Using Predictive Analytics for Early Identification of Short-Term Disability Claimants Who Exhaust Their Benefits

Area of Science:

• Occupational health
• Insurance analytics
• Rehabilitation medicine

Background:

• Early interventions are crucial for short-term disability insurance (STDI) claimants to return to work after off-the-job medical conditions.
• Effective targeting of interventions requires identifying claimants likely to exhaust STDI and transition to long-term disability insurance (LTDI) without support.

Purpose of the Study:

• To identify factors predicting STDI exhaustion and LTDI transfer.
• To evaluate if a waiting period before intervention enhances targeting efficiency for disability claims.

Main Methods:

• Utilized a large US private employer-sponsored STDI claims database.
• Employed logistic regression and ROC curve analysis on a split sample.
• Assessed predictive accuracy by omitting claims resolving within 2, 4, and 6 weeks.

Main Results:

• Age, primary diagnosis, and employer industry were significant predictors of STDI exhaustion and LTDI transfer.
• Targeting accuracy increased substantially with waiting periods: STDI exhaustion prediction rose from 63.2% to 82.9% by week 6.
• LTDI transfer prediction accuracy improved from 63.7% to 83.0% by week 6.

Conclusions:

• A brief waiting period (e.g., 6 weeks) can significantly enhance the accuracy of early intervention targeting for disability claims.
• Predictive modeling, combined with a waiting strategy, efficiently narrows the target population, potentially reducing intervention costs.
• Consideration of potential trade-offs related to delayed intervention is essential before implementing a waiting strategy.