Doctors judge female surgeons more harshly when patients die

By THE DATA TEAM

GENDER bias can be hard to measure, especially when it comes in subtle forms. A recently published working paper by Heather Sarsons of Harvard University provides an unusually clear example. It shows that, with all other factors held constant, the gender of surgeons influences how likely doctors are to refer patients to them.

Ms Sarsons examined data on surgical referrals in America that took place in 2008-12. She measured how referrals from a given doctor to a given surgeon changed when the outcome of a particular referral to that surgeon was unusually good or bad. An unusually good referral was one in which a patient from the riskiest 1% survived the operation and was not readmitted within a month. An unusually bad one was when the patient died within a week.

Ms Sarsons used a statistical matching procedure to create pairs of surgeons—one male and one female. Both members of the pair had performed the same surgery on similar patients (taking into account sex, age, race and pre-existing medical conditions) with either unusually good or unusually bad outcomes. And at the time of the referrals in question, the surgeons had similar professional characteristics, such as length of experience and number of procedures completed.

She then looked, in each case, at the subsequent relationships between the referring doctor and the surgeon in question. Because of the matching process, the female and male surgeons in each pair had, in the year leading up to the successful or failed procedure being examined, almost the same number of referrals from the doctor who had chosen them for that procedure. If the procedure resulted in rapid death, the number of referrals made by the doctor to the surgeon in question fell by 34%—but only if the surgeon was a woman. Male surgeons saw no lasting fall in subsequent referral rates. Even when a female surgeon did receive subsequent referrals from the doctor, these tended to be for less complicated cases than those referred to men. Nor was this bias affected by the sex of the referring doctor. Female doctors exhibited it as much as male doctors did.

Bias in favour of the abilities of male surgeons also emerged from the data on those surgeons who had had an exceptionally good patient outcome. Following such an event, the doctors who had referred the lucky patients more than doubled their referrals if the surgeon was male. They also increased them if the surgeon was female, but only by 72%.

That women are judged more harshly than men for the same mistake and rewarded less for the same accomplishment may come as no surprise. The more sobering implication of the study is that many patients who might otherwise be treated by a brilliant female surgeon are referred to a mediocre male one instead.