Researchers are increasingly studying how cost sharing affects healthcare utilization, costs, and patient outcomes, especially with the rise of high-deductible health plans (HDHPs) in the US. In 2023, 29% of covered workers had a HDHP. Claims data can be valuable for analyzing HDHPs, but there are challenges in using them. A recent study by Cliff et al. (2024) assesses methods to predict plan deductibles using Optum Labs data, with promising results.
The study evaluates four different imputation approaches for predicting deductibles, including parametric predictions and modal methods. Surprisingly, the simple “mode method” performed best in classifying enrollees into HDHP vs. non-HDHP plans and predicting deductible spending accurately.
For more details and insights, you can access the full paper here.
