Abstract: Clinical order patterns derived from data-mining electronic health records can be a valuable source of decision support content. However, learned patterns can be compromised by underlying provider experience as clinical orders of less experienced providers (trainees) can exhibit greater variability and worsened patient outcomes than those of experienced providers (attendings).We find that attending and trainee clinical order patterns learned by association rule episode mining converge to comparable results, aligned with clinical practice guidelines.
Learning Objective 1: Evaluate the influence of clinician experience on the consistency of clinical order patterns learned from data-mining electronic health record data.
Learning Objective 2: Use clinical practice guidelines as a reference standard for evaluating machine-learned decision support content.
Jason Wang (Presenter)
Alejandro Schuler, Stanford University
Nigam Shah, Stanford University
Jonathan Chen, Stanford University