Estimating hospital costs by diagnosis for population-based analysis

Area of Science:

• Health Services Research
• Health Economics
• Hospital Administration

Background:

• Accurate hospital cost data is crucial for analyzing population-based acute care utilization.
• Direct cost information for all discharges is often unavailable, necessitating reliable estimation methods.

Purpose of the Study:

• To compare alternative methods for estimating hospital discharge costs using available data.
• To identify the most accurate estimation model for population-based health services research.

Main Methods:

• Compared estimation models using hospital name, total inpatient costs, diagnosis, and length of stay from discharge data.
• Utilized data from multiple states and verified estimates with Maine's detailed charge data.
• Evaluated models based on their ability to explain variability in hospital charges and costs per case.

Main Results:

• An estimate using hospital name, total costs, diagnosis, and length of stay explained 77.3% of the variability in average charges per case.
• A simpler model, multiplying cost per day by length of stay, explained 76.1% of individual charges and 91.9% of average costs per case.
• Both models demonstrated high reliability in estimating hospital expenditures.

Conclusions:

• Combining hospital discharge data with routinely reported per diem costs provides a reliable method for estimating per capita expenditures on hospital care.
• This approach supports population-based analysis of acute hospital utilization and healthcare spending for specific diagnoses.