Letters to the Editor
We appreciate this opportunity to clarify why we concluded that the salary differences we observed are difficult to justify. Dr. Ambati raises three thoughtful issues.
First, he notes that we did not measure relative value units (RVUs), suggesting that perhaps the salary differences observed resulted from unmeasured differences in clinical productivity. However, we did include other measures of clinical productivity—work hours and the relative level of pay within the respondent’s specialty (a proxy for work that results in a high level of RVUs per time spent). Moreover, we deliberately selected a cohort whose K-awards require that most (usually 75%) of their effort is devoted to research rather than clinical activity, thus minimizing the expected impact of differences in clinical productivity on salaries within this sample.
Dr. Ambati also notes that we did not consider philanthropic grants or fundraising by our respondents. In a study of midcareer physician–researchers,1 we did adjust for attainment of other grants, and we nevertheless observed a similar gender difference in salary. We did not include this variable in this early-career study because individuals at such an early point in their careers (still holding active career development awards) are unlikely to have already attained substantial other grants or raised philanthropic funds to an extent that would meaningfully impact salary.
Finally, Dr. Ambati notes that we considered salary but not benefits, and women might prefer “to take a greater proportion of their compensation in the form of health insurance and other fringe benefits.” In academic medicine, benefits packages are generally standardized and set at the institutional level; we do not believe that differences in individual benefits elections meaningfully affect salary determinations. Moreover, although most men in our sample were married to part-time or nonworking spouses (whose jobs would be unlikely to provide health insurance), the vast majority of women in our sample had full-time working spouses. Thus, if anything, we suspect that the women in this sample would be less likely than the men to need to take advantage of employer-provided benefits.
Therefore, we continue to find our observations concerning. It is possible, though unlikely, that the differences observed were the result of unmeasured confounding, so we encourage institutions to conduct internal reviews with more granular data wherever possible, in order to ensure transparency, accountability, and equity in the determination of salaries in academic medicine.
Reshma Jagsi, MD, DPhil
Associate professor and associate chair, Department of Radiation Oncology, University of Michigan, Ann Arbor, Michigan; firstname.lastname@example.org.
1. Jagsi R, Griffith KA, Stewart A, Sambuco D, DeCastro R, Ubel PA. Gender differences in the salaries of physician researchers. JAMA. 2012;307:2410–2417