Key Takeaways
- Early KCC evaluations (PY2022–23) show unfavorable net results for Medicare primarily because participant program payments outweighed often modest gross savings or amplify non-statistically significant losses.
- KCC’s PY2026 payment-policy changes (benchmark discount, QCP reduction, and kidney transplant bonus elimination) would materially reduce program payments (roughly 40%–60% by year in retrospective estimates).
- We estimate that if PY2026 program redesign changes were applied retrospectively, the KCC program would have generated net savings of $26 million in PY2023 and $90 million in PY2024.
- If we also retroactively apply global payment parameters to KCEs that in PY2026 are participating in the global track, in addition to retroactively applying the 6% ESRD global discount, net savings to CMS would result in $112 million in PY2023 and $118 in PY2024.
- Literature evidence from major CMS population-management models (e.g., MSSP) consistently shows that savings are rarely immediate and typically increase over time as organizations build capabilities and care processes mature.
- MSSP did not generate net savings until year five, and subsequent program redesign (Pathways to Success) and participant maturity were associated with higher savings rates, lower adverse utilization, and stronger quality.
- Although the latest evaluation does not yet show net savings or reductions in hospitalizations, readmissions, or ED visits, it indicates program progress consistent with objectives and notes it may be too early to see acute-care impacts, which typically represent significant drivers of total costs. To this end, one- and two-year evaluations of the KCC program are likely to understate program effectiveness.
Introduction
The first two evaluations of the Kidney Care Choices Model (KCC) performed by the Lewin Group have shown unfavorable program results in terms of savings to the Medicare Trust Fund. The first evaluation for Performance Year (PY) 2022, which marked the first year of the program, estimated the KCC model led to a non-statistically significant net increase to Medicare spending of $52 million, driven primarily by program payments to participants amounting to $84 million, and $32 million in savings (not statistically significant). The second evaluation report for PY2023 shows a statistically significant increase to net Medicare spending of $305 million, which stems from non-statistically significant gross losses amounting to $116 million and program payments of $189 million.
In this memo, we will address the appropriateness of these results for predicting future financial performance of the KCC program, and potential concerns around the United States Centers for Medicare & Medicaid Services (CMS)’s weighing results from early evaluations too heavily. To this end, our approach consisted of three parts:
- Analysis of financial reconciliation data from the KCC program to assess how PY2026 KCC payment changes result in CMS savings when applied retroactively to past performance years.
- A review of literature on how population management programs, especially those from the Center for Medicare and Medicaid Innovation (CMMI), take time to mature and generate savings.
- An examination of how program savings evolve over time, using the Medicare Shared Savings Program (MSSP) as a case example.
PY2026 Programmatic Changes
We conducted analysis of financial reconciliation data from the KCC program for performance years 2022 to 2024. We used the KCC public use file from PY2022, participant data provided by client for PY2023, and publicly available financial reconciliation data for PY2024. We assessed how PY2026 program changes would have affected KCC payments by retroactively applying them to the historical financial reconciliation data.
We focused on three programmatic changes: (1) the 1% discount applied to both CKD and ESRD benchmarks for all risk tracks, effectively raising the shared savings threshold; (2) 50% reduction in QCP; and (3) elimination of the kidney transplant bonus (KTB). The overall impact of these changes when applied to historical PYs materially reduces the CMS’ KCC program payments as shown in Figure 1.
Figure 1. KCC 2026 Changes Applied to Historical PYs (millions)
Across performance years, PY2026 changes reduced KCC total payments relative to actuals by approximately 60% in PY2022, 45% in PY2023, and 40% in PY2024. This should have the effect of drastically increasing the likelihood of savings over the program trajectory, much like in the 2019 overhaul of the MSSP program.
To put this into perspective, Figure 2 (next page) compares the estimated net costs to CMS when the payment reductions from PY2026 are applied as compared to actuals. In this calculation, we consider all major program payment components: net shared savings, leakage and alignment withhold balance, Health Professional Shortage Area Bonus, home dialysis true-up, high performers pool, nephrology claims, QCP and retrospective recoupments, and estimated total KTB payments over the three-year payment period (based on the number of transplants done during each performance year). The difference in columns represent these program payments less gross savings from Total Cost of Care (TCoC), with positive / negative values reflecting net costs to / savings for CMS, respectively.
