Oncology Care Model Linked to Reduced Therapy in Poor-Prognosis Cancer

Value-Based Reimbursement and Systemic Therapy Utilization

The clinical landscape of oncology is currently defined by a tension between rising incidence rates for several major cancers and a steady decline in overall mortality driven by therapeutic advances (8, 9). As the population of cancer survivors grows and the demographic ages, the healthcare system faces increasing pressure to deliver high-quality care while managing escalating costs (11, 14). Historically, fee-for-service structures have been criticized for prioritizing the volume of interventions over the value of patient outcomes. To address this, the Centers for Medicare and Medicaid Services introduced alternative payment models designed to align financial incentives with efficient, evidence-based practice (17, 18). However, concerns remain that episode-based payments might inadvertently encourage clinicians to initiate systemic therapy more frequently to trigger care management fees. A recent study examines whether these reimbursement shifts actually influenced the likelihood of starting systemic treatment across different patient populations.

Evaluating the Impact of Episode-Based Payments

The Centers for Medicare and Medicaid Services Oncology Care Model functioned as an episode payment model for patients with cancer, where clinical episodes were triggered specifically by the receipt of systemic cancer therapy. To support enhanced coordination, the program provided monthly care management payments to participating clinics. During the early years of the program, which began in July 2016, all practices were engaged in 1-sided risk (a financial arrangement where providers could share in any realized savings but were not held liable for financial losses). This structure raised clinical concerns that the model might inadvertently incentivize an increase in episode volume by prompting physicians to initiate systemic therapy more frequently to secure care management fees. To investigate this possibility, researchers conducted a quasi-experimental study using a matched difference-in-differences analysis of serial cross sections of Medicare beneficiaries. This statistical method compares the changes in outcomes over time between a group receiving an intervention and a control group to isolate the effect of the program itself. The study period included patients with an index visit for cancer from January 2010 to December 2019, with each patient followed for one year. Data analysis for this project was conducted from October 2021 to November 2025, comparing 197 intervention practices that adopted the Oncology Care Model against 197 matched comparison practices that continued under traditional reimbursement structures. The researchers evaluated two distinct groups: an incident population of newly diagnosed patients and a poor-prognosis cohort. The incident population included 754,182 patient episodes among 750,483 patients with a mean age of 74.1 years (standard deviation [SD] 9.0) and a female majority of 467,071 (62.2%). The poor-prognosis cohort consisted of 517,858 patients with a mean age of 72.4 years (SD 9.7), of whom 270,416 (52.2%) were female. By analyzing these large datasets, the study sought to determine if the financial incentives altered the likelihood of initiating systemic therapy across different clinical stages of disease.

Treatment Initiation and Financial Outcomes

The primary analysis of the incident population, consisting of newly diagnosed patients, demonstrated that the Oncology Care Model did not lead to the anticipated surge in treatment volume. Specifically, there was no statistically significant differential change in the initiation of systemic therapy within this group, with a recorded -0.9 percentage point difference (95% confidence interval [CI], -2.2 to 0.3 percentage points; P = .14). This finding suggests that the financial structure of the model, despite providing monthly care management payments for active treatment episodes, did not inadvertently drive physicians to initiate chemotherapy or other systemic agents in cases where they might not have otherwise. Furthermore, the financial impact on this incident population was relatively modest; the researchers observed a non-statistically significant relative decrease in spending of -$898.26 (95% CI, -$1890.31 to $93.80; P = .08) in the year following the index diagnosis. In contrast, the impact of the payment model was more pronounced among patients with late-stage disease. Among those in the poor-prognosis cohort, the study identified a statistically significant differential decrease in the likelihood of systemic therapy initiation of 1.5 percentage points (95% CI, -2.8 to -0.2 percentage points; P = .03). This reduction in treatment initiation for patients with limited life expectancy or advanced disease suggests a shift toward more conservative management, perhaps reflecting a greater clinical emphasis on palliative care and quality of life over aggressive systemic intervention. This clinical shift was accompanied by substantial cost savings. In the poor-prognosis cohort, there was a statistically significant relative decrease in spending of -$2192.15 (95% CI, -3559.66 to -833.63; P = .002) per patient in the year after the index visit. These results indicate that the Oncology Care Model may have achieved its goal of reducing low-value care and overall expenditures specifically within the most clinically vulnerable patient populations.

Clinical Implications for End-of-Life Care

The findings address a primary concern regarding episode-based payment models: the risk that clinicians might increase treatment volume to trigger additional payments. Because the Oncology Care Model provided monthly care management payments only when systemic therapy was administered, some analysts feared it would incentivize unnecessary treatment. However, the data from 754,182 patient episodes in the incident population (750,483 patients; mean age 74.1 years) demonstrate that the Oncology Care Model was not associated with an increase in the likelihood of initiating systemic therapy episodes among patients with incident cancers, showing a non-significant -0.9 percentage point difference (95% CI, -2.2 to 0.3 percentage points; P = .14). This suggests that for newly diagnosed patients, the model did not disrupt standard clinical decision-making or drive inappropriate utilization of systemic agents. The clinical impact was most pronounced in the cohort of 517,858 patients with poor-prognosis cancers (mean age 72.4 years; 52.2% female). In this group, the Oncology Care Model was associated with less chemotherapy initiation and lower spending among patients with poor-prognosis cancers, evidenced by a 1.5 percentage point decrease in therapy initiation (95% CI, -2.8 to -0.2 percentage points; P = .03) and a significant spending reduction of -$2192.15 (95% CI, -3559.66 to -833.63; P = .002). These results indicate that the model may have successfully encouraged a shift toward higher-value care, potentially prioritizing palliative goals over aggressive systemic interventions in patients with advanced disease. Furthermore, the researchers concluded that by not examining changes in chemotherapy initiation, the original Oncology Care Model evaluation may have underestimated savings related to the model. This suggests that future value-based oncology frameworks must account for treatment initiation rates to fully capture their economic and clinical impact on end-of-life care.