Paying dialysis providers for transplants: the kidney order priced outcomes they only partly controlled
The July 2019 kidney order and CMS’s mandatory ETC proposal tied dialysis payments to home dialysis and transplant rates. Those rates also depend on patient support, transplant centers and organ supply, and the model ended without moving home dialysis use, waitlisting or living donation.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
The July 2019 kidney order and CMS’s mandatory ETC proposal tied dialysis payments to home dialysis and transplant rates. Those rates also depend on patient support, transplant centers and organ supply, and the model ended without moving home dialysis use, waitlisting or living donation.
- The week On Wednesday, July 10, 2019, President Trump signed Executive Order 13879, Advancing American Kidney Health.
- How the payments worked The proposal paired a carrot with a carrot-and-stick.
- The patient’s informed choice The first link is the one ETC’s paid parties could reach directly.
- Support to dialyze at home The second link is support.
- Access to the transplant waiting list and living donors The third link is transplant access.
- The supply of deceased donor organs The fourth link is organ supply.
- Objections One objection holds that ETC’s purpose was to test, not to succeed, and that a mandatory, randomized design gave Medicare a clean answer.
- How it played out by type of organization A dialysis organization with facilities both inside and outside the selected regions faced a portfolio effect: some facilities had adjustments and others did not.
- With hindsight The final rule for the ETC Model was published in the Federal Register on September 29, 2020.
The week
On Wednesday, July 10, 2019, President Trump signed Executive Order 13879, Advancing American Kidney Health. The order opened with the government’s own figures: kidney disease was the ninth-leading cause of death in the United States in 2017; approximately 37 million Americans had chronic kidney disease and more than 726,000 had end-stage renal disease (ESRD); more than 100,000 began dialysis each year; and nearly 100,000 were on the waiting list for a kidney transplant. It called the state of care for these patients unacceptable.
The order set three policies: prevent kidney failure where possible, increase patient choice through affordable alternative treatments for ESRD, and increase access to transplants by modernizing organ recovery and transplantation. It then gave the Secretary of Health and Human Services a list of tasks with deadlines measured in days from the order. The Secretary was to select a payment model for patients with advanced kidney disease who were not yet on dialysis, including flexible advance payments for nephrologists, and a second payment model to create incentives for home dialysis and kidney transplants. The second model was to adjust payments based on the percentage of a provider’s attributed patients on home dialysis or with a transplant, and to include a learning system.
The order also directed a proposed rule to revise the rules and metrics for organ procurement organizations (OPOs), which recover organs from deceased donors in assigned service areas, with more transparent, reliable and enforceable measures of performance. It directed work to speed kidney matching and reduce discards, a proposed rule to remove financial barriers to living donation by covering lost wages and child and elder care, and steps to encourage artificial kidneys through FDA review and the KidneyX partnership with the American Society of Nephrology.
The same day, the Centers for Medicare & Medicaid Services issued a proposed rule for two mandatory payment models, which went on public inspection that day and was published in the Federal Register on July 18. One was a Radiation Oncology Model. The other was the ESRD Treatment Choices (ETC) Model. Under the proposal, ESRD facilities and Managing Clinicians, the clinicians who bill Medicare’s monthly capitation payment for dialysis patients, would be required to participate if they were located in randomly selected Hospital Referral Regions. CMS proposed selecting a random sample of 50 percent of those regions, stratified by Census region, plus every region with at least 20 percent of its zip codes in Maryland.
Participants would receive a positive Home Dialysis Payment Adjustment on home dialysis claims for the first three years. They would also receive a Performance Payment Adjustment, upward or downward, on dialysis and dialysis-related payments, based on their rates of home dialysis and of kidney and kidney-pancreas transplants among attributed patients. Rates would be measured from Medicare claims and administrative data and from the Scientific Registry of Transplant Recipients. The rule cited a 2018 analysis by the US Renal Data System that ranked the United States 27th of the countries and regions compared on the share of dialysis patients dialyzing at home, at 12 percent in 2016. Comments were due on September 16, 2019.
