On Thursday, the journal Nature published a study that purportedly shows that telemed chemical abortions pose few health risks. The study identified over 6,000 women who had obtained chemical abortion pills via telehealth. It found that they had a complication rate similar to that of women who had obtained chemical abortions after an in-person medical exam. Unsurprisingly, this study quickly received sympathetic coverage from several mainstream media outlets including the New York Times, National Public Radio, Forbes, Bloomberg, and U.S. News and World Report.

The main methodological problem with the study is that it relies on survey data and therefore fails to track health outcomes from all of the women who had obtained chemical abortions.  Specifically, the study identifies over 6,154 women who had obtained them. However, it received a follow-up contact for only about 4,613. That means that the researchers have no idea what happened to nearly a fourth of the women who had obtained chemical abortions via telehealth. That clearly biases the results, as women who suffered adverse effects after obtaining a chemical abortion would probably be less likely to respond to a follow-up survey.