Observational Data and Clinical Trials: Editorial
One of the signal events in epidemiology in the last decade is a mammoth randomized trial that failed to confirm an established epidemiologic finding. The observational finding was the protection provided by hormone replacement therapy against cardiovascular disease, and the trial was of course the Women's Health Initiative.1 A protective effect had been consistently observed in cohort studies, supported by clinical evidence, and trusted by physicians and millions of their patients. According to the clinical trial, mortality from cardiovascular disease among the treated was, in fact, worse.
In this issue of Epidemiology we publish an ambitious paper by Hernán (who also serves as one of our editors) et al,2 in which a new analysis of the Nurses’ Health Study provides results consistent with the randomized trial. This paper is followed by 3 commentaries (by Hoover,3 Stampfer,4 and Prentice5) and responses from the authors.6,7
The questions raised are fundamental, and the stakes are high. Do the methods of Hernán et al narrow the gap between cohort studies and clinical trials? Have more conventional analyses of cohort studies been misleading? How much of the discrepancy between the trial and the cohort results is due to differences in study setting and populations, to differences in handling discontinuation of replacement hormones, or to differences in the scientific and clinical questions being asked?
Readers be forewarned: these papers lay open a controversy that will not quickly be resolved. The fact that the authors chose to respond in 2 separate rejoinders shows how deeply the controversy extends. The relation of randomized trials and observational studies has long been fertile territory for methodologic discussion. Is this new reanalysis a distracting detour, or does it augur a change in the way we do our work?
1. Writing Group for the Women’s Health Initiative Investigators.Risks and benefits of estrogen plus progestin in healthy postmenopausal women. Principal results from the Women’s Health Initiative randomized controlled trial. JAMA
2. Hernán MA, Alonso A, Logan R, et al. Observational studies analyzed like randomized experiments: an application to postmenopausal hormone therapy and coronary heart disease. Epidemiology
. 2008;19:766– 779.
3. Hoover R. The sound and the fury: was it all worth it? Epidemiology
. 2008;19:780 –782.
4. Stampfer MA. “Intention to treat” for observational data—worst of both worlds? Epidemiology
. 2008;19: 783–784.
5. Prentice R. Data analysis methods and the reliability of analytic epidemiologic research. Epidemiology
6. Hernán MA, Robins JM. Author response part 1: observational studies analyzed like randomized experiments– Best of both worlds. Epidemiology
. 2008;19:789 –792.
7. Willett WC. Author response part 2. Epidemiology