Quantifying Electronic Medical Record Usability to Improve Clinical Workflow (California)
Health care providers are increasingly constrained in the time they have to assess, diagnose, and treat patients during face-to-face office visits. Electronic medical records (EMRs) have the potential to improve effectiveness, safety, efficiency, and patient-centered aspects of care. However, EMRs can also interfere with existing clinical workflow, impede physician-patient communication, and increase providers' cognitive load. The potential negative impacts of EMR use are often linked to an incomplete understanding of clinical workflows and actual real-world EMR use patterns.
This prospective study aims to increase understanding of how clinical work is actually done in outpatient clinics that use EMRs. The study will take place in two different outpatient settings, primary care clinics, and medical specialty clinics. Indicators assessed will include EMR usage, workflow, physician-patient communication, cognitive load, and user satisfaction. Associations between indicators will be explored across study sites, provider types, patient visits, and EMR vendor. The study uses a variety of methods, including observation via video recording, surveys, usability software, and the NASA Task Load Index. This multifaceted approach is designed to provide a comprehensive assessment of usability, workflow, communication, and cognitive load.
The specific aims of the study are to:
- Measure and compare EMR use patterns.
- Measure and compare clinical workflow and physician-patient communication.
- Measure satisfaction and cognitive load.
- Explore associations between the three above aims.
This study will enable the systematic and simultaneous evaluation of multiple domains of EMR and physician-patient interaction. The interdisciplinary study design borrows methods and knowledge from communications sciences, cognitive sciences, human-computer interaction, software and systems engineering, and medical informatics and health services research. Data and methods will be shared via de-identified datasets and publicly available online open-source code repositories.