As shown below, applying retrospective PY2026 changes improves the likelihood of savings to CMS; from PY2022 to PY2024 actual payments exceeded gross savings, but when PY2026 changes are considered, the resulting differences now show net savings to CMS in PY2023 and PY 2024 of $26 million and $90 million respectively, and slight losses of $8 million in PY 2022. While these figures are not evaluations nor represent controlled research studies, results indicate that PY2026 changes should improve financial results from future evaluations.
Figure 2. KCC 2026 Payment Changes Relative to Gross Savings
Since alignment and leakage balance amounts are based on CMS withholding a percentage of the QCPs, and under PY2026 QCPs are being reduced by 50%, it is likely that this mechanic would also result in alignment and leakage balances being reduced by 50%. In Figure 3, we consider this scenario. The results are directionally the same, though show slightly larger savings to CMS.
Figure 3. KCC 2026 Payment Changes Relative to Gross Savings (leakage and alignment balance payments also reduced by 50%)
Savings are larger if considering KCEs transition to Global and the Global ESRD discount
In addition to the PY2026 programmatic changes mentioned above, we also modelled the shift to global status for KCEs in PY2026 and the 6% discount to the ESRD benchmark for global participants. We used PY2022 to PY2024 participant data and identified KCEs that are participating in the global track for PY2026. For those KCEs, we applied PY2026 global model parameters to their PY2022 to PY2024 performance, including the PY2026 ESRD discount of 6%. Additionally, we also applied the PY2026 ESRD discount to KCEs that were already in the global track in PY2022, PY2023 and PY2024. As expected, applying global track parameters to KCEs’ historical performance data as well as the 6% ESRD benchmark discount decreased participant benchmarks, making it less likely to generate shared savings, which in turn reduced payments from CMS to KCEs (as seen in Figure 4). Had CMS implemented these program design elements to past performance years, it would have saved roughly $100 million in PY2022, close to $190 million in PY2023, and $130 million in PY2024.
Figure 4. KCC 2026 Changes, PY2026 KCE Global Status, and PY2026 ESRD Discount Applied to Historical PYs
In Figure 5 (next page), we compare CMS payments resulting from these changes in addition to all other major program payment components. The difference in columns represent program payments less gross savings from Total Cost of Care (TCoC), with positive / negative values reflecting net costs to / savings for CMS, respectively. As seen in Figure 5, applying KCE global status to past performance years for KCEs that in PY2026 are in the global track as well as the PY2026 ESRD discount to global KCEs in past performance years results in savings to CMS across all years from $33 million in PY 2022 to almost $120 million in PY 2024. In Figure 6 (next page), we show the same results but considering that the alignment and leakage balance amounts would also be reduced by 50% since QCP is reduced by 50%. The results are directionally the same, though show slightly larger savings to CMS.
Figure 5. KCC 2026 Payment Changes with PY2026 KCE Global Status and ESRD Discount Applied Relative to Gross Savings
Figure 6. KCC 2026 Payment Changes with PY2026 KCE Global Status and ESRD Discount Applied Relative to Gross Savings (leakage and alignment balance payments also reduced by 50%)
Review of Literature
Savings take time and grow over time
Our literature review focused on empirical research describing how population management programs evolve over time and generate savings. The largest and most policy-relevant evidence comes from evaluations of Accountable Care Organization (ACO) programs, particularly the Medicare Shared Savings Program (MSSP); however, across all programs and populations, the most consistent finding is that savings are rarely realized immediately and rather grow over time.
In a longitudinal study of ACOs, Bond et al. (2025) analyzed 2010 to 2019 Medicare data covering more than 41 million person years and found that ACOs were associated with larger savings over time, finding mean annual per-patient spending reductions of $234 in year three to $584 in year six, a more than two-fold increase.1 Earlier ACO evaluations showed the same pattern of increasing returns over time. McWilliams et al. (2018) found that physician-group ACOs entering MSSP in 2012 achieved mean annual spending reductions of $474 per patient by 2015, compared to only $156 for 2014 cohort entrants, strongly suggesting that longer participation yields larger savings.2 This cohort-entry effect — in which earlier entrants uniformly outperform later entrants — is consistent with organizational maturation and program preparedness, as we later address in this report.
