Related Experiment Videos
Development and validation of a psychiatric case-mix system.
Kevin L Sloan1, Maria E Montez-Rath, Avron Spiro
1VA Puget Sound Health Care System, and the Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, Washington 98108-1597, USA. Kevin.Sloan@med.va.gov
Medical Care
|May 19, 2006
Summary
A new case-mix system, the PsyCMS, effectively predicts mental health and substance abuse healthcare costs and utilization. This validated model offers improved accuracy over existing systems for psychiatric populations.
Area of Science:
- Health Services Research
- Psychiatric Epidemiology
- Health Economics
Background:
- Risk adjustment for mental health (MH) and substance abuse (SA) populations presents significant challenges.
- Existing risk-adjustment models are not specifically designed for patients with psychiatric disorders.
Purpose of the Study:
- To develop and validate the "PsyCMS" (Psychiatric Case-Mix System).
- To predict concurrent and future MH/SA healthcare costs and utilization.
Main Methods:
- Utilized data from 914,225 veterans in the Veterans Administration (VA) healthcare system for fiscal year 1999.
- Derived diagnostic categories from ICD-CM codes based on DSM-IV definitions and clinical input.
- Developed weighted least-squares regression models and compared PsyCMS predictive ability against other case-mix systems using R-squares and MAPEs.
Main Results:
- The PsyCMS demonstrated predictive ability for MH/SA costs (R=0.11 concurrent, R=0.06 prospective) and utilization (outpatient R=0.25 concurrent, R=0.07 prospective; inpatient R=0.13 concurrent, R=0.05 prospective).
- The PsyCMS outperformed other examined case-mix systems, showing higher R-squares and lower Mean Absolute Prediction Errors (MAPEs).
Conclusions:
- The PsyCMS provides clinically meaningful categories.
- It exhibits strong predictive capability for MH/SA costs and utilization.
- Represents a valuable tool for predicting mental health service needs and expenditures.