AWARDEES: Peter Bearman, Barbara Entwisle, Kathleen Mullan Harris, Ronald Rindfuss, and Richard Udry
SCIENCE: The National Longitudinal Study of Adolescent to Adult Health
FEDERAL FUNDING AGENCY: National Institutes of Health
It’s hard to believe now that the most important adolescent health research study ever conducted was almost over before it got underway. What began as a study of just one stage of life—driven by social science curiosity and public-health concerns—has become a landmark example of how longitudinal research can yield extraordinary and unexpected insights.
The National Longitudinal Study of Adolescent to Adult Health—Add Health, for short—is a gold standard basic research study on human health, funded by the National Institutes of Health (NIH), which has had profound impacts on society and helped shape the national conversation around adolescent health. With its revolutionary open-access model – nearly all of the survey data is publicly available as soon as it has been collected and processed – it has made possible an enormous range of scientific studies that have strengthened our understanding of the importance of family connectedness to adolescent health, allowed us to track and scrutinize the rising tide of the obesity epidemic, and demonstrated the social, behavioral, and biological importance of adolescence to lifelong health and well-being.
The Add Health study has been to social sciences what a major telescope facility would be to the astronomical sciences. But unlike a typical telescope, which can observe in only one narrow wavelength range at a time, Add Health has the ability to observe many, many wavelengths of human health and behavior at once. The nationally representative sample and longitudinal social, behavioral, environmental, biological, and genetic data, spanning over 20 years, has enabled over 10,000 individual researchers to investigate myriad unanswered questions and publish almost 3,000 research articles on human health.
Yet the scientists behind Add Health faced significant political opposition before they even began their initial proposed study of adolescent sexual behavior because, to some, it sounded like the federal government was funding a “teen sex study.” To some a teen sex study seemed inappropriate, especially using federal funds. Critics feared that asking teens questions about their sexuality could do more harm than good. Some caricatured the early work as a simple survey of whether or not teens were having sex, which they ridiculed as wasteful; the then high prevalence of teen pregnancy, they said, answered that question.
The truth was much more complex, and this research was badly needed. In the 1980s and into the ‘90s, the world was in the grips of a terrifying AIDS epidemic. Tens of thousands of Americans were dying each year, and there was very little understanding of how the disease was spreading, especially among one of the youngest and most vulnerable of populations—adolescents.
The American Teenage Study
In response to reports from scientific advisory bodies, the NIH issued a call for proposals from researchers to design a comprehensive study of adolescent sexual behavior. In 1987, University of North Carolina at Chapel Hill researchers Ronald Rindfuss, Richard Udry, Barbara Entwisle, and Peter Bearman won one of these competitive contracts to design a study that could answer many questions about adolescent behavior, with particular attention to sexual and other risky behaviors. And that is where the seeds of scientific progress—and the political problems—began.
This proposed “teen sex study” raised red flags for Members of Congress and conservative organizations, and both began investigating what the NIH was funding. So it wasn’t until 1991 that the UNC team was finally able to submit a proposal and win a grant—the largest ever issued in the social and behavioral sciences at the time—to perform the comprehensive study they had worked so hard to design: the American Teenage Study. But just months after the grant was awarded, the ongoing political inquiries led to a nearly unprecedented action by Secretary of Health and Human Services, Dr. Louis Sullivan: he rescinded funding just as the researchers were preparing to conduct their first wave of research interviews.
After a series of heated arguments in Congress and dueling newspaper opinion pieces, 1993 saw three important amendments included in critical legislation enacted to reauthorize the NIH. The first forbade the NIH from funding the American Teenage Study—along with another study on adult sexual behavior. The second required that any future studies of sexual behavior address public health concerns and be approved by both peer review and ethics boards. The third required NIH to fund a “prospective longitudinal study on adolescent health” that would consider all behaviors related to their health—thus implicitly including sexual behavior. This new law buried the American Teenage Study for good, but it meant that NIH would move forward with a broader study that could explore the important open questions affecting all social and behavioral aspects of adolescent health.
Enter Add Health
In 1994, a team led by Drs. Udry and Bearman, joined by newcomer Dr. Kathleen Mullan Harris, built upon the years of thorough planning and design work by Drs. Udry, Bearman, Rindfuss, and Entwisle to design and propose what would become The National Longitudinal Study of Adolescent Health, or Add Health. Their new, expanded study was selected through the NIH’s competitive peer review process as the best proposal for meeting the goals of the congressionally mandated study on adolescent health.
