SummaryAn Avalere analysis found that lengthening episode duration from 6 months to a year would not have a large impact on the relationship between episode expenditures and benchmark costs, meaning that performance on longer episodes would not improve relative to shorter episodes.
The Oncology Care Model (OCM) is a voluntary, episode-based, oncology-focused payment model from the Centers for Medicare & Medicaid Services (CMS) that aims to improve care coordination and reduce costs for Medicare fee-for-service (FFS) beneficiaries. The model evaluates the total cost of care for 6-month episodes initiated by chemotherapy treatments that are attributed to oncology practices participating in the model. In each performance period (PP), CMS calculates whether practices earn performance-based payments or owe recoupments based on their total expenditures relative to a benchmark.
The 6-month episode duration in the OCM program presents many benefits, such as standardized observation windows and quicker feedback provided to participating practices, but does not necessarily reflect the typical length of chemotherapy treatment a Medicare beneficiary might undergo. Avalere analyzed the difference between the first day and last day of chemotherapy for beneficiaries treated for a single cancer type in the 3 years between 2016 and 2018 and found that beneficiaries in 8 cancer types had average treatment durations of longer than 184 days, or the length of a 6-month episode (Figure 1). Beneficiaries in these 8 cancer types made up 45% of the total number of beneficiaries treated with chemotherapy for a reconciliation eligible cancer type in this time. Therefore, almost half of all beneficiaries in the OCM program should be expected to receive chemotherapy treatment for longer than a single episode and will likely “re-trigger” subsequent episodes.
Instead of allowing these beneficiaries to re-trigger a follow-up episode, Avalere adjusted the OCM methodology to create year-long episodes, meaning that the treatment of many beneficiaries would only occur in 1 episode instead of 2. Avalere aligned these year-long periods to overlap with the 6-month OCM periods, so 2 OCM performance periods would overlap roughly with a single year-long period. To cleanly compare OCM methodologies and time periods, Avalere limited the analysis to OCM performance periods 3 and 4 (PP3-4) and the second year-long period. The change in episode length caused the number of total episodes to decrease by about 33%, and the average episode expenditures to increase by about 60%. However, the average benchmark costs also increased by about 60%, meaning that the performance on episodes, measured by the percent difference between expenditures and benchmark prices, was not substantially impacted (Figure 2). These findings indicate that while longer episodes may more closely mirror a patient’s trajectory of service use, they may not lead to significant changes in practice performance. As episodic payment models in oncology continue to advance, considerations for episode length to more appropriately align with advancements in treatment and duration of clinical response are critical areas to evaluate.
Avalere performed this analysis using Medicare Part A/B FFS claims and Part D prescription drug event data under a CMS research data use agreement. The total sample of beneficiaries observed in this analysis does not exceed 20% of total Medicare beneficiaries. For the traditional performance period episodes, Avalere replicated the OCM payment methodology developed by CMS, including the attribution of episodes to practices, the assignment of cancer type, and the calculation of episode expenditures and benchmark prices. For the adjusted year-long episodes, Avalere changed the duration of the episode from 6 months to 12 months, updated the practice attribution and cancer assignment, and recalculated the episode expenditures. Avalere did not adjust the prediction model variables that measure services occurring in the episode other than changing the episode duration and did not change at all the methodology for the demographic variables, including age, dual status, and HCC count. However, Avalere did change the definition of several time-sensitive variables in the prediction model to account for the change in duration, including the episode length and clean period variables.
In this analysis, Avalere observed the reconciliation eligible cancer types, which are the only types within the OCM program that count towards a participant’s performance. For the length of chemotherapy treatment analysis, Avalere observed the difference between the first and last days of chemotherapy administered, defined by CMS’ OCM Initiating Therapies list. For the performance analysis, Avalere aggregated all participants’ episodes, and averaged the episode expenditures and benchmark prices to calculate the percent difference to benchmark metric. Avalere did not observe the individual performances of any individual practices.
To receive Avalere updates, connect with us.
produces measurable results. Let's work together.