Introduction: The proposed Radiation Oncology Alternative Payment Model (RO-APM) aims to test prospective episode-based payments for radiotherapy episodes. Practices will need a tool that can calculate historical episode reimbursements to succeed in this new model. An automated software-based technology was created to calculate historical episode reimbursements within a large Network of community oncology practices.
Materials and methods: Claims data between January 1, 2017, and July 31, 2019, were cleaned, organized into episodes, and analyzed with a series of Python computer programs per proposed RO-APM methodology. Averaged Winsorized historical episode reimbursements were first calculated over the entire Network, then over 24 of the largest Practices, and then rerun after application of Clinical Rules to remove misattributed episodes.
Results: A total of 79,418 RO-APM-defined episodes were generated from 6,512,375 claims lines. A total of 7,086 episodes (8.9%) were removed because of no treatment delivery code within 28 days of treatment planning. The Network of practices had more bone metastases, and breast, cervical, and uterine cancers but less lung and prostate cancer than the RO-APM dataset. Combination-modality episodes were more costly and required more providers than single-modality episodes. Clinical Rules reattributed 2,495 episodes (3.4%) and increased episode reimbursement by +5.8% over all disease sites (+3.7% using volume weighting; P = .001).
Conclusion: As payment models continue to shift from volume to value, practices will need an automated analytics technology to measure historical costs and prepare for operational and financial transformation. This automated approach can be adapted to future versions of the RO-APM. Our analysis suggests that future iterations of the RO-APM could incorporate Clinical Rules to remove misattributed palliative care episodes and could implement a separate payment for episodes with multiple radiation therapy modalities.