs to be given the charge and clear authority to reshape the National Children's Study. They would benefit from carefully examining the lessons learned thus far, but not be obligated to sustain any features of the ongoing effort other than the fundamental goal of the National Children's Study itself—conducting a large, comprehensive, geographically representative study of the health of children in the United States.
Given the importance of the study to children's health, and the substantial amount of money and time invested, aggressive efforts are needed to ensure that the National Children's Study moves forward toward attaining its goals. New leadership of the National Children's Study is now examining key elements of the study methods, but the very organization of the study calls for reconsideration and possibly fundamental change. While change in something as large and complex as the National Children's Study inevitably will cause consternation, the corrective actions are straightforward and entirely feasible. We cannot predict the many opportunities and challenges that will unfold over the life of the National Children's Study, but they will surely come, and the study must be ready to address them openly, wisely, and efficiently.
ABOUT THE AUTHORS
DAVID SAVITZ chaired the National Children's Study Panel on Sample Selection in 2004 that recommended the study design. He served as the initial Principal Investigator of the Duplin County, North Carolina Vanguard Center and was a member of the Executive Steering Committee. ROBERTA NESS was the Principal Investigator for the Westmoreland and Marion County Pennsylvania National Children's Study sites, and also served on the Executive Steering Committee. Both authors have directed multidisciplinary NIH-funded prospective cohort studies of pregnancy outcomes.
1. Scheidt P, Dellarco M, Dearry A. A major milestone for the National Children's Study [commentary]. Environ Health Perspect. 2009;117:A13.
2. Landrigan PJ, Trasande L, Thorpe LE, et al. The National Children's Study: a 21-year prospective study of 100,000 American children. Pediatrics. 2006;118:2173–2186.
3. Klebanoff MA. The Collaborative Perinatal Project: legacy 50-year retrospective. Paediatr Perinat Epidemiol. 2009;23:2–8.
5. Park CH. Vanguard study recruitment and study visit update. Paper presented at: NCS Federal Advisory Committee Meeting; January 14, 2010; Rockville, MD.
6. Olsen J, Melbye M, Olsen SF, et al. The Danish National Birth Cohort—its background, structure, and aim. Scand J Public Health. 2001;29:300–307.
7. Magnus P, Irgens LM, Haug K, et al. Cohort profile: the Norwegian Mother and Child Cohort Study (MoBa). Int J Epidemiol. 2006;35:1146–1150.
8. National Research Council and Institute of Medicine. The National Children's Study research plan: a review. Panel to review the National Children's Study Research plan. Committee on National Statistics, Division of Behavioral and Social Science and Education. Board on Children, Youth, and Families and Board on Population Health and Public Health Practice, Institute of Medicine. Washington, DC: The National Academies Press; 2008.
© 2010 Lippincott Williams & Wilkins, Inc.