Using Electronic Health Records to Identify the Daily Primary Provider During Hospitalization

Area of Science:

• Health Services Research
• Clinical Informatics
• Healthcare Quality Improvement

Background:

• Provider impact on clinical outcomes is significant but often unmeasured in healthcare research.
• Current methods for identifying responsible providers (manual chart review, claims data analysis) are inefficient or inaccurate.
• Accurate provider attribution is crucial for understanding outcome variations.

Purpose of the Study:

• To develop and validate an algorithm using electronic health record (EHR) data.
• To accurately identify the responsible provider for each patient-day during hospitalization.
• To overcome limitations of existing provider identification methods.

Main Methods:

• Multicenter retrospective cohort study.
• Development of an automated algorithm using EHR data.
• Validation against manual chart review as the gold standard.

Main Results:

• Manual chart review demonstrated 100% inter-rater reliability.
• The automated algorithm achieved 93% accuracy in identifying the responsible provider compared to manual review.
• High agreement was observed across 3626 patient-days from 200 randomly selected patients.

Conclusions:

• Electronic health record (EHR) data can reliably and accurately assign patients to providers daily.
• This methodology offers a scalable solution for healthcare research.
• Enables identification of non-intervention-related sources of outcome variation.