The kidney order and the ETC proposal were a bet on a mechanism: change how Medicare pays the people who manage dialysis, and they will change how many patients dialyze at home or get a transplant. The bet was reasonable on its face. Dialysis facilities and nephrologists see patients for years and influence their choices. CMS itself said it saw ESRD facilities and Managing Clinicians as the key providers managing modality options and helping patients through the transplant process.
Our argument is that the ETC design paid two parties for outcomes produced by a longer chain, and that the parts of the chain outside their control set the limit on what the payments could achieve. Home dialysis and transplant depend on four links, which together cover the path from a failing kidney to a treatment. First, the patient’s informed choice. Second, the support a patient needs to dialyze at home and keep doing so. Third, access to the transplant waiting list and to living donors, which runs through transplant centers. Fourth, the supply of deceased donor organs, which runs through organ procurement organizations and allocation. The paid parties controlled much of the first, some of the second and little of the third and fourth. After a short account of how the payments worked, we take each link in turn, then test the argument against objections, and then look at how it played out for different kinds of organization.
The sector here is healthcare, and the lens is economics, because the question is how a payment design allocates incentive and risk. Organizations should check how any Medicare model or rule applies to them with counsel and their Medicare administrative contractor.
How the payments worked
Two adjustments
The proposal paired a carrot with a carrot-and-stick. The Home Dialysis Payment Adjustment raised payments on home dialysis claims for the first three years, to help participants start or expand home programs. The Performance Payment Adjustment applied to dialysis and dialysis-related payments, home and in-center alike, and could be positive or negative. CMS proposed that the size of the positive and negative adjustments would increase over the life of the model. That design front-loaded support and back-loaded risk.
Scores, benchmarks and the higher of two
Each participant would receive a Modality Performance Score built from its home dialysis rate and its transplant rate. The rates would be risk-adjusted and reliability-adjusted. CMS proposed two benchmarks for each rate: an achievement benchmark built from historical rates in comparison areas, and an improvement benchmark based on the participant’s own history. A participant would get the higher of its achievement score and its improvement score for each rate. CMS said it intended to raise achievement benchmarks over the duration of the model.
The higher-of-two rule is generous by design. It rewards a participant that starts low and improves, and one that starts high and holds steady. It also means a participant can score well on improvement when its rate rises for reasons that have nothing to do with its own effort, such as a national trend toward home dialysis. That feature mattered later, when CMS explained why the model cost more than projected.
Timing and attribution
Performance would be measured over twelve-month measurement years that overlapped by six months. Each measurement year had a six-month payment adjustment period that began six months after it ended. Patients would be attributed to facilities and to Managing Clinicians based on the care they received, and Managing Clinicians would also be credited with pre-emptive transplants, those that happen before a patient starts dialysis. The data would come from Medicare claims, Medicare administrative data and the transplant registry. A participant’s adjustment would therefore arrive at least six months after the measurement year in which the care was given, which weakens any incentive’s link to day-to-day decisions.
The patient’s informed choice
What the clinicians controlled
The first link is the one ETC’s paid parties could reach directly. A patient choosing between in-center hemodialysis, home dialysis and transplant needs to understand the options, ideally before dialysis begins. Nephrologists and dialysis facility staff are positioned to provide that education. If they do it well and early, more patients may choose home dialysis or seek a transplant evaluation. A payment adjustment tied to home dialysis and transplant rates gives them a reason to invest in that education.
The model’s design reinforced the point. CMS’s model page later said dialysis facilities and doctors in the model were encouraged to offer patients education to support their choice of treatment. The CMS Innovation Center’s third annual evaluation found that the model led to an additional 360 patients training for home dialysis per year. That is a measurable effect on the link the paid parties controlled best.
The education benefit
Medicare already had a tool aimed at this link. Since the Medicare Improvements for Patients and Providers Act of 2008, Part B has covered kidney disease education services for beneficiaries with Stage 4 chronic kidney disease, up to six one-hour sessions, as the proposed rule described. CMS proposed, and the final rule adopted, waivers of certain requirements of that benefit so that Managing Clinicians in ETC had more flexibility to furnish and bill for it. The proposed rule gave the reason: “education is a key component of assisting beneficiaries with making such selections.” When CMS ended ETC, it listed a waiver of the same benefit among the ideas being tested in the Kidney Care Choices Model. Both designs treated education as a lever. The benefit is aimed at Stage 4 disease, before dialysis begins, while most of the patients ETC’s rates were calculated on were already on dialysis. The education tool and the payment incentive therefore met the patient at different points.
