Abstract
Importance
The Oncology Care Model (OCM) was the Centers for Medicare & Medicaid Services’ first cancer-focused alternative payment model, running from 2016 to 2022. The OCM aimed to reduce Medicare spending and improve quality of care for patients receiving chemotherapy.
Objective
To evaluate the association of the OCM with changes in Medicare spending, utilization, and quality of care.
Design, Setting, and Participants
Difference-in-differences (DID) regression analysis of 6-month chemotherapy episodes attributed to practices voluntarily participating in the OCM or propensity-matched comparison practices, adjusted for beneficiary, episode, practice, and regional characteristics. Episodes for fee-for-service Medicare beneficiaries were grouped into baseline (initiated January 2014-January 2016) and intervention (initiated July 2016-June 2022) periods.
Main Outcomes and Measures
Total episode payments (Medicare spending for Parts A, B, and D, excluding OCM Monthly Enhanced Oncology Services [MEOS] payments); episode payments for Medicare Parts A, B, and D, hospitalizations, emergency department visits, and measures of quality.
Results
The study population included 739 735 Medicare beneficiaries (mean age, 73.2 [SD, 8.6] years; 59.4% female; 1 746 368 episodes) undergoing chemotherapy (ie, traditional cytotoxic therapy, targeted therapy, immunotherapy, and hormonal therapy) at 202 OCM practices and 830 165 beneficiaries (mean age, 73.1 [SD, 8.8] years; 56.6% female; 1 919 516 episodes) at 534 comparison practices. Total episode payments increased from $29 206 (baseline period) to $36 190 (intervention period) for OCM episodes and from $28 788 to $36 388 for comparison episodes, for an OCM-associated spending change of −$616 [90% CI, −$912 to −$321]). Reductions in total episode payments increased over time (−$1282 in the final 6-month performance period). Statistically significant spending reductions were observed for Part A (DID, −$176 [90% CI, −$288 to −$63]) and Part B (DID, −$340 [90% CI, −$529 to −$149]) but not for Part D (DID, −$53 [90% CI, −$216 to $111]). The OCM was not associated with significant differences in hospitalizations, emergency department visits, or quality. Accounting for MEOS payments and performance-based incentive payments, the OCM resulted in an estimated net loss to Medicare of $639 million over 6 years.
Conclusions and Relevance
The OCM was associated with modest reductions in Medicare payments during cancer treatment episodes without significant changes in care quality; payment reductions increased during the program’s last 3 years. However, the OCM incurred a net loss because these estimated savings were exceeded by enhanced services payments and performance-based payments to practices.
This difference-in-differences regression analysis evaluates the association of the Medicare Oncology Care Model (OCM) with changes in Medicare spending, utilization, and quality of care in oncology practices voluntarily participating in the OCM vs propensity-matched comparison practices.
Key Points
Question
Did the Medicare Oncology Care Model (OCM) lead to reduced spending or improved quality of care during 6-month episodes of care for Medicare fee-for-service beneficiaries receiving chemotherapy?
Findings
The OCM resulted in a statistically significant $616 relative decrease in total episode payments; however, payment reductions were offset by program payments for care coordination and performance-based incentives, yielding net losses to Medicare. There were no statistically significant differences in most measures of utilization or quality.
Meaning
Reductions in Medicare payments associated with the OCM increased over time but were insufficient to fully offset model incentive payments over the course of the model.
Introduction
The Centers for Medicare & Medicaid Services (CMS) launched the Oncology Care Model (OCM) in July 2016, aiming to reduce Medicare spending during systemic therapy for cancer while improving quality of care.1 The OCM—CMS’s first oncology-focused alternative payment model—was structured around voluntary participation by oncology physician practices. Under the OCM, 6-month episodes of care were initiated (or renewed) when a patient at a participating practice received a qualifying anticancer therapy in the outpatient setting. Participating practices could bill Medicare for Monthly Enhanced Oncology Services (MEOS) payments ($160/mo) for patients receiving chemotherapy (ie, traditional cytotoxic therapy, targeted therapy, immunotherapy, and hormonal therapy) during an active episode; this monthly payment was intended to facilitate care coordination activities that practices were required to implement for model participation.2,3 Additionally, participating practices could earn performance-based payments by reducing spending, conditional on meeting quality-of-care benchmarks.
The OCM created an array of new incentives for participating practices. Before the OCM, most US oncology practices were reimbursed under the fee-for-service model, with payments for costly infused chemotherapy drugs representing a critical revenue stream. While fee-for-service payments continued under the OCM, participating practices also faced new incentives to improve efficiency and care coordination and reduce low-value spending (eg, for avoidable hospitalizations and low-value drugs, including chemotherapies). More than 200 practices in 33 states participated in the OCM, including 122 practices that remained engaged in the OCM until its completion in June 2022. During the model period the OCM covered approximately one-fourth of all systemic cancer treatment in fee-for-service Medicare.4
A 2021 evaluation of the OCM through June of 2019 showed that the OCM was associated with an estimated savings of $297 per episode.5 Although statistically significant, these savings represented only 1% of baseline spending, and gross savings to CMS did not offset model-related spending, including MEOS payments and performance-based payments to practices. Since OCM participation required major transformation of cancer care delivery at oncology practices, it was expected that OCM-attributable savings would grow over time, as participating practices learned from their experience. This study describes the mature outcomes of the OCM—encompassing Medicare spending, utilization, and quality of care—through the end of the model in 2022.6
Methods
Study Design and Data
Using a difference-in-differences design, we estimated the association of the OCM with changes in Medicare spending, utilization, and quality throughout the duration of the OCM. Data included Medicare fee-for-service claims (Parts A and B); Part D Prescription Drug Event files; Medicare enrollment and coverage data from January 1, 2014, through September 28, 2022; and data on incentive payments provided by the CMS. Supplemental data included National Plan and Provider Enumeration System data; CMS Health Professional Shortage Area and Area Health Resource data; academic medical school affiliation data7,8; and a commercially available office-based physician file. The Abt Global institutional review board approved the study; informed consent was waived under the public benefit exemption of the Privacy Act of 1974. This report follows the STROBE guideline for reporting of a cohort study.9
Study Population
The study population comprised Medicare fee-for-service beneficiaries with cancer of any type and a qualifying 6-month chemotherapy episode attributed to an OCM or comparison practice, as identified in Medicare claims.2,5 Episodes were categorized by cancer diagnosis into 1 of 24 episode types, as per OCM rules. Episodes were included if beneficiaries were continuously enrolled in fee-for-service Medicare Parts A and B, had Medicare as the primary payer, and were not Medicare end-stage kidney disease beneficiaries for all 6 months of their episode (or until death, for which episodes could be shorter than 6 months). Episodes were attributed to the physician practice providing the plurality of cancer-related evaluation and management visits (office visits) during the episode, based on tax identification numbers.
