Systems Learning from Physician Performance Data.

Area of Science:

• Health Services Research
• Primary Care Medicine
• Quality Improvement

Background:

• Physician performance data, including Health Plan Employer Data and Information Set (HEDIS) measures, were collected in 1998.
• Reports were distributed to 194 primary care physicians across 25 medical centers.
• Research aimed to assess the reliability of performance measures and identify practice factors influencing patient outcomes.

Purpose of the Study:

• To evaluate the reliability of individual and aggregated physician performance measures.
• To identify practice components associated with improved patient outcomes.
• To inform strategies for enhancing primary care delivery and patient care.

Main Methods:

• Retrospective analysis of administrative data on physician performance and practice structures.
• Utilized 23 HEDIS measures, including cancer screening, diabetic management, patient satisfaction, and ambulatory costs.
• Employed bivariate and multiple regression models, controlling for patient and physician characteristics.

Main Results:

• Aggregated HEDIS measures demonstrated higher reliability for physician performance assessment compared to individual measures.
• Practice coordination, characterized by shared practice, team tenure, and clinic size, correlated with better cancer screening, diabetic management, and patient satisfaction.
• Physician performance assessment for those with reduced or part-time hours was linked to enhanced cancer screening and diabetic management.

Conclusions:

• Collective analysis of physician performance data provides valuable insights into clinical practice.
• Identifying effective primary care practice structures and processes through performance data analysis can significantly benefit patient care.