Estimating annual charges for ambulatory care from limited utilization data

Area of Science:

• Health Services Research
• Healthcare Economics
• Medical Informatics

Background:

• Accurate estimation of ambulatory care charges is crucial for healthcare financial management.
• Existing methods may lack aggregate utilization measures, especially when charge or cost data is absent.
• This study addresses the need for reliable utilization-based charge estimation.

Purpose of the Study:

• To identify essential utilization information for estimating annual ambulatory care charges.
• To develop and validate predictive models for aggregate utilization measures.
• To provide a method for estimating charges in the absence of direct cost or charge data.

Main Methods:

• Utilized charge and utilization data from the RAND Health Insurance Experiment.
• Grouped services by California Relative Value Studies (CRVS) codes and clinical meaningfulness.
• Developed and validated linear regression models to predict annual charges.
• Assessed model predictive accuracy using adjusted R2-values and cross-validation.

Main Results:

• A comprehensive model explained 84% of the variance in ambulatory care charges.
• Key predictors included provider visits (medical and mental health), prescription days, and procedure/test utilization.
• Validated models predicted 77% of variance and mean charges within 102% of actual values.
• A simplified model with four categories explained 77% of charge variance.

Conclusions:

• Broad service utilization categories are effective for predicting annual ambulatory care charges.
• These models offer a viable approach for estimating healthcare costs when specific data is limited.
• The findings support the use of utilization data as an aggregate measure for financial planning.