Comparison practices were selected by propensity score matching to the 202 OCM practices in the pre-OCM baseline period.5,10 We estimated the propensity for practice participation in the OCM based on episode, practice, and market factors; each OCM practice was matched to up to 10 non-OCM practices. The comparison group provided a counterfactual to estimate the effect of the OCM.
Outcomes
Prespecified outcome measures (described below) were calculated at the episode level, except for measures of end-of-life care, which were calculated at the beneficiary level.
Episode Payments
The primary outcome assessing Medicare spending was total episode payments, defined as the sum of all fee-for-service Medicare Parts A, B, and D payments during an episode. Total episode payments exclude MEOS payments and performance-based payments. We also evaluated subcomponents of episode spending by Medicare Part (A, B, and D) and by Part B subcomponents.
Utilization
We examined occurrence of any acute care hospitalization, occurrence of any emergency department visit without hospitalization, number of outpatient office visits, and number of cancer-related outpatient office visits. The hospitalization and emergency department measures were incorporated in the OCM Performance-Based Payment Quality Score calculation (the hospitalization measure was retired as of July 2, 2018).
Quality of Care
Chemotherapy-associated emergency department visits and hospitalizations were evaluated using specifications adapted from the CMS Hospital Outpatient Quality Reporting Program chemotherapy measure, OP-35. Timeliness of adjuvant chemotherapy for colon and breast cancer was defined as chemotherapy initiation within 60 days after surgery.11,12 We assessed 2 measures of guideline-recommended use of supportive care medications: use of prophylactic white blood cell growth factors during breast cancer chemotherapy with high risk for fever and neutropenia13 and use of potent antiemetic regimens during chemotherapy with high risk for nausea and vomiting.6 Guideline-concordant antiemetic regimens included either a neurokinin 1 receptor antagonist and any serotonin antagonist or palonosetron and olanzapine.14 We examined end-of-life care quality for beneficiaries who died during or within 90 days of the end of an episode, including hospice enrollment 3 days or more before death and 3 measures of high-intensity care: hospitalizations in the last 30 days of life, 2 or more emergency department visits in the last 30 days of life, and Part B chemotherapy in the last 14 days of life.15,16,17 Enrollment in hospice 3 days or more before death was included in the OCM Performance-Based Payment Quality Score.
Analyses
We examined spending, utilization, and quality for 6-month episodes initiated from July 2, 2014, through January 1, 2016 (pre-OCM baseline), and July 1, 2016, through June 29, 2022 (intervention period, comprising eleven 6-month performance periods), extending an earlier analysis.5 Our intention-to-treat analytic design retained episodes from any of the original 202 OCM practices, even if that practice later exited the model; 122 practices remained in the model through June 2022, after accounting for practice dropout and mergers.
We used difference-in-differences regression analyses18 to estimate changes in payments, utilization, and quality associated with the OCM, adjusting for observable factors unrelated to the OCM that could influence outcomes. Difference-in-differences analyses compare changes in an outcome from the baseline to intervention periods for the treatment group (OCM episodes), relative to a comparison group (comparison episodes). We adjusted for episode, practice, and market-level factors to account for time-varying influences that affected both the OCM and comparison groups, including local COVID-19 severity (see eAppendix in Supplement 1 for full list of covariates). Information about beneficiary race and ethnicity was collected for descriptive purposes, using the Research Triangle Institute race code variable in Medicare administrative records.
As previously reported,5 we assessed for nondifferential trends in the baseline period by testing the null hypothesis that trends in measures were similar for OCM and comparison episodes during the 18-month baseline period; the parallel trends assumption was not rejected for any of the reported measures (all P > .05).10,19 For each difference-in-differences result we present point estimates with 90% CIs. The OCM evaluation used 2-sided P values less than .10 to reduce the chance of not identifying a true association (type II error). Difference-in-differences results for spending outcomes are described as changes associated with the OCM. Study findings for outcomes other than total episode payments (primary outcome) should be considered exploratory, due to the multiplicity of outcomes evaluated.
Subgroup Analyses
Because care and opportunities for savings differ for different types of cancer, we conducted additional analyses for several subgroups of episodes5; subgroups were defined as per OCM program rules.2 The subgroup of lower-risk cancer episodes was composed of (1) breast cancer episodes with hormone therapy only (anastrozole, exemestane, letrozole, and tamoxifen), (2) prostate cancer episodes with hormone therapy only (bicalutamide, degarelix, flutamide, goserelin, histrelin, leuprolide, nilutamide, and triptorelin), and (3) bladder cancer episodes with intravesicular therapy only; all other episodes were considered higher risk. Selected analyses by specific cancer type were conducted.