Further evidence that savings take time is shown in an ACO case study by Madden et al. (2022), in which they describe an initial period of financial losses from 2015–2017 followed by shared savings from 2018–2020, alongside improvements in quality scores and the implementation of strategies such as risk stratification, care coordination, and more precise diagnostic coding. Importantly, these strategies do not just come overnight but take time to build and mature as functioning organization capabilities, which in turn produce savings.3
The evidence demonstrating that savings from population management interventions take time is not just isolated to MSSP historical experience. A study by Maddox et al. (2020) evaluating the Bundled Payments for Care Improvement (BPCI) program, found that early cohort participants, Wave 1 hospitals, had no significant savings at 0–6, 6–12, 12–18, or 18–24 months before achieving significant savings at 24–30 months.4 Additionally, a six-year evaluation of the Comprehensive Primary Care (CPC) program found that hospitalization reductions emerged only in years five and six, with ED visit reductions appearing in years three through six, while the effects were entirely absent the first two years; in contrast, the two-year CPC evaluation showed no significant savings of any kind5 and later participant joiners did not produce any savings.6 To summarize, material reductions in medical events that typically drive healthcare costs are not generally achieved early in the program.
In Figure 7, we illustrate how major CMS programs have evolved over time in terms of savings generation and program performance.
Figure 7. Savings Trajectory Across Major Medicare Care Coordination Evaluations
Organizational capabilities, population risk, and program design influence timing and savings generation
Organizational learning and capability building require time for organizations to develop the necessary infrastructure, workflows, and physician engagement to produce savings. Maddox et al. (2020) illustrate this by finding that Wave 1 hospitals in the BPCI program achieved significant savings only at 24 to 30 months of participation, while Wave 2 and Wave 3 hospitals did not achieve savings regardless of interval.7 Because all waves faced the same payment model and patient populations, the differences can be attributed to maturation of internal competencies rather than program design. Additionally, earlier ACO cohorts have shown consistently greater savings than new entrants.8
Another factor influencing program financial performance is population risk structure. The CMS synthesis of 21 CMMI models found that models targeting specific high-cost populations (i.e., terminal illness, joint replacements) were more likely to show gross savings and had larger, more favorable utilization impacts than primary care and population management models serving broader, healthier populations.9 While the KCC model is built around population health management for beneficiaries with late-stage chronic kidney disease and ESRD, beneficiary disease progression can vary widely, hence making the cost curve and timing of savings more nuanced as opposed to treating other high-risk more acute populations with clearer savings opportunities in the horizon.
Lastly, program financial incentive structures directly affect how program savings are derived and hence the narrative around program success. The CMS synthesis of evaluation results across CMMI programs found that only six of twelve programs achieved net savings, noting that while generous bonus structures offered strong financial incentives for participation, they made it difficult for the models to demonstrate net returns particularly in the short run when reductions in spending are limited.10 The Comprehensive ESRD Care Model illustrates this – the program reduced spending among early participants; however, shared savings payments exceeded gross savings, resulting in a net loss to Medicare of approximately $46 million overall.11
Implications for KCC program evaluations
The KCC model was built upon the experience of the Comprehensive End-Stage Renal Disease Care Model (CEC), which ran from 2015 to 2021. The fifth program evaluation of the CEC model found reductions in hospital admissions and readmissions and lower Medicare expenditures per beneficiary per month.12 As KCC participants continue to further organizational capabilities, its savings trajectory should similarly improve. The latest KCC program evaluation for PY2023 has already shown empirical evidence that participants are implementing changes in line with overall program goals. The evaluation notes:
- An increase in evaluation and management payments of $2 per patient per month (PPPM) relative to the comparison group, pointing out that “providers may be increasing evaluation and management visits to improve their oversight of patient care and avoid higher-cost complications in the future”13; however, this investment takes time to materialize as shown by the literature.
- An increase of home dialysis payments of $16 PPPM driven by peritoneal dialysis (PD); the evaluation notes that increases in Part B payments among KCC patients relative to the comparison group, “corresponded to KCC impacts on dialysis modality [home dialysis] and largely aligned with overall model goals”.14
These findings are important, because while the evaluation does not currently show program savings, it acknowledges expected results in line with program objectives, which in turn should produce savings into the future. Additionally, while there are no significant reductions in hospitalizations, readmissions, and ED visits, which are typical drivers of cost savings, the evaluation notes that it “may be too early in the model to observe effects on acute care use. Changing the culture and processes of a practice in a way that improves outcomes can take time”.15
In short, as the KCC program matures and participants build operational capabilities, we would expect future evidence of savings generation, consistent with the broader literature showing that savings often take several years to materialize. When coupled with program redesign elements implemented in PY2026, future evaluations can reasonably be expected to demonstrate more favorable program findings.
