QnAs with Shoham Choshen-Hillel and Alex Gileles-Hillel

Pain is a ubiquitous experience that can harm mental and physical well-being.In healthcare settings, physicians must make decisions about pain management largely based on a patient's self-reported pain level.Biases held by physicians have the potential to skew their decisions.In a recent PNAS article ( 1 ), Shoham Choshen-Hillel, at the Business School of the Hebrew University of Jerusalem in Israel, Alex Gileles-Hillel, at Hadassah Medical Center in Israel, and their colleagues explored potential sex-based biases in emergency department pain management.The authors analyzed emergency room (ER) data covering 21,851 patients with pain complaints in Israel and the United States.The analysis indicated that nurses and physicians were consistently less likely to provide pain-relief medication to female patients compared with male patients.In an experiment involving 109 healthcare practitioners, the authors found that participants judged the pain experienced by female patients in clinical vignettes to be less intense than pain experienced by male patients.Such sex bias has significant societal and medical implications.PNAS spoke with Choshen-Hillel and Gileles-Hillel about their research. PNAS: What was the motivation for your research?Choshen-Hillel: It all started with a prior paper that we published in PNAS (2).We were interested in investigating how night shifts affect physicians' empathy and pain management decisions.We did find a pretty large effect, where physicians in our experiments had less empathy [] after a night shift than when they slept at home.During the night at the emergency department, physicians prescribed less pain medication for the same pain complaints.We controlled for many factors there, to try and convince first ourselves, then the journal, that it is really night shifts that drive physicians' empathy and pain management decisions.My PhD student, Mika Guzikevits, who is the first author on the new paper (1), said, "Look, in the regression analysis where you control for all these variables, you see that some of the variables you control for are actually predictive on their own, for example, gender."She started searching the literature, but she hardly found any papers looking at gender.We decided to collect more data and run controlled experiments.