Net Savings and Losses
The primary outcome measure of total episode payments omitted MEOS payments and performance-based payments. An additional analysis was conducted to estimate the overall net financial effects of OCM after incorporating these additional payments.
Analyses were conducted from January 2016 through October 2023, using SAS Enterprise Guide version 7.1 (SAS Institute Inc) and Stata/MP version 18 (StataCorp).
Results
This study evaluated 1 746 368 OCM episodes among 739 735 beneficiaries attributed to 202 OCM practices, and 1 919 516 comparison episodes among 830 165 beneficiaries attributed to 534 comparison practices (across the baseline and intervention periods). Demographic characteristics of OCM and comparison episodes were generally similar in both baseline and intervention periods (Table 1; eTable 1 in Supplement 1). Episodes for low-risk breast cancer constituted the most common episode type, comprising 21% to 23% of episodes in the baseline and intervention periods. Weighted by episodes, OCM practices tended to have more physicians and higher episode volumes than comparison practices (Table 1). OCM practices were less likely than comparison practices to be hospital-owned and similarly likely to be affiliated with an academic medical center.
Table 1. Characteristics of Oncology Care Model and Comparison Beneficiary-Episodesa.
| OCM | Comparison | |||
|---|---|---|---|---|
| Baseline | Intervention | Baseline | Intervention | |
| No. of episodes | 345 881 | 1 400 487 | 405 605 | 1 513 911 |
| No. of beneficiaries | 227 113 | 603 123 | 267 285 | 667 306 |
| Beneficiary characteristics | ||||
| Sex, % | ||||
| Male | 39.7 | 40.9 | 42.2 | 43.8 |
| Female | 60.3 | 59.1 | 57.8 | 56.2 |
| Age in y, % | ||||
| <65 | 9.9 | 8.0 | 11.2 | 9.1 |
| 65-69 | 25.1 | 24.1 | 24.4 | 23.8 |
| 70-74 | 23.6 | 25.4 | 23.0 | 24.8 |
| 75-79 | 19.2 | 20.3 | 18.7 | 19.9 |
| 80-84 | 12.6 | 13.0 | 12.8 | 12.8 |
| ≥85 | 9.5 | 9.2 | 9.9 | 9.5 |
| Race and ethnicity, % | ||||
| Hispanic | 4.8 | 4.6 | 4.4 | 4.4 |
| Non-Hispanic Black | 9.0 | 8.2 | 9.2 | 7.8 |
| Non-Hispanic White | 82.7 | 82.7 | 82.6 | 82.5 |
| Otherb | 3.4 | 4.5 | 3.8 | 5.3 |
| Medicare-Medicaid dual eligible, % | 14.4 | 12.9 | 16.8 | 14.8 |
| Cancer type, % | ||||
| Low-risk breast | 23.8 | 23.3 | 23.1 | 21.3 |
| High-risk breast | 10.6 | 9.9 | 9.5 | 9.0 |
| Lung | 9.3 | 9.3 | 8.8 | 9.0 |
| Low-intensity prostate | 8.3 | 8.4 | 11.3 | 11.2 |
| Lymphoma | 6.7 | 5.6 | 5.8 | 5.1 |
| Multiple myeloma | 6.1 | 6.3 | 5.8 | 5.9 |
| Colorectal | 5.5 | 5.0 | 5.1 | 4.7 |
| Non–reconciliation eligible | 4.2 | 5.3 | 4.9 | 6.5 |
| High-intensity prostate | 3.8 | 4.2 | 4.1 | 4.5 |
| Chronic leukemia | 3.4 | 3.6 | 3.3 | 3.3 |
| Otherc | 18.3 | 19.1 | 18.3 | 19.5 |
| Higher-risk episodes, %d | 67.2 | 67.7 | 64.3 | 66.5 |
| HCC risk score, %e | ||||
| 0-0.99 | 22.2 | 20.5 | 21.9 | 20.0 |
| 1-1.99 | 23.5 | 21.1 | 23.7 | 21.0 |
| 2-3.99 | 33.0 | 32.1 | 33.2 | 32.6 |
| ≥ 4 | 21.4 | 26.3 | 21.3 | 26.5 |
| Practice characteristics f | ||||
| Hospital owned, % | ||||
| Yes | 27.7 | 16.3 | 50.2 | 32.4 |
| No | 72.3 | 82.2 | 49.6 | 66.6 |
| Missing | 0.1 | 1.5 | 0.2 | 1.0 |
| Affiliated with academic medical center, % | ||||
| Yes | 18.4 | 21.2 | 16.7 | 22.3 |
| No | 81.6 | 78.8 | 83.3 | 77.7 |
| Practice has <4 oncologists, % | ||||
| Yes | 2.8 | 1.9 | 14.6 | 10.0 |
| No | 97.2 | 98.1 | 85.4 | 90.0 |
| Total No. of episodes, quartile, %g | ||||
| First | 1.0 | 0.6 | 5.0 | 3.5 |
| Second | 3.4 | 3.1 | 15.7 | 11.1 |
| Third | 14.5 | 8.8 | 25.5 | 21.0 |
| Fourth | 81.2 | 87.5 | 53.8 | 64.4 |
Abbreviations: HCC, Hierarchical Condition Category; OCM, Oncology Care Model.
The pre-OCM baseline period comprised episodes initiated July 2, 2014, to January 1, 2016; the OCM intervention period comprised episodes initiated July 1, 2016, to June 29, 2022. All characteristics are presented at the episode level.
Includes patients with Research Triangle Institute race codes Asian/Pacific Islander, American Indian/Alaska Native, Unknown, or Other.
The most common cancer types in this group include pancreatic, bladder (high- and low-risk), gastroesophageal, head and neck, and ovarian.