MSSP Program Experience
To further examine program maturation and its effect on savings, we analyzed historical data from the MSSP, which began over a decade ago and represents one of CMMI’s most successful value-based payment initiatives. Figure 8 (next page) illustrates historical net gains or losses for each program year as well as running totals representing cumulative savings. As seen below, MSSP did not generate net savings until its fifth year, 2017, thereafter producing consistent savings netting CMS about $12 billion through 2024. This pattern is consistent with literature in that program savings take time to realize and are rare in the first few years of program maturity.
Figure 8. MSSP Net Program Savings/ Losses Over Time
The MSSP underwent a major redesign in 2019 with the “Pathways to Success” overhaul, which fundamentally restructured the program to push ACOs more quickly into downside risk and helped CMS generate net savings. We examined how ACOs that started in 2019 have improved over time in Figure 9 (next page). In their first performance year, 2019 starters recorded their lowest savings rate at approximately 2.3%, which more than doubled to 5.0% by performance year 2024. During this same period, the percentage of ACOs with positive savings increased from 76% to nearly 90%. This growth illustrates how ACOs that stayed in the program over time gained experience and performed financially better, which in turn helped overall program performance.
Figure 9. MSSP 2019 Starters Savings Performance by Year
This observation regarding ACO performance over time is not just isolated to 2019 starters. In Figure 10, we demonstrate the consistency of this maturation curve across cohorts; specifically, earlier years in the program yielded the lowest savings rates followed by increasing rates in later years across cohorts.
Figure 10. Savings Maturation by Cohort
Since MSSP program rules and methodology are regularly changed as the program evolves, we examined MSSP program results for PY2024 alone by participant program duration in Figure 11 (next page). Veteran ACOs entering the program earlier (2019 or prior) and those with longer participation durations have consistently demonstrated stronger financial and utilization performance than later entrants. The 2019-or-earlier cohort achieved the highest mean savings (5.1%) alongside the lowest ED visit (615.4 annual utilization / 1,000) and SNF admission rates (40.8), indicating more effective cost and utilization management. In contrast, more recent entrants, particularly the 2024 cohort, show the lowest savings (2.7%) and higher utilization levels across IP admissions, ED visits, and SNF use.
Notably, quality scores trend higher among longer-tenured ACOs, suggesting that experience in the program may contribute to both improved efficiency and maintained or enhanced care quality over time. Further, the improvements in savings and utilization metrics discussed above materialize despite risk scores slightly increasing with program duration.
Figure 11. PY2024 Metrics by ACO Program Duration
Early evaluations do not capture extent of program maturity
Participant organizations perform better as the program ages. In the case of the MSSP, the program took five years to produce CMS net savings. MSSP earlier cohorts consistently performed better over later years in the program, with program duration being a key factor in increasing savings, reducing adverse utilization metrics, and improving quality measure performance. To this end, one- and two-year evaluations of the KCC program likely understate program effectiveness. As KCC matures and participants build experience, future evaluations are more likely to capture savings. To this end, KCC entities have already made considerable improvements in quality by increasing optimal ESRD starts by 31%, home dialysis rates by 10%, and decreasing hemodialysis catheter use (considered less safe) by 29%.16
1 Amelia M. Bond, Yasin Civelek, William L. Schpero, et al (2025) Long-Term Spending of Accountable Care Organizations in the Medicare Shared Savings Program. Journal of the American Medical Association (JAMA). https://doi.org/10.1001/jama.2025.3870
2 Michael J. McWilliams, Laura A. Hatfield, Bruce E. Landon, et al (2018) Medicare Spending after 3 Years of the Medicare Shared Savings Program. New England Journal of Medicine. https://doi.org/10.1056/NEJMsa1803388
