Patient Experience Surveys Reveal Gender-Biased Descriptions of Their Care Providers

Dylan Haynes, Anusri Pampari, Christina Topham, Kathryn Schwarzenberger, Michael Heath, James Zou, Teri M. Greiling

Research output: Contribution to journalArticlepeer-review

Abstract

Patient experience surveys (PES) are collected by healthcare systems as a surrogate marker of quality and published unedited online for the purpose of transparency, but these surveys may reflect gender biases directed toward healthcare providers. This retrospective study evaluated PES at a single university hospital between July 2016 and June 2018. Surveys were stratified by overall provider rating and self-identified provider gender. Adjectives from free-text survey comments were extracted using natural language processing techniques and applied to a statistical machine learning model to identify descriptors predictive of provider gender. 109,994 surveys were collected, 17,395 contained free-text comments describing 687 unique providers. The mean overall rating between male (8.84, n = 8558) and female (8.80, n = 8837) providers did not differ (p = 0.149). However, highly-rated male providers were more often described for their agentic qualities using adjectives such as “informative,” “forthright,” “superior,” and “utmost” (OR 1.48, p < 0.01)—whereas highly-rated female providers were more often described by their communal qualities through adjectives such as “empathetic,” “sweet,” “warm,” “attentive,” and “approachable” (OR 2.11, p < 0.0001). PES may contain gender stereotypes, raising questions about their impact on physicians and their validity as a quality metric which must be balanced with the need for unedited transparency. Future prospective studies are needed to further characterize this trend across geographically and racially diverse healthcare providers.

Original languageEnglish (US)
Article number90
JournalJournal of Medical Systems
Volume45
Issue number10
DOIs
StatePublished - Oct 2021

Keywords

  • Bias
  • Gender
  • Patient experience
  • Survey

ASJC Scopus subject areas

  • Medicine (miscellaneous)
  • Information Systems
  • Health Informatics
  • Health Information Management

Fingerprint

Dive into the research topics of 'Patient Experience Surveys Reveal Gender-Biased Descriptions of Their Care Providers'. Together they form a unique fingerprint.

Cite this