Evaluating a Model to Predict Primary Care Physician-Defined Complexity in a Large Academic Primary Care...

Area of Science:

• Health Services Research
• Clinical Informatics
• Predictive Analytics

Background:

• Effective patient risk stratification is crucial for optimizing healthcare resource allocation.
• Current risk prediction models may not fully capture patient complexity.
• Physician insight offers a valuable perspective in risk assessment.

Purpose of the Study:

• To evaluate a physician-defined complexity prediction model (ePDC).
• To compare ePDC against the outpatient Charlson score (OCS) and a commercial risk predictor (CRP).
• To assess the models' ability to predict suboptimal quality and utilization outcomes.

Main Methods:

• A predictive model for estimated physician-defined complexity (ePDC) was developed using data from 4,302 adult primary care patients.
• The patient population (143,372 patients) was categorized using ePDC, OCS, and CRP.
• Outcomes measured included incomplete cancer screening, uncontrolled HbA1c, uncontrolled LDL, emergency department visits, and hospital admissions.

Main Results:

• Physician-defined complexity (ePDC) showed modest agreement with OCS (36.7%) and CRP (39.6%).
• ePDC-identified complex patients had significantly higher rates of incomplete cancer screening, uncontrolled HbA1c, and uncontrolled LDL.
• ePDC-complex patients also exhibited higher rates of emergency department visits and hospital admissions.

Conclusions:

• The estimated physician-defined complexity (ePDC) measure favorably compares to existing risk-prediction approaches.
• ePDC demonstrates superior ability in identifying patients at risk for suboptimal quality and utilization outcomes.
• This physician-informed model enhances patient risk stratification for improved healthcare management.