Medicare reimbursement and regression to the mean

Area of Science:

• Health economics
• Health services research
• Medicare policy

Background:

• Medicare's payment formula for Health Maintenance Organizations (HMOs) may lead to over or underpayment due to selection bias.
• Previous research suggests that costs for biased groups tend to regress toward the population mean, implying temporary payment inaccuracies.

Purpose of the Study:

• To investigate the impact of biased selection on Medicare payments to HMOs.
• To determine if reimbursement rates regress toward the mean for different types of biased selection.
• To simulate the financial implications of HMO-favorable selection for Medicare over time.

Main Methods:

• Analysis of Medicare reimbursement data for HMOs experiencing biased selection.
• Comparison of reimbursement regression toward the mean for cohorts biased by medical use versus demographic factors.
• Simulation modeling to assess Medicare's financial performance under HMO-favorable selection scenarios.

Main Results:

• Reimbursement regression toward the mean was observed for HMO cohorts biased by medical use.
• This regression effect was not found for cohorts biased by demographic factors.
• Simulations showed initial Medicare losses in HMO-favorable selection scenarios, but these were recouped by year seven due to regression toward the mean.

Conclusions:

• Medicare reimbursement for HMOs can be temporarily inaccurate due to selection bias, particularly concerning medical use.
• The regression toward the mean phenomenon helps to mitigate initial financial losses for Medicare over the long term.
• Medicare's payment formula may require adjustments to account for the nuances of selection bias and its temporal financial effects.