3 K. Madden, Arnold DoRosario, Ryan T. O’Connell, et al (2022) Community-based, Hospital-Affiliated Medical Group Becomes a Successful Accountable Care Organization. American Journal of Medical Quality. https://doi.org/10.1097/JMQ.0000000000000085
4 Karen E. Joynt Maddox, E. Orav, Jie Zheng, Arnold M. Epstein (2020) Learning and the “Early Joiner” Effect for Medical Conditions in Medicare’ s Bundled Payments for Care Improvement Program. Medical Care. https://doi.org/10.1097/MLR.0000000000001395
5 Dale S, Ghosh A, Peikes D, et al. Two-Year Costs and Quality in the Comprehensive Primary Care Initiative. NEJM. 2016. doi:10.1056/NEJMsa1414953
6 Fu N, Singh P, Dale S, et al. Long-Term Effects of the Comprehensive Primary Care Model on Health Care Spending and Utilization. JGIM. 2021. doi:10.1007/s11606-021-06952-w
7 Karen E. Joynt Maddox, E. Orav, Jie Zheng, Arnold M. Epstein (2020) Learning and the “Early Joiner” Effect for Medical Conditions in Medicare’ s Bundled Payments for Care Improvement Program. Medical Care. https://doi.org/10.1097/MLR.0000000000001395
8 Michael J. McWilliams, Laura A. Hatfield, Bruce E. Landon, et al (2018) Medicare Spending after 3 Years of the Medicare Shared Savings Program. New England Journal of Medicine. https://doi.org/10.1056/NEJMsa1803388
9 (2022) Synthesis of Evaluation Results across 21 Medicare Models, 2012-2020 Synthesis of Evaluation Results across 21 Medicare Models, 2012-2020
10 (2022) Synthesis of Evaluation Results across 21 Medicare Models, 2012-2020 Synthesis of Evaluation Results across 21 Medicare Models, 2012-2020
11 Centers for Medicare & Medicaid Services (CMS). Comprehensive ESRD Care (CEC) Model: Performance Year 2 Annual Evaluation Report. Baltimore, MD: CMS Innovation Center, 2019
12 Boyer, G. J., Duvall, T., & Wells, C. (2022, January 18). The Medicare CEC model: Using lessons learned to improve value-based kidney care. Health Affairs Forefront. https://www.healthaffairs.org/content/forefront/medicare-cec-model-using-lessons-learned-improve-value-based-kidney-care
13 Refer to page 22 of the KCC Model Second Annual Evaluation Report, Performance Year 2023. Centers for Medicare & Medicaid Services. (2026). Kidney Care Choices (KCC) model second annual evaluation report: Performance year 2023. U.S. Department of Health and Human Services. https://www.cms.gov/priorities/innovation/data-and-reports/2026/kcc-2nd-annual-report
14 Refer to page 19 of the KCC Model Second Annual Evaluation Report Summary of Findings. Centers for Medicare & Medicaid Services. (2026). Kidney Care Choices (KCC) model second annual evaluation report: Performance year 2023 (Preview version). U.S. Department of Health and Human Services. https://www.cms.gov/priorities/innovation/data-and-reports/2026/kcc-2nd-annual-report-preview
15 Refer to page 19 of the KCC Model Second Annual Evaluation Report, Performance Year 2023. Centers for Medicare & Medicaid Services. (2026). Kidney Care Choices (KCC) model second annual evaluation report: Performance year 2023. U.S. Department of Health and Human Services. https://www.cms.gov/priorities/innovation/data-and-reports/2026/kcc-2nd-annual-report
16 KCC Model Second Annual Evaluation Report Summary of Findings. Centers for Medicare & Medicaid Services. (2026). Kidney Care Choices (KCC) model second annual evaluation report: Performance year
Share
Accountable for Health's Kidney Care Coalition letter urges CMS to extend the proposed MOD2 modifier to nephrologists in the Comprehensive Kidney Care Contracting Model.
A4H’s Kidney Care Coalition outlines a roadmap to strengthen and sustain specialty-focused kidney care models in Medicare. The paper offers policy recommendations to improve stability, expand participation, strengthen accountability, and ensure continued access to value-based kidney care beyond 2027.
A4H’s Kidney Care Coalition released new resources showing that accountable kidney care is improving outcomes and has stronger savings potential than early evaluations suggest. With the CKCC Model set to end in 2027, CMS should build on this progress by creating a stable, long-term pathway for kidney care providers to continue delivering coordinated, patient-centered kidney care.
