A two-stage super learner for healthcare expenditures

Abstract

AbstractObjectiveTo improve the estimation of healthcare expenditures by introducing a novel estimation method that is well-suited to situations where data exhibit strong skewness and zero-inflation.Data SourcesSimulations, and two sources of real-world data: the 2016-2017 Medical Expenditure Panel Survey (MEPS) and the Back Pain Outcomes using Longitudinal Data (BOLD) datasets.Study DesignSuper learner is an ensemble machine learning approach that can combine several algorithms in order to improve estimation. We propose a two-stage super learner that is well suited for use with healthcare expenditure data by separately estimating the probability of any healthcare expenditure and the mean amount of healthcare expenditure conditional on having healthcare expenditures. These estimates can be combined to yield a single estimate of expenditures for each observation. The method can flexibly incorporate a range of individual estimation approaches for each stage of estimation, including both regression-based approaches and machine learning algorithms such as random forests. We compare the performance of the proposed two-stage super learner with a one-stage super learner, and with multiple individual algorithms for estimation of healthcare cost under a broad range of data settings in simulated and real data. The predictive performance of alternative strategies was compared using Mean Squared Error and R2.Principal FindingsOur results indicate that the two-stage super learner has better performance compared with a one-stage super learner and individual algorithms, for healthcare cost estimation under a wide variety of settings in both simulations and empirical analyses. The improvement of the two-stage super learner over the one-stage super learner was particularly evident in settings when zero-inflation is high.ConclusionsThe two-stage super learner provides researchers an effective approach for healthcare cost analyses in environments where they cannot know the best single algorithm a priori.