Everyone in public health has heard of the first Nurses’ Health Study (NHS). This cohort of 121 700 nurses assembled in 1976 has generated a substantial part of what we know today about women’s health and the prevention of chronic diseases. A brilliant and competent team of investigators from Harvard University originated it 40 years ago and have conducted it since then, but as every epidemiologist knows, wonderful ideas do not necessarily translate into great studies. The NHS is the encounter of a wonderful idea (i.e., enrolling nurses in a large cohort study) and the historically outstanding response from nurses, who made the study theirs. Nurses joined, persisted, and used their unique and essential skills to make this cohort study an exemplar.
BEYOND DOCTORS
The inspiration for the NHS came from the British Doctor’s Study (BDS), which originally involved just a half-page questionnaire about smoking that was filled out in 1951 by around 30 000 male doctors, whose subsequent deaths were tracked with the help of the British Medical Association. In 1966, Frank Speizer went as Research Associate to the Statistical Research Unit, in London, with Richard Doll, who was the leading investigator of the BDS. Back at Harvard in 1968, Speizer wanted to study the health effects of oral contraceptives (OCs). This was a timely question because in 1960 virtually no women used oral contraception. By 1976, most young women had used oral contraceptives at some point in their lives. Unraveling OCs’ long-term health effects warranted investment in a large cohort study. A first attempt, which took place in 1971, consisted in enrolling the wives of US doctors. But too many (maybe 15%–20%) of the questionnaires sent to the wives were filled out by their husbands. This design did not work. That is when the idea of nurse participants came about.
The design itself was extremely audacious. The 1970s and 1980s were the heydays of case-control studies. Once properly understood—thanks to the great insights of people such as Jerome Cornfield and Olli Miettinen—the design was enormously attractive. With only a few thousand participants, one could estimate the same risk ratios for which a cohort design would have required tens of thousands of participants and many additional years to assess. Proposing the NHS went against the current. The project had to be tightly originated and designed to pass the scrutiny of funding review committees. In the early 1980s, a new generation of young epidemiologists—including Walter Willett—expanded the scope of the NHS beyond contraception to diet and nutrition and lifestyle factors.
BEYOND CASE-CONTROL STUDIES
But the challenge still remained. A cohort study is only as good as its ability to retain its participants. Losses to follow-up usually do not occur at random, and this introduces potential biases in the studied associations. The precursor of the NHS, the BDS, was protected against selection bias by its passive follow-up. The British Medical Association tracked the life status of its members and informed the investigators of deaths as they occurred. As a result, almost no doctor was lost for reasons other than death over more than 60 years. But the design of the NHS consisted in actively contacting all cohort participants every two years. Such studies are often rapidly depopulated as people get tired of being requested to fill out questionnaires and undergo clinical examinations. In this context, the approximate 90% response rate of the NHS women at each of the 20 follow-up cycles is simply astonishing. Overall, because participants may skip a cycle but still remain in the study, the NHS estimates that it has retained more than 94% of its original population. Or, better said, 94% of the original nurses have kept their original commitment to the study.
The other classical limitation of cohort studies is the little time that can be dedicated to measuring exposures in participants. But the nurses were self-organized. They could manage completing relatively complex questionnaires, taking their own blood, saliva, toenails, etc., with a very high level of technical quality.
Thanks to both the competence and skills of the investigators and the grit and collective will of the nurses, the NHS was built on two strong methodological foundations: high quality exposures and almost exhaustive ascertainment of outcomes.
An additional advantage of cohort studies over case-control studies is to offer opportunities to study many different outcomes. The narrative reviews assembled in this issue of AJPH attest that the NHS has provided a probably unrivaled perspective on the effects on women's health of lifestyle, social and mental health determinants, biomarkers, genes, and proteins on most chronic diseases, including skin disorders, mental health, occupational health, ocular health, kidney disorders, reproductive health, neurodegenerative disorders, life course exposures, and more.