Data the clinician needs
Education aimed at the right patients at the right time requires data. A nephrology practice needs to know which of its patients have advanced kidney disease, how fast their kidney function is declining, and whether they have been given information on home dialysis and transplant. A dialysis facility needs to know which of its in-center patients might be candidates for home treatment or transplant evaluation. Claims data are unlikely to carry most of that. It comes from clinical records, laboratory trends and documentation of conversations, and they need to be pulled into a registry or worklist that staff actually use.
For a health system that employs nephrologists, building that worklist may be a modest analytics project with a clear purpose. For independent practices, it may require help from an electronic health record vendor or a data partner. Either way, the investment sits with the clinicians, and its payoff under ETC depended on the later links in the chain working too.
Timing of the decision
The second feature of this link is timing. Some patients reach ESRD without long preparation. A patient who starts dialysis urgently, in a hospital, is likely to begin in-center. Changing that path requires earlier identification of patients with advanced chronic kidney disease, which happens before ETC’s Managing Clinicians may be paid for dialysis care. The executive order recognized this by calling for a separate model aimed at patients not yet on dialysis, with advance payments for nephrologists. ETC alone paid on patients already in the dialysis system, after many choices had been made.
For health systems, this is where primary care and hospital medicine come in. Identifying patients with declining kidney function, referring them to nephrology, and preparing them for a planned start belong to the system, not to the dialysis facility. A payment to the dialysis facility for home dialysis rates does not reach the primary care physician who could have referred a patient two years earlier.
Choice has non-clinical reasons
Patients choose for reasons beyond clinical advice. Home dialysis requires space, stability, a willingness to manage treatment and, for some modalities, a care partner. A patient may prefer the routine and social contact of a center. The order spoke of patient choice; the payment rewarded a particular outcome. Those can diverge. A model that rewards home dialysis rates must guard against pressure on patients for whom home dialysis is a poor fit, and the evaluation design must measure whether patients felt informed or pushed. The third evaluation reported that the model preserved patient quality of life and experience of care for patients dialyzing at home and in-center, which suggests that particular risk did not materialize at scale.
Equity in who gets the choice
Access to home dialysis and transplant may not be evenly spread. Patients with fewer resources may lack stable housing, space for supplies, a care partner or reliable transport to transplant evaluations. CMS acknowledged this in the model’s later years. According to the evaluation summary, from July 2023 CMS provided additional support to participants treating patients dually eligible for Medicare and Medicaid or receiving the Part D low-income subsidy. That change recognized that the same payment produces different results depending on the patient’s circumstances, which is another way of saying the outcome depends on more than the paid party.
Support to dialyze at home
Training is a start, not retention
The second link is support. A patient who trains for home dialysis must then succeed at it, month after month. That depends on equipment and supply delivery, technical support, nursing follow-up, management of complications, and the patient’s home and social circumstances. A patient who stops home dialysis and stays on dialysis goes back to a center. A model that counts the share of dialysis time spent at home rewards retention as well as starts.
The evaluation’s numbers point to a gap here. More patients trained, but the evaluation found no impact on overall home dialysis use: home dialysis kept growing in ETC areas and in comparison areas at a similar pace. Some combination of trainees not starting, starting and stopping, or growth in comparison areas closing the gap would produce that pattern. The at-a-glance report listed barriers drawn from interviews: education on treatment options, reliable transportation, mental health support services and communication during the transplant process. Those are mostly support problems, and several sit outside the dialysis facility.
Different home modalities, different support
Home dialysis is not one thing. Peritoneal dialysis and home hemodialysis differ in training, equipment, supply logistics and the role of a care partner. When CMS ended ETC, it pointed to the Kidney Care Choices Model’s focus on peritoneal dialysis as the dominant home modality, among other features being tested there. A program that wants to grow home dialysis needs to know which modality it is growing and build support to match. A single home dialysis rate, as used in ETC, hides that distinction from the payer and from the provider’s own management reports.