Lower-risk episodes include low-risk breast cancer (breast cancer episodes with hormonal therapy only), low-intensity prostate cancer (prostate cancer episodes with hormonal therapy other than enzalutamide, abiraterone, apalutamide), and low-intensity bladder cancer (bladder cancer episodes treated with intravesicular therapy only). All other episodes are considered higher-risk episodes.
Calculated from a risk-adjustment model originally designed to estimate future health care costs for Medicare beneficiaries based on age, sex, and diagnoses. It ranges from 0.19 to 20.80; higher values reflect higher risk. Scores are normalized to a value of 1.0 among all Medicare beneficiaries; individuals with risk scores less than 1 are expected to be less costly than the average beneficiary; those with risk scores greater than 1 are expected to be more costly than the average beneficiary. A risk score of 0.5 means that costs are expected to be one-half that of an average beneficiary; a score of 2.0 means costs are expected to be twice that of an average beneficiary.
Practice-level characteristics are presented at the episode level, which is the unit of analysis for difference-in-differences models. However, the propensity score matching considered these variables at the practice level.
There were not good comparison practices for 2 very large practices that participated in the OCM; difference-in-differences findings were robust to sensitivity analyses excluding these 2 very large practices.
Total Episode Payments
Adjusted total episode payments in the baseline period were $29 206 for OCM episodes and $28 788 for comparison episodes. In the intervention period, total episode payments increased to $36 190 for OCM episodes and $36 388 for comparison episodes (Table 2). Increased spending over time was driven primarily by Part B chemotherapy drugs and Part D drugs (Table 2; eFigure in Supplement 1). The OCM-associated change in total episode payments was −$616 (90% CI, −$912 to −$321; P < .001), corresponding in magnitude to 2.1% of the OCM baseline total episode payment amount. Figure 1 shows OCM-associated changes in total episode payments by 6-month performance periods. Estimated reductions in total episode payments in the first 6 performance periods (through 2019) ranged from −$268 to −$37. The magnitude of estimated payment reductions in each of performance periods 7 through 11 (2020-2022), ranging from −$1371 to −$653, was greater than in any of the first 6 payment periods. Payment reductions exceeded $960 (the maximum possible sum of $160 MEOS payments over a 6-month episode) in 3 of the last 4 performance periods.
Table 2. Association of the Oncology Care Model With Episode Payments (Excluding Monthly Enhanced Oncology Service Payments) and Health Care Utilizationa.
| Measure | OCM | Comparison | Difference-in-differences estimate (90% CI) | Change, %b | ||
|---|---|---|---|---|---|---|
| Baseline mean | Intervention period mean | Baseline mean | Intervention period mean | |||
| Change in total episode payments (without MEOS), $ | ||||||
| Overall | 29 206 | 36 190 | 28 788 | 36 388 | −616 (−912 to −321) | −2.1 |
| Lower-risk episodes | 7339 | 7583 | 7388 | 7613 | 18 (−95 to 132) | 0.2 |
| Higher-risk episodes | 40 304 | 50 604 | 39 578 | 50 776 | −898 (−1314 to −483) | −2.2 |
| Part A payments | 6229 | 5712 | 6078 | 5736 | −176 (−288 to −63) | −2.8 |
| Part B payments | 17 286 | 21 673 | 17 029 | 21 756 | −340 (−529 to −149) | −2.0 |
| Part D payments | 6699 | 10 547 | 6736 | 10 636 | −53 (−216 to 111) | −0.8 |
| Part B payment subcomponents, $ | ||||||
| Chemotherapy drugs | 7704 | 12 095 | 7534 | 11 889 | 35 (−112 to 183) | 0.5 |
| Nonchemotherapy drugs | 2668 | 2578 | 2423 | 2620 | −288 (−370 to −207) | −10.8 |
| Supportive care drugsc | 2258 | 2077 | 2066 | 2126 | −241 (−319 to −163) | −11.0 |
| Noncancer office visits | 906 | 891 | 885 | 886 | −16 (−30 to −2) | −1.7 |
| Radiation therapy | 867 | 815 | 966 | 914 | 0 (−23 to 23) | 0 |
| Imaging | 821 | 856 | 815 | 876 | −26 (−38 to −13) | −3.0 |
| Chemotherapy administration | 637 | 663 | 670 | 696 | 0 (−13 to 14) | 0 |
| Laboratory tests | 454 | 539 | 418 | 500 | 3 (−11 to 17) | 0.7 |
| Cancer-related office visits | 393 | 388 | 355 | 353 | −3 (−14 to 7) | −0.8 |
| Other (not including MEOS)d | 2801 | 2809 | 2908 | 2961 | −46 (−90 to −2) | −1.6 |
| Hospitalizations and ED visits | ||||||
| With acute care hospital inpatient stay, %e | 27.9 | 24.0 | 26.5 | 22.7 | −0.1 (−0.4 to 0.3)f | −0.2 |
| With ED visit not resulting in an inpatient stay, %e | 23.5 | 22.1 | 24.1 | 22.7 | 0.0 (−0.4 to 0.3)f | −0.1 |
| Outpatient office visits per episode, % | ||||||
| Office visits | 21.1 | 18.3 | 20.3 | 17.8 | −0.3 (−0.8 to 0.3) | −1.3 |
| Cancer-related office visits | 5.3 | 4.9 | 5.0 | 4.7 | 0.0 (−0.1 to 0.1) | −0.5 |
Abbreviations, DID, difference-in-differences; ED, emergency department; MEOS, Monthly Enhanced Oncology Services; OCM, Oncology Care Model.
