Reference letters play an important role for both postgraduate residency applications and medical faculty hiring processes. This study seeks to characterise the ways in which gender bias may manifest in the language of reference letters in academic medicine. In particular, we conducted a systematic review in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. We searched Embase, MEDLINE and PsycINFO from database inception to July 2020 for original studies that assessed gendered language in medical reference letters for residency applications and medical faculty hiring. A total of 16 studies, involving 12 738 letters of recommendation written for 7074 applicants, were included. A total of 32% of applicants were women. There were significant differences in how women were described in reference letters. A total of 64% (7/11) studies found a significant difference in gendered adjectives between men and women. Among the 7 studies, a total of 86% (6/7) noted that women applicants were more likely to be described using communal adjectives, such as “delightful” or “compassionate”, while men applicants were more likely to be described using agentic adjectives, such as “leader” or “exceptional”. Several studies noted that reference letters for women applicants had more frequent use of doubt raisers and mentions of applicant personal life and/or physical appearance. Only one study assessed the outcome of gendered language on application success, noting a higher residency match rate for men applicants. Reference letters within medicine and medical education exhibit language discrepancies between men and women applicants, which may contribute to gender bias against women in medicine.

Background Wearable consumer technology allows for the collection of a growing amount of personal health data. Through the analysis of reflected LED light on the skin, heart rate measurement and arrhythmia detection can be performed. Given that melanin alters skin light absorption, this study seeks to summarize the accuracy of cardiac data from wrist-worn wearable devices for participants of varying skin tones. Methods We conducted a systematic review, searching Embase, MEDLINE, CINAHL, and Cochrane for original studies that stratified heart rate and rhythm data for consumer wearable technology according to participant race and/or skin tone. Results A total of 10 studies involving 469 participants met inclusion criteria. The frequency-weighted Fitzpatrick score for skin tone was reported in six studies (n = 293), with a mean participant score of 3.5 (range 1-6). Overall, four of the ten studies reported a significant reduction in accuracy of heart rate measurement with wearable devices in darker-skinned individuals, compared to participants with lighter skin tones. Four studies noted no effect of user skin tone on accuracy. The remaining two studies showed mixed results. Conclusions Preliminary evidence is inconclusive, but some studies suggest that wearable devices may be less accurate for detecting heart rate in participants with darker skin tones. Higher quality evidence is necessary, with larger sample sizes and more objective stratification of participants by skin tone, in order to characterize potential racial bias in consumer devices.

No abstract

No abstract