A case-mix classification system for medical rehabilitation

Area of Science:

• Medical Rehabilitation
• Health Services Research
• Health Economics

Background:

• Medicare's current payment system for rehabilitation care faces dissatisfaction, necessitating a new prospective payment system.
• Developing a robust patient classification system is a prerequisite for implementing a rehabilitation prospective payment system.
• The Uniform Data System for Medical Rehabilitation provided data for 36,980 patients from 125 facilities.

Purpose of the Study:

• To create a classification system for medical rehabilitation patients.
• To support Medicare's proposed rehabilitation prospective payment system.
• To facilitate facility planning and research on rehabilitation outcomes, quality, and cost.

Main Methods:

• Utilized clinical judgment and a recursive partitioning algorithm to form classification rules.
• Developed the Functional Independence Measure version of the Function Related Groups (FIM-FRGs).
• Identified four key predictor variables: diagnosis, admission Functional Independence Measure (FIM) motor and cognitive scores, and patient age.

Main Results:

• The FIM-FRG system comprises 53 distinct groups.
• The system explains 31.3% of the variance in the natural logarithm of length of stay in a validation sample.
• The FIM-FRG classification system demonstrated conceptual simplicity and stability in validation testing.

Conclusions:

• The FIM-FRG system offers a manageable number of groups for classifying rehabilitation patients.
• This classification system may provide a viable solution for payment, facility planning, and research in medical rehabilitation.
• The FIM-FRG system is a promising tool for improving the understanding and management of rehabilitation care.