The pre-OCM baseline period comprised episodes initiated July 2, 2014, to January 1, 2016; the OCM intervention period comprised episodes initiated July 1, 2016, to June 29, 2022. Part A payments include (in order of magnitude): inpatient care at acute care hospitals, hospice care, and postacute care. Part B payments include infused and injected drugs (including chemotherapy and supportive care drugs), physician services, radiation therapy, imaging, other outpatient services, and durable medical equipment. Part D payments are for pharmacy-dispensed prescription drugs.
Percent change of the DID estimate from the OCM baseline value.
Supportive care drugs is a subset of nonchemotherapy drugs, comprising growth factors, antiemetics, and bone-modifying drugs.
Other payments include payments for services such as ambulance, chiropractor, physical therapy, occupational therapy, vision, hearing and speech services, durable medical equipment, ambulatory surgical care facility fees, anesthesia.
Contributes to OCM Performance-Based Payment Quality Score.
Difference-in-differences estimates are presented as absolute percentage (percentage point) differences.
Figure 1. Association of the Centers for Medicare & Medicaid Services Oncology Care Model With Total Episode Payments, by Performance Period.
Whiskers represent 90% confidence intervals. Performance period (PP) 1 began July 1, 2016. Each subsequent calendar year had two 6-month PPs, from January through June and from July through December.
OCM-associated changes in total episode payments varied substantially across lower- vs higher-risk episodes and by cancer episode type (Table 2; eFigure in Supplement 1). OCM was not associated with any difference in total episode payments for lower-risk episodes (including hormonal therapies for breast and prostate cancer and intravesicular therapies for bladder cancer). Among higher-risk episodes, the OCM-associated change in total episode payments was −$898 (90% CI, −$1314 to −$483). When stratified by specific cancer episode types, model-associated changes in total episode payments were statistically significant for 4 common cancer episode types: high-risk breast cancer, lung cancer, lymphoma, and colorectal cancer (eTable 2 in Supplement 1). The OCM was not associated with significant differences in total episode payments for other cancer episode types (P > .10).
Payment Components
Part B payments accounted for the largest proportion of OCM total episode payments ($21 673 during the intervention period), followed by Part D payments ($10 547) and Part A payments ($5712). The OCM was associated with statistically significant reductions in Part A payments (−$176 [90% CI, −$288 to −$63]) and Part B payments (−$340 [90% CI, −$529 to −$149]) but not Part D payments (Table 2).
Within the category of Part B payments, the OCM was associated with statistically significant reductions in payments for nonchemotherapy drugs (−$288 [90% CI, −$370 to −$207]) and imaging services (−$26 [90% CI, −$38 to −$13]), but not for chemotherapy drugs or radiation therapy services (Table 2). More than 80% of the OCM-related payment reductions for Part B nonchemotherapy drugs were attributable to a small number of supportive care drugs, including white blood cell growth factors, bone-modifying agents, and antiemetic therapies.
Utilization and Quality
The proportion of cancer treatment episodes with any acute inpatient hospitalization declined nominally from the baseline to the intervention period for both OCM and comparison episodes, with no significant association of the OCM (Table 2). Likewise, the OCM had no significant association with the proportion of episodes with any emergency department visit not leading to inpatient admission, the number of evaluation and management visits per episode (Table 2), or the proportion of episodes with chemotherapy-associated hospitalizations or emergency department visits (Table 3).
Table 3. Association of the Oncology Care Model With Chemotherapy Use, Quality of Care, and Outcomes.
| No. of episodes | OCM, % | Comparison, % | Difference-in-differences estimate (90% CI)a | Change, %b | ||||
|---|---|---|---|---|---|---|---|---|
| OCM | Comparison | Baseline mean | Intervention period mean | Baseline mean | Intervention period mean | |||
| Proportion with chemotherapy-associated hospitalizations and ED visitsc | ||||||||
| Chemotherapy-associated hospitalization | 1 177 695 | 1 264 973 | 12.8 | 12.0 | 12.3 | 11.3 | 0.1 (−0.1 to 0.4) | 1.1 |
| Chemotherapy-associated ED visit | 1 177 695 | 1 264 973 | 16.8 | 15.9 | 16.7 | 16.0 | −0.2 (−0.5 to 0.0) | −1.4 |
| Proportion starting adjuvant chemotherapy within 60 d of surgery | ||||||||
| Breast cancer | 22 645 | 23 446 | 71.9 | 71.1 | 74.1 | 72.2 | 1.1 (−0.8 to 3.0) | 1.5 |
| Colorectal cancer | 17 783 | 18 127 | 60.4 | 61.9 | 61.4 | 63.2 | −0.3 (−2.2 to 1.7) | −0.4 |
| Proportion receiving guideline-recommended supportive care medications | ||||||||
| GCSF for high-risk breast cancer chemotherapy | 17 565 | 17 806 | 85.7 | 91.0 | 87.6 | 91.6 | 1.3 (−1.0 to 3.7) | 1.6 |
| Recommended antiemetics for highly emetogenic chemotherapy | 26 423 | 30 319 | 79.2 | 83.9 | 74.5 | 78.2 | 1.1 (−1.4 to 3.6) | 1.3 |
| Proportion with specific care at the end of life | ||||||||
| Hospice enrollment ≥3 d before deathd | 213 351 | 228 714 | 58.7 | 60.1 | 57.6 | 58.9 | 0.1 (−0.8 to 1.0) | 0.2 |
| Chemotherapy in last 14 d of lifee | 213 351 | 228 714 | 11.9 | 10.6 | 11.6 | 10.4 | −0.1 (−0.5 to 0.3) | −0.5 |
| Any hospitalization in the last 30 d of life | 213 351 | 228 714 | 52.9 | 51.5 | 53.0 | 52.2 | −0.6 (−1.4 to 0.2) | −1.1 |
| ED use (≥2 visits) in last 30 d of life | 213 351 | 228 714 | 15.0 | 15.4 | 15.6 | 16.4 | −0.3 (−0.8 to 0.3) | −1.8 |
Abbreviations: ED, emergency department; GCSF, granulocyte colony–stimulating factor; OCM, Oncology Care Model.
