Automated Big Data Analytics for the Radiation Oncology Alternative Payment Model Proposal Using a Novel Health Care Software Technology

Abstract

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.

Impact of Radiation Oncology Alternative Payment Model on Community Cancer Centers

“…Little has been published about RO-APM outside of the final document. 1 However, the existing literature suggest that RO-APM will result in reimbursement reductions, 4,5,[7][8][9][10] particularly affecting cancers more common in underserved populations 5 and cancers treated with advanced technology. 9 Specific examination of RO-APM's effect on FS community practices suggests that episode reimbursement is too low and should be increased by 5.8% over all disease sites.…”

“…9 Specific examination of RO-APM's effect on FS community practices suggests that episode reimbursement is too low and should be increased by 5.8% over all disease sites. 10 An additional factor affecting FS community centers is that combination-modality external beam and brachytherapy often have multiple sites of service, which could lead to under-reimbursement under RO-APM. It is unclear how these multiple sites of service and or multiple providers may be handled under RO-APM, but the final rule encourages participants to access instructions through the Medicare Learning Network (MLN Matters) publications, webinars, and the RO Model website.…”

Payment Methodology for the Radiation Oncology Alternative Payment Model: Implications for Practices and Suggestions for Improvement

“…The inclusion of palliative cases with definitively treated disease sites does not only ultimately simplify the payment methodology but also decrease the calculated National Base Rate payment for each disease site by averaging in lower-cost episodes. 5,18 Furthermore, this payment methodology uses a 90%/10% weighting of the historical experience versus the National Base Rate. In the present model, historically efficient practices (eg, early adopters of hypofractionation) would face more substantial negative adjustments relative to inefficient practices (ie, practices with high-cost episodes at baseline because of the use of more expensive technology, higher unit reimbursement such as HOPDs, or longer fractionation schema).…”

“…Such algorithms have been shown to reattribute 3.4% of episodes and increase overall reimbursement by 5.8%. 18 CMS could then separately create a palliative care episode for these reattributed episodes. 5 Furthermore, creating combined adjustment factors that are disease site-specific could better incorporate the historical costs required to manage complex cases that may be seen in low frequency.…”

Radiation Oncology Alternative Payment Model's Impact on Small and Rural Practices

“…They were told that because CMS tried to exclude extremely rural areas, they are probably not as rural as they think they are without providing any further explanation as to how CMS defined rural. 10 13 Such uncertainty in financial impact is worrisome to a small practice and contradicts CMS's goal of providing predictable payments.…”