Adjusting for comorbidities in disease expenditure studies: An application to United States national inpatient data

One of the major challenges in tracking health expenditure is allocating expenditure for a health care event to a single disease when comorbidities are present. Disease expenditure studies tend to ignore comorbidities although it is increasingly an issue, in part reflecting the epidemiologic transition from acute to chronic diseases. In this study, we propose a regression-based method for comorbidity adjustment, particularly relevant for a health system encounter-based approach. Encounter-based approaches assign expenditure to disease based on the coded diagnoses. Without comorbidity adjustment, the expenditure is assigned entirely to the primary diagnosis and the estimates are biased.

We apply this framework for comorbidity adjustment to the National Inpatient Survey data in the United States for years 1996-2010. For each disease, comorbidity adjustment factors are generated separately by age and sex. Our preliminary estimates suggest that mental health problems such as unipolar depressive disorders have the greatest increase in expenditure after adjustment, whereas dementia has the greatest decrease in expenditure after comorbidity adjustment.

Our methodology takes a unified approach to comorbidity, which not only accounts for excess expenditure that needs to be distributed away from a primary diagnosis, but also the inflows of expenditures that needs to be redistributed towards each disease acting as a comorbidity for other primary diagnoses. Correcting for comorbidities provides a more accurate estimates of the expenditure incurred by specific diseases.