Difference-in-differences estimates are presented as absolute percentage (percentage point) differences.
Percent change of the difference-in-differences estimate from the OCM baseline value.
Assessed among higher risk episodes.
Contributes to OCM Performance-Based Payment Quality Score.
Part B chemotherapy only.
The OCM had no significant association with the use of guideline-recommended antiemetic therapies during chemotherapy with high risk for nausea and vomiting (Table 3). Likewise, the OCM had no significant association with the use of guideline-recommended white blood cell growth factors during chemotherapy with high risk for fever and neutropenia (Table 3). Last, the OCM had no significant association with timely initiation of adjuvant chemotherapy for breast or colorectal cancer (Table 3).
Among beneficiaries who died during or shortly after an episode, the proportion of decedents with 3 days or more of hospice care before death increased nominally from the baseline to the intervention period for both OCM and comparison episodes (increasing from 58.7% in the baseline period to 60.1% in the intervention period, for OCM episodes). The OCM had no significant association with receipt of 3 or more days of hospice care or with the three measures of high-intensity end-of-life care (Table 3).
Medicare Net Savings and Losses
We calculated net savings and losses to Medicare attributable to the OCM as the estimated impact of OCM on total episode payments (gross Medicare savings), less the sum of OCM incentive payments (MEOS payments and performance-based payments to practices). Time trends in these quantities over the course of the OCM are shown in Figure 2. Overall, the OCM resulted in an estimated net loss to Medicare of $639 million over the 6-year model period. Estimated net losses in the last 5 performance periods (2020-2022) were smaller than in any of the first 6 performance periods (2016-2019), and model incentive payments were nearly offset by spending reductions in the final performance period.
Figure 2. Net Impact of the Centers for Medicare & Medicaid Services Oncology Care Model on Medicare Spending, by Performance Period.
Performance period (PP) 1 began July 1, 2016. Each subsequent calendar year had two 6-month PPs, from January through June and from July through December. MEOS indicates Monthly Enhanced Oncology Services; OCM, Oncology Care Model.
Discussion
Over 6 years, the OCM was associated with a modest statistically significant reduction in Medicare total episode payments for 6-month cancer treatment episodes. This reduction was $616, or 2.1% of baseline episode spending, with savings split across Part A (ie, spending for hospital-based care) and Part B (where reduced spending for supportive care medications and faster adoption of biosimilar drugs were primary drivers of savings).4,13 Importantly, reductions in total episode payments increased during the later years of the OCM, with estimated savings exceeding $1000 per episode for 3 of the model’s last four 6-month performance periods. Still, the OCM did not achieve the goal of reducing total Medicare outlays, because costs of model administration (monthly enhanced oncology services payments and performance-based incentive payments to practices) contributed to an estimated net loss of $639 million over the 6-year model period.
The OCM was not associated with changes in quality of care—positive or negative. Despite pay-for-performance quality measures targeting prevention of emergency department visits and timely hospice referral, the OCM was not associated with measurably better outcomes for either of these domains or for other measures of quality. On the other hand, no evidence was found that OCM incentives to restrain spending were associated with adverse changes, such as care delays or withholding of recommended care. Moreover, prior analyses have suggested that the OCM may have improved some aspects of care (such as adherence to high-priced oral anticancer therapies) in historically marginalized subpopulations.20
The increasing magnitude over time of OCM-associated reductions in total episode payments is a key finding of this analysis. The OCM played out in a context of rapid evolution of cancer treatment and massive disruptions in health care delivery precipitated by the COVID-19 pandemic. Total episode spending during the OCM was roughly 25% higher than in the baseline period, with chemotherapy being the major driver of spending growth in both Part B and Part D. At the same time, Part A spending for acute hospital care declined from the baseline period for both OCM and comparison episodes. Given these changes, it is understandable that oncology practices participating in the OCM would need time to identify and refine strategies and tactics for value-based care delivery transformation. Compared with modest reductions in total episode spending seen in the first 3 years of the OCM,5 the greater magnitude of savings in the last 3 years of the model could indicate that participating practices were able to learn and grow from their experience in the model. This interpretation is supported by the finding that OCM-related savings were clustered in a small number of high-cost, high-volume episode types (high-risk breast cancer, lung cancer, lymphoma, and colorectal cancer), consistent with OCM practices focusing their value-based care initiatives on these episodes.
Because this study used an intention-to-treat analytic framework, the findings described here likely underestimate per-episode savings at the 122 practices that remained in the OCM until its end. A separate OCM evaluation report that used an as-treated analytic framework found that most of the savings in total episode payments was attributable to the 24 practices that ever elected to enroll in 2-sided risk during the course of the OCM.21 These practices, which accounted for 34% of OCM episode volume, were at risk for owing repayments to Medicare if they failed to meet episode spending targets.
Several other investigator teams have also studied the OCM.22,23,24,25,26 Manz and colleagues27 independently confirmed that the OCM was not associated with differences in the use of novel therapies. At least 2 analyses have described apparently favorable spillover effects of the OCM in commercially insured populations.28,29 Evidence of spillover effects beyond the Medicare fee-for-service population may help motivate oncology practices to seek out value-based care initiatives with a broader range of payers.
