Electronic health record use factors linked to efficiency and productivity: an explainable machine learning analysis

Area of Science:

• Health Informatics
• Medical Practice Management
• Machine Learning in Healthcare

Background:

• Physician efficiency and productivity are critical for healthcare delivery.
• Electronic Health Records (EHRs) significantly impact physician workflows.
• Understanding EHR use patterns can optimize clinical operations.

Purpose of the Study:

• To investigate the correlation between specific ambulatory physician EHR usage characteristics and proxies for physician efficiency.
• To identify key EHR utilization metrics associated with enhanced physician productivity and performance.

Main Methods:

• A longitudinal cohort study analyzed physician-month EHR metadata from 413 US organizations (May 2019-April 2022).
• A multi-model machine learning classifier was developed to predict physician efficiency.
• Physician efficiency was measured by same-day chart completion rate; productivity by daily patient visit volume.

Main Results:

• The machine learning model achieved 0.74 accuracy in classifying high chart completion efficiency, associated with <1.5 day inbox turnaround and <25 min/day after-hours documentation.
• A secondary analysis (0.84 accuracy) linked high visit volumes to <4.1 min/visit EHR time outside hours and <3.2 min/visit clinical review.
• Included 218,610 physicians with over 5.19 million observations, showing average chart completion efficiency of 72.9% and 10.8 visits/day.

Conclusions:

• Specific EHR use metrics, including inbox management and after-hours documentation, can be targeted for interventions.
• Actionable insights from EHR data can foster balanced, efficient work environments.
• Optimizing EHR use can improve patient care and reduce physician time spent on EHR tasks.