Geographic variation of procedure utilization. A hierarchical model approach

Area of Science:

• Biostatistics
• Health Services Research
• Epidemiology

Background:

• Understanding variations in medical care utilization is crucial for healthcare policy and resource allocation.
• Previous methods like standardization do not fully account for differences in sample sizes or complex hierarchical data structures.

Purpose of the Study:

• To introduce hierarchical statistical models for analyzing variations in medical care utilization.
• To quantify and explain interstate variations in coronary angiography use among Medicare patients experiencing acute myocardial infarction.

Main Methods:

• Employed hierarchical statistical models to differentiate between-state and within-state variations in coronary angiography utilization.
• Modeled within-state variation using logistic regression (age, sex as predictors) and between-state variation using a multivariate normal distribution for state-specific coefficients.
• Assessed model fit and compared alternative computation approaches.

Main Results:

• Demonstrated substantial interstate variation in coronary angiography rates for Medicare patients with acute myocardial infarction.
• Found only moderate interstate variation in the influence of age and sex on angiography decisions.
• The hierarchical approach provided more detailed insights than standardization and accounted for varying state sample sizes.

Conclusions:

• Hierarchical models offer a powerful method for analyzing complex healthcare utilization data, revealing significant geographic disparities.
• Future research should incorporate patient and state characteristics to further refine these models and inform targeted interventions.