In July 2023, the CMS launched the Enhancing Oncology Model (EOM) as a successor to the OCM.30 Like the OCM, the EOM is a voluntary, episode-based alternative payment model focused on systemic therapy for cancer. However, the EOM targets a narrower range of common, high-cost cancer types, based on the finding from the OCM that savings were greatest for such high-cost episodes. Additionally, the financial incentives in the EOM include smaller monthly payments and mandatory 2-sided risk for participating practices, reflecting experiences from the OCM as well as the CMS goal for the model to be cost-neutral (or cost-saving). Perhaps related to the relatively modest initial participation (41 practices as of May 2024 and 28 practices as of March 2026),31 the CMS announced changes to the EOM in May 2024 that increase monthly payments and extend the model,31 illustrating the fine balance of running a voluntary participation model that is attractive to payers and oncology practices alike.
Limitations
This evaluation has limitations. First, the OCM was a nonrandomized voluntary model, and practices that chose to participate may differ from those that did not; the study findings may not generalize to all practices. While the difference-in-differences study design with a matched comparison group accounts for constant differences between participating and nonparticipating practices, this approach is not robust to differential changes in unobserved factors associated with outcomes of interest (eg, changes in referral patterns). Second, this analysis did not include detailed clinical information, such as cancer stage and histology, making clinically comprehensive risk adjustment infeasible. Third, many analyses were performed, and adjustments were not made for multiple testing. A threshold of P < .10 was used to define statistical significance to avoid false-negative assessment of program effects (type II error); however, this approach increases the likelihood that some statistically significant results were in fact due to chance.
Conclusions
The OCM was associated with modest payment reductions during 6-month episodes for Medicare beneficiaries receiving chemotherapy for cancer. OCM-attributable payment reductions increased over the 6-year model period, but the costs of model administration exceeded reductions in episode spending. There were no statistically significant differences for any of the reported measures of utilization or quality. Additional research is needed to determine whether the Medicare follow-on Enhancing Oncology Model, launched in 2023, can achieve spending and quality goals.
eAppendix. Analyses
eTable 1. Additional Beneficiary, Practice and Market Characteristics of OCM and Comparison Episodes
eTable 2. Association of OCM With Total Episode Payments, by Episode Type
eFigure. Changes in Spending for Higher- and Lower-Risk Cancer Episodes Before and During OCM, by Category of Spending
Data Sharing Statement
References
- 1.Kline RM, Muldoon LD, Schumacher HK, et al. Design challenges of an episode-based payment model in oncology: the Centers for Medicare & Medicaid Services Oncology Care Model. J Oncol Pract. 2017;13(7):e632-e645. doi: 10.1200/JOP.2016.015834 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.RTI International. OCM Performance-Based Payment Methodology Version 7.1. Centers for Medicare & Medicaid Services . Published July 2021. Accessed January 29, 2026. https://www.cms.gov/priorities/innovation/files/x/ocm-pp3beyond-pymmeth.pdf
- 3.Oncology Care Model Evaluation Team. Oncology Care Model (OCM) Care Transformation Insights. Centers for Medicare & Medicaid Services . Accessed January 29, 2026. https://www.cms.gov/priorities/innovation/data-and-reports/2023/ocm-evaluation-pp1-9-transf-ins
- 4.Oncology Care Model Evaluation Team. Evaluation of the Oncology Care Model: Performance Periods 1-9. Centers for Medicare & Medicaid Services . Published June 2023. Accessed November 21, 2025. https://www.cms.gov/priorities/innovation/data-and-reports/2023/ocm-evaluation-pp1-9
- 5.Keating NL, Jhatakia S, Brooks GA, et al. ; Oncology Care Model Evaluation Team . Association of participation in the Oncology Care Model with Medicare payments, utilization, care delivery, and quality outcomes. JAMA. 2021;326(18):1829-1839. doi: 10.1001/jama.2021.17642 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Oncology Care Model Evaluation Team. Evaluation of the Oncology Care Model: Final Report. Centers for Medicare & Medicaid Services . Published May 2024. Accessed November 21, 2025. https://www.cms.gov/priorities/innovation/data-and-reports/2024/ocm-final-eval-report-2024
- 7.Welch WP, Bindman AB. Town and gown differences among the 100 largest medical groups in the United States. Acad Med. 2016;91(7):1007-1014. doi: 10.1097/ACM.0000000000001240 [DOI] [PubMed] [Google Scholar]
- 8.Keating NL, Huskamp HA, Schrag D, et al. Diffusion of bevacizumab across oncology practices: an observational study. Med Care. 2018;56(1):69-77. doi: 10.1097/MLR.0000000000000840 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Vandenbroucke JP, von Elm E, Altman DG, et al. ; STROBE Initiative . Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007;4(10):e297. doi: 10.1371/journal.pmed.0040297 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Oncology Care Model Evaluation Team. First Annual Report from the Evaluation of the Oncology Care Model: Baseline Period. Centers for Medicare & Medicaid Services . Published February 1, 2018. Accessed November 21, 2025. https://downloads.cms.gov/files/cmmi/ocm-baselinereport.pdf
- 11.Raphael MJ, Biagi JJ, Kong W, Mates M, Booth CM, Mackillop WJ. The relationship between time to initiation of adjuvant chemotherapy and survival in breast cancer: a systematic review and meta-analysis. Breast Cancer Res Treat. 2016;160(1):17-28. doi: 10.1007/s10549-016-3960-3 [DOI] [PubMed] [Google Scholar]
