Stakeholders' Attitudes on Automating Data Abstraction in a Surgical Quality Improvement Program

National surgical quality improvement programs rely on a labor-intensive process of manual data abstraction. There have been few efforts to understand stakeholders' attitudes on programmatic workflows, data collection processes, or identification of modifications to workflows that might enhance efficiency. Given contemporary data science methods, the attitudes of stakeholders regarding data collection automation could have implications for future national quality program modernization efforts.

Qualitative interviews were conducted with US Department of Veterans Affairs (VA) surgical quality nurses (SQNs) and researchers who use VA Surgical Quality Improvement (VASQIP) data from a national integrated health system. All transcripts were analyzed using thematic analysis.

Data were obtained from 42 interviews (26 with SQNs and 16 with researchers) from 37 medical facilities across the United States. The authors found five themes pertaining to participant's attitudes regarding (1) automating data abstraction, (2) interpreting variable definitions, (3) the need for standardizing documentation, (4) factoring in human review of data collected automatically, and (5) SQNs' job security concerns.

SQNs and VASQIP researchers expressed general interest and optimism about developing data science methods to automate data collection. However, participants considered human review a necessary part of data collection. More work is needed on how to automate data collection within existing surgical quality improvement programs.