Evaluating differential cost growth across individual cancers: Insights from Oncology Care Model data.

Abstract

e18853 Background: Analyzing cost growth in oncology often faces significant challenges, including fragmentation of data for infused vs. oral agents and ascertaining the cost of inpatient care. How differential cost growth at the individual cancer level contributes to aggregate cost growth in oncology is often opaque. We leverage the integrated claims data provided by Medicare for its Oncology Care Model (OCM) pilot to analyze variations across individual cancers with respect to cost growth. Methods: OCM’s innovative methodology creates a natural experiment where costs are compared against a counterfactual comprised of Medicare patients seen in non-OCM practices. Leveraging this differential counterfactual provides an opportunity to gain insight into cost growth for individual cancers compared to oncology as a whole. Specifically, cost growth is measured each Performance Period (PP) with respect to the Baseline Period, from 2012 to 2015. We analytically decomposed and remodeled key quantitative factors in OCM associated with cost dynamics in oncology, including the Trend Factor (TF), which represents non-OCM cost growth. From 124,896 episodes, we sampled with replacement 19,191 episodes from 17 practices between PP1 and PP6 using empirical data distributions. We assumed neutral novel therapy and experience adjustments and then compared the overall TF to the remodeled cancer-level TF, reaggregated from the individual episodes. Results: Reallocating the TF reveals that cost growth among cancer types is highly variable in the broader Medicare population. Cost growth at the individual cancer level varied from +99.3% to -14.3%. Of the 21 OCM cancers, 18 have TFs greater than zero, indicating cost growth in the non-OCM Medicare population since the Baseline Period. Four have a TF greater than 50%. Three cancer types show decreasing costs relative to the baseline period: intestinal cancer, MDS, and CNS tumor. Conclusions: Significant variation exists across individual cancer subtypes in terms of cost growth. Aggregate analyses of cancer at large have insufficient specificity to rationalize payment mechanisms. Payment reform efforts within cancer care should directly address dynamics at the individual cancer or cancer subtype level in order to provide more valid considerations for expected resource utilization, including in future payment policies. [Table: see text]