Why

Chapter 7. Experimentation

How can we find causes by intervening on people and systems?

Many claims about health seem to be reversed if we wait long enough. One of the most stunning shifts is in our understanding of the link between hormone replacement therapy (HRT) and heart attacks: while initial studies found it prevented them, later studies found either no effect or even an increase in heart attacks.

The first evidence for HRT’s benefits came from the Nurses’ Health Study,¹ which drew its power from surveying an enormous cohort of registered nurses (nearly 122,000), who were followed up with every two years since the initial survey in 1976. Analysis of the data in 1997 found that postmenopausal HRT users had a 37% lower risk of death and that this was due in large part to fewer deaths from coronary heart disease.

While guidelines emerged suggesting that HRT can be used to reduce the risk of cardiovascular disease,² a study published just one year after the Nurses’ Health Study found it had no effect on coronary heart disease. Unlike the Nurses’ study, which just observed people’s behavior, the HERS trial³ randomly assigned patients to either HRT or a placebo. While the study followed only 2,763 women over four years, it raised questions because the incidence of heart attacks in the HRT group actually increased in the first year of the study (though this effect was reversed in the last two years). The Women’s Health Initiative randomized controlled trial recruited a larger ...