- 12.Argilés G, Tabernero J, Labianca R, et al. ; ESMO Guidelines Committee . Localised colon cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2020;31(10):1291-1305. doi: 10.1016/j.annonc.2020.06.022 [DOI] [PubMed] [Google Scholar]
- 13.Brooks GA, Landrum MB, Kapadia NS, et al. Impact of the Oncology Care Model on use of supportive care medications during cancer treatment. J Clin Oncol. 2022;40(16):1763-1771. doi: 10.1200/JCO.21.02342 [DOI] [PubMed] [Google Scholar]
- 14.National Comprehensive Cancer Network . NCCN Clinical Practice Guidelines in Oncology: Antiemesis. Version 2.2025. Published 2025. Accessed November 21, 2025. https://www.nccn.org/guidelines/guidelines-detail?category=3&id=1415
- 15.Earle CC, Neville BA, Landrum MB, Ayanian JZ, Block SD, Weeks JC. Trends in the aggressiveness of cancer care near the end of life. J Clin Oncol. 2004;22(2):315-321. doi: 10.1200/JCO.2004.08.136 [DOI] [PubMed] [Google Scholar]
- 16.Morden NE, Chang CH, Jacobson JO, et al. End-of-life care for Medicare beneficiaries with cancer is highly intensive overall and varies widely. Health Aff (Millwood). 2012;31(4):786-796. doi: 10.1377/hlthaff.2011.0650 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.National Quality Forum . Geriatrics and Palliative Care, Spring 2022 Cycle: CDP Report. Published January 30, 2023. Accessed January 16, 2026. https://digitalassets.jointcommission.org/api/public/content/02d025736ea44f2f8d24d4774692a603?v=815d8b8a
- 18.Dimick JB, Ryan AM. Methods for evaluating changes in health care policy: the difference-in-differences approach. JAMA. 2014;312(22):2401-2402. doi: 10.1001/jama.2014.16153 [DOI] [PubMed] [Google Scholar]
- 19.Oncology Care Model Evaluation Team. Evaluation of the Oncology Care Model: Performance Periods 1-5. Centers for Medicare & Medicaid Services . Published January 2021. Accessed November 21, 2025. https://www.cms.gov/priorities/innovation/data-and-reports/2021/ocm-evaluation-pp1-5
- 20.Keating NL, Brooks GA, Landrum MB, et al. The Oncology Care Model and adherence to oral cancer drugs: a difference-in-differences analysis. J Natl Cancer Inst. 2022;114(6):871-877. doi: 10.1093/jnci/djac026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Oncology Care Model Evaluation Team. Evaluation of the Oncology Care Model: Impacts by Risk Arrangement. Centers for Medicare & Medicaid Services . Published October 2024. Accessed January 20, 2026. https://www.cms.gov/priorities/innovation/data-and-reports/2024/ocm-tot-report
- 22.Walker B, Wilfong L, Frytak J, Robert N. Practice patterns among oncologists participating in the oncology care model after three years. J Cancer Policy. 2021;29:100294. doi: 10.1016/j.jcpo.2021.100294 [DOI] [PubMed] [Google Scholar]
- 23.Patt DA, Patel AM, Bhardwaj A, et al. Impact of remote symptom monitoring with electronic patient-reported outcomes on hospitalization, survival, and cost in community oncology practice: the Texas Two-Step Study. JCO Clin Cancer Inform. 2023;7:e2300182. doi: 10.1200/CCI.23.00182 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ghosh N, Emond B, Lafeuille MH, Côté-Sergent A, Lefebvre P, Huang Q. Treatment patterns among patients with mantle cell lymphoma and comparison of healthcare resource utilization and costs among relapsed/refractory patients treated with ibrutinib or chemoimmunotherapy: a real-world retrospective study. Clin Ther. 2021;43(8):1285-1299. doi: 10.1016/j.clinthera.2021.06.012 [DOI] [PubMed] [Google Scholar]
- 25.Franklin EF, Nichols HM, Bohannon L. Exploring the experiences of patients in the Oncology Care Model. JCO Oncol Pract. 2020;16(12):e1433-e1440. doi: 10.1200/OP.20.00324 [DOI] [PubMed] [Google Scholar]
- 26.Gajra A, Jeune-Smith Y, Balanean A, et al. Reducing avoidable emergency visits and hospitalizations with patient risk-based prescriptive analytics: a quality improvement project at an Oncology Care Model practice. JCO Oncol Pract. 2023;19(5):e725-e731. doi: 10.1200/OP.22.00307 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Manz CR, Tramontano AC, Uno H, Parikh RB, Bekelman JE, Schrag D. Association of oncologist participation in Medicare’s Oncology Care Model with patient receipt of novel cancer therapies. JAMA Netw Open. 2022;5(9):e2234161. doi: 10.1001/jamanetworkopen.2022.34161 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mullangi S, Ukert B, Devries A, et al. Association of participation in Medicare’s Oncology Care Model with spending, utilization, and quality outcomes among commercially insured and Medicare Advantage members. J Clin Oncol. 2025;43(2):133-142. doi: 10.1200/JCO.24.00502 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Mullangi S, Chen X, Pham T, et al. Association of patient, physician, and practice-level factors with uptake of payer-led oncology clinical pathways. JAMA Netw Open. 2023;6(5):e2312461. doi: 10.1001/jamanetworkopen.2023.12461 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Chong A, Witherspoon E, Honig B, Ela E, Cavanagh H, Strawbridge L. Reflections on the Oncology Care Model and looking ahead to the Enhancing Oncology Model. JCO Oncol Pract. 2022;18(10):685-690. doi: 10.1200/OP.22.00329 [DOI] [PubMed] [Google Scholar]
- 31.Centers for Medicare & Medicaid Services . EOM (Enhancing Oncology Model). Published August 25, 2025. Accessed November 21, 2025. https://www.cms.gov/priorities/innovation/innovation-models/eom
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix. Analyses
eTable 1. Additional Beneficiary, Practice and Market Characteristics of OCM and Comparison Episodes
eTable 2. Association of OCM With Total Episode Payments, by Episode Type
eFigure. Changes in Spending for Higher- and Lower-Risk Cancer Episodes Before and During OCM, by Category of Spending
Data Sharing Statement
