Assessment of Bias in Patient Safety Reporting Systems Categorized by Physician Gender, Race and Ethnicity, and Faculty Rank A Qualitative Study

IMPORTANCE Patient safety reporting systems (PSRSs) are designed to decrease the risk of harm to patients due to medical errors. Owing to the voluntary nature of PSRSs, implicit bias of the reporter may affect the management of safety events reported. Stanford Alert For Events (SAFE) is the PSRS used at Stanford Health Care. OBJECTIVE To examine whether variation exists in the content of SAFE reports based on demographic characteristics of physicians who are the subject of the event report. DESIGN, SETTING, AND PARTICIPANTS This retrospective qualitative analysis from a single academic medical center evaluated SAFE reports from March 2011to February 2020. Event reports were coded by theme and categorized by severity (scale of 1 to 3, with 1 being the lowest and 3 the highest). The reports were then analyzed from October 2020 to February 2022 and categorized by physician gender, race and ethnicity, and faculty rank. A total of 501 patient safety events were collected from the adult hospital during the study period, and 100 were excluded owing to incompleteness of information. MAIN OUTCOMES AND MEASURES This qualitative study had no planned outcome. RESULTS A qualitative analysis was performed on 401 reports representing 187 physicians (138 [73.8%] male and 49 [26.2%] female). In terms of race and ethnicity, 4 physicians (2.1%) were African American, 49 (26.2%) were Asian; 7 (3.7%), Hispanic or Latinx; 108 (57.7%), White; and 19 (10.2%), declined to state. Female physicians had disproportionate representation among reports referencing communication and conversational issues and the lowest severity level. Male physicians had disproportionate representation for ignoring or omitting procedures, process issues, and physical intimidation. African American physicians had disproportionate representation for lack of communication and process issues. Asian physicians had disproportionate representation for lack of communication, process issues, conversational conduct, and the lowest severity level. Latinx physicians had disproportionate representation for conversational conduct. White physicians had disproportionate representation for ignoring or omitting procedures, verbal abuse, physical intimidation, and the highest severity level. CONCLUSIONS AND RELEVANCE In this qualitative study, female physicians and physicians who were members of racial and ethnic minority groups were more likely to be reported for low-severity communication issues compared with their male and White counterparts, respectively. These findings suggest that there may be a lower threshold for reporting events when the subject of the report is female and/or a member of a racial or ethnic minority group. Restructuring the reporting and management of patient safety events may be needed to facilitate conflict resolution in a manner that reduces implicit bias and fosters team cohesion.