Predicting Disability among Community-Dwelling Medicare Beneficiaries Using Claims-Based Indicators

Area of Science:

• Health Services Research
• Biostatistics
• Gerontology

Background:

• Identifying disability in community-dwelling populations is crucial for resource allocation and policy.
• Existing claims-based algorithms have limitations in accurately identifying individuals with disabilities.

Purpose of the Study:

• To evaluate the feasibility of using current claims-based methods to identify disabled Medicare beneficiaries.
• To enhance disability identification by developing predictive models combining multiple claims indicators.

Main Methods:

• Utilized data from 12,415 community-dwelling fee-for-service Medicare beneficiaries (2003-2006).
• Employed logistic regression models predicting self-reported disability using six claims-based indicators.
• Assessed model performance using Receiver Operating Characteristic (ROC) curves.

Main Results:

• Regression-based predictive models outperformed individual claims-based indicators.
• For beneficiaries aged 65+, sensitivity was 0.72 and specificity 0.65.
• For beneficiaries under 65, sensitivity was 0.54 and specificity 0.67.
• Identified strategies to improve predictive performance for specific disability groups.

Conclusions:

• Predictive models integrating multiple claims-based indicators offer a superior tool for identifying individuals with disabilities in Medicare claims data.
• This enhanced methodology aids researchers in more accurately identifying disability populations.