Return to Work Prediction

August 15, 2020

Leading Life & Disability insurance company in the US

Problem Statement

• Inadequate system to predict accurately the length of disability for a complicated injury or illness because of which the employer is not able to provide a more adequate replacement resource for the job duties

Analytics Led Approach

• Predictive analytical model using Neural Networks to predict the time taken for return to work on incoming disability claims
• Given are high level process steps which were  followed
  • Data Pre-processing
  • Logistic Regression Model creation
  • New short-term disability claim
• Output of the model provides an estimation time for return to work

Business Impact

• A utility was designed to score the incoming disability claims against the Return – to – work predictive model

Critical Success Factors

• Helps to manage and determine claims reserving more efficiently
• Helps to predict the possible duration of a claim and total expected loss and adjustment expenses associated with it
• Recommend for better nursing care in advance to improve customer centricity and profitability
• Provide a more adequate replacement resource for the job duties