When the Add Health team began its work, they set out to answer the question: what factors are important to adolescent health and health behaviors? With a nationally representative sample of more than 20,000 adolescents, recruited through their schools, the study focused on the social contexts in which adolescents live their lives. Much prior work had been done studying the health of individuals, but no one had taken a comprehensive look at the social context. Researchers, it turned out, were missing the forest for the trees. As the Add Health study has demonstrated, the forest is critically important.
One of the most significant aspects of Add Health has proven to be its unexpected longevity, enabling researchers to track not just social influences on health but also the impact of adolescence on future phases of life. But had it not been for some serendipitous “sabotage” along the way, Add Health might never have been able to follow its subjects beyond the first two years of the survey.
The serendipitous road to longitudinal research
In 1996, the team was preparing to do its first set of follow-up interviews with the adolescents they had first surveyed two years earlier. In Wave II, as they referred to it, they planned to also collect biological samples of urine and saliva to test for sexually transmitted infections, including HIV. But possessing positive test results would have required reporting to the study participants’ state health offices, and the researchers had promised the participants that their names would never be revealed in connection to any of the data being collected. So the researchers planned to immediately and permanently destroy all identifying information once the biological data were collected. This would allow the researchers to get direct measures of the prevalence of sexually transmitted infections in their nationally representative sample while still honoring their participants’ privacy.
The field interviewers contracted to collect the data had other ideas.
As social survey workers—not biological data collectors—the field interviewers proceeded to flout the carefully planned protocols during a pretest of the Wave II interviews. They sent in poor quality biological samples—they were too little, or not kept at the proper temperature, or kept so long the needed lab tests could not be run. The Add Health team thus abandoned biological data collection in Wave II. Without test results that required reporting, the researchers were able to keep the identifying information on their adolescent cohort. Thanks to this act of unplanned “sabotage,” Add Health continues today and is moving into its fifth wave of interviews with their original adolescent cohort, who are moving through their fourth decade of life, yielding insights the original design team never imagined.
This was not the only moment of serendipity involved in the project. In order to address Dr. Udry’s interest in the impact of genetic factors on human health, the project intentionally oversampled siblings of all types, especially twins. For some twins, the team did not have sufficient information to tell if they were identical or fraternal. But there was one surefire way to tell: DNA. So in the early days of Add Health, the team collected DNA for a small subset of its twins. And seeing the tremendous value in this added dimension of data, Add Health has expanded its DNA collection in each of Waves III and IV, now encompassing the study’s entire cohort. This unique, robust dataset allows researchers to examine the interplay of genetic and environmental factors that influence health and behavior over more than two decades.
There is yet another important element of serendipity related to this research: the mid-to-late 1990s saw the rise of the obesity epidemic in the United States, one of the most important human health developments of the past 20 years. This hit adolescents just as Add Health was starting up, and it can be seen clearly in the data. At Wave I in 1994-95, just 11 percent of the surveyed adolescents were obese; at Wave III, six years later, in 2001-02, the percentage doubled to 22 percent; and by the time of Wave IV, in 2008, a whopping 37% were obese. Hundreds of scientific studies based on Add Health data have delved into the causes and consequences of obesity, including critical work on the ways in which genetic and environmental factors interact to increase the risks of obesity for young people.
With its integrated social and biological data across over two decades of life, Add Health has provided many surprising insights. Using Add Health data, researchers have discovered that unhealthy conditions like obesity can actually spread among friend groups in ways analogous to contagious viruses, and they’ve shown that relationships with family, friends, neighbors, and schoolmates actually help adolescents cope with the stressors of daily life and protect them against health risks like obesity, diabetes, hypertension, and inflammation into adulthood some 15 years later.
Today, some of the very same organizations that opposed the original survey on which Add Health was based recognize the value of the resulting research and frequently cite its data. Thanks to the persistence of the researchers, their innovative integrated social and biological study design, a series of serendipitous events, and the persistent support of the National Institutes of Health and several champions in Congress, Add Health is the largest and most comprehensive study ever of adolescent health and its implications for adult well-being, and it continues to profoundly impact the ways we understand and act to improve human health.