Who pays for support
Medicare pays ESRD facilities under the ESRD prospective payment system and pays Managing Clinicians a monthly capitation payment. ETC adjusted those payments. We think the adjustments may have looked small beside the cost of building home programs, such as training staff, home visit capacity and logistics. The Home Dialysis Payment Adjustment, positive for the first three years, helped with start-up. After it ended, the Performance Payment Adjustment could be positive or negative depending on performance against benchmarks.
A dialysis organization deciding whether to invest in a home program would weigh the expected adjustment against the cost and the risk that other factors limit uptake. Where support services such as mental health or transportation depend on community resources, the dialysis organization cannot buy them with the adjustment. The economics of the second link were only partly within ETC’s reach.
Staffing and supply chains
Home programs depend on home training nurses, technical support and reliable delivery of supplies. Each of those is a constraint a dialysis organization must solve before it can serve more patients at home. Staffing shortages can limit training capacity. Supply disruptions can force patients back to a center. A payment adjustment does not create nurses or trucks. It can fund them, if it is large and predictable enough, and if the organization is confident the patients will come. The overlapping measurement years and the lag between care and adjustment made that confidence harder to build.
Measuring retention
ETC’s home dialysis rate counted the share of dialysis treatment time that attributed patients spent at home. That measure captures retention indirectly: a patient who trains and then returns to a center contributes little. For an organization managing its own program, a direct measure is more useful. It should track, for each cohort of patients who start home dialysis, how many are still at home after three, six and twelve months, and why the others stopped. Those reasons separate clinical causes from support failures, and support failures are the ones a program can fix.
Rural patients
The evaluation found greater gains in home dialysis among patients in rural areas, and no other pattern of different effects across the populations studied. That is worth dwelling on. For rural patients, the alternative to home dialysis is long travel to a center several times a week. The incentive and the patient’s own interest point the same way. Where the patient’s alternative is a center nearby, the incentive has less to work with. The rural finding fits the argument that payment works best where the rest of the chain already leans in the same direction.
Access to the transplant waiting list and living donors
A link controlled by transplant centers
The third link is transplant access. A patient becomes a transplant candidate through referral, evaluation and acceptance onto the waiting list by a transplant center. Living donation requires a donor, evaluation of that donor, and a transplant center willing and able to proceed. The dialysis facility and the nephrologist can refer and encourage. They do not decide who is listed, and they do not perform the transplant.
The proposed rule measured transplant rates, which depend heavily on factors beyond the paid parties. The final rule, published in September 2020, changed the measure. It adjusted payments based on rates of home dialysis, transplant waitlisting and living donor transplants. Waitlisting and living donation are earlier in the chain and closer to what a dialysis provider can influence through referral and support. Even so, waitlisting requires a transplant center to complete an evaluation and accept the patient.
Pre-emptive transplant
A transplant before dialysis begins spares that patient dialysis entirely, and the executive order noted that transplantation may eliminate the need for dialysis for many patients. The proposed rule credited Managing Clinicians with pre-emptive transplants for that reason. CMS’s model page later recorded how rare they were: in 2016, only 2.8 percent of incident ESRD patients received a pre-emptive transplant, and 29.6 percent of prevalent ESRD patients had a functioning transplant. A pre-emptive transplant requires referral, evaluation, listing or a living donor, and surgery, all before kidney failure. The nephrologist can start that process. Everything after the referral sits with the transplant center and the donor.
What the evaluation found
Through its first three years, the model had no impact on waitlisting rates or living donor transplant rates, according to the third evaluation. Overall transplant rates grew faster in ETC areas, driven by deceased donor transplants. The model also led to 102 home dialysis patients per year being transplanted. Interviews pointed to communication during the transplant process as a barrier.
The pattern supports the argument. The measures that depend on the transplant center’s decisions, waitlisting and living donation, did not move. The measure that rose, deceased donor transplants, depends on organ supply and allocation, which the next link covers. A dialysis provider’s incentive to get patients listed meets a transplant center’s own capacity, criteria and incentives. ETC put no payment on the transplant center.
Transplant centers have their own constraints
A transplant center decides whom to evaluate and list based on its clinical criteria