Population choice and variable selection in the estimation and application of risk models

Area of Science:

• Health economics
• Biostatistics
• Healthcare management

Background:

• Risk adjustment models are crucial for healthcare policy and resource allocation.
• Accurate prediction of medical expenditures is essential for health plan management.
• The transferability of risk models across diverse populations remains a challenge.

Purpose of the Study:

• To evaluate the transferability of risk adjustment models across different populations.
• To assess the predictive power of demographic, employment, and administrative variables for medical expenditures.
• To inform policymakers on the appropriate application of risk models.

Main Methods:

• Development and comparison of risk adjustment models using demographic and employment variables.
• Analysis of models across populations from a single employer in different health plans.
• Evaluation of models across populations from a single health plan with different employers.
• Statistical assessment of variable predictive power for medical expenditures.

Main Results:

• Risk adjustment models showed statistically significant differences when applied to different populations.
• Demographic and employment variables demonstrated limited transferability between populations.
• Employment-based variables, such as length of employment, had minimal predictive power.
• Administrative variables were found to be not useful in predicting medical expenditures.

Conclusions:

• Risk adjustment models are not easily transferable across populations due to significant variations.
• Employment-based variables should not be included in future risk models for large employers.
• Policymakers must exercise caution when applying risk models to different demographic and employment groups.