GLOBAL IMPACT
Over the past 40 years the NHS has had several offspring. NHS II was launched in 1989 for the purpose of studying the health effects of oral contraceptive use and other risk factors during early reproductive life in 116 430 nurses aged 25 to 42 years. In 1996 and 2004, a cohort of 27 805 children (aged 9–14 years) of NHS II nurses were enrolled in the Growing Up Today Study (GUTS) to investigate factors that influence weight change. In 2010, NHS3 enrolled nearly 40 000 female nurses aged 19 to 49 residing throughout the United States and Canada with 13% self-identifying as members of a racial or ethnic minority. NHS3 has included male nurses since 2015. So it is entirelyappropriate that the NHS acronym now refers to the Nurses' Health Studies. The NHS have a male companion study in the Health Professionals Follow-Up Study (HPFS), which began in 1986 and was designed to examine similar hypotheses as the original NHS.
Beyond the United States, investigators from most continents have acknowledged their debt to the pioneering role of the NHS, as epidemiologists from Australia, Brazil, and Norway do in this issue of AJPH.
The impact of the NHS on public health recommendations, such as US Dietary Guidelines, Physical Activity Guidelines, and Nutrition and Physical Activity Guidelines for Cancer Survivors, has also been outstanding. The trans fat story, recounted here by Angell et al. (p1537), is telling. The deleterious effect of these solid when cold and liquid when heated types of oil on cardiovascular health was first detected in the NHS, then confirmed in other studies, transformed into policy in New York City, and then nationally, resulting in their quasi-elimination from fast food. The elimination of trans fat correlates with a reduction in population levels of cholesterol.
A PUBLIC HEALTH ANNIVERSARY
A full recounting of the NHS contributions should include a critical appraisal of its limitations and failures. The greatest criticism made of the NHS has been, paradoxically, that it can potentially be the victim of its own success. The bounty of carefully collected data was tempting to dredge; that is, to overanalyze in search of associations, some of which could be chance finds. Controversies also arose about whether the homogeneity of behaviors among US nurses hindered the identification of associations that required larger exposure variability, in particular with respect to dietary habits or social class.
But in this issue of AJPH we have opted to stress one of the historical contributions of the NHS, and perhaps the most unique: its standing out as the greatest active contribution an occupational group has ever made to science and to public health. I can only think of a couple of other situations in which industrial workers, civil servants, or employees actively contributed to scientific investigations about their workplace or about the health effects of their occupational exposures. But none which has covered as many exposures and diseases, or for which the impact has been so global. The articles published in this issue about or from the NHS are listed in Table 1. Nurses rallying around their study explain its success. It provided a model for many other cohort studies launched on other continents. This is indeed a very noteworthy public health anniversary.
TABLE 1—
Articles Published in This Issue About or From the Nurses’ Health Study
| Rubric | Article |
| Perspectives from other cohorts | The Australian Longitudinal Study on Women’s Health |
| Elsa-Brasil | |
| Norway | |
| Impact on Policy | Trans fat elimination |
| Prevention, translation, and control | |
| Evidence-based nutritional policies and guidelines | |
| Methods | Origin, methods, and evolution of the three nurses’ health studies |
| Methods for eliminating bias resulting from measurement error and misclassification | |
| Narrative Reviews | Alcohol consumption |
| Cancer: breast | |
| Cancer: colorectal | |
| Cancer: endometrial, ovarian, pancreatic and hematologic cancers | |
| Cardiovascular disease | |
| Cataract: age-related macular degeneration and glaucoma | |
| Exogenous hormone use: oral contraceptives, postmenopausal hormone therapy | |
| Genomics: telomere length, epigenetics, and metabolomics | |
| Nephrolithiasis | |
| Neurodegenerative diseases | |
| Obesity | |
| Psychiatric, psychological, and social determinants of health | |
| Reproductive health | |
| Skin outcomes | |
| Type 2 diabetes |
In 1976, 121 700 US nurses committed to a scientific adventure that shook up public health. Please join AJPH to celebrate the 40th anniversary of this scientific revolution with a clear, resounding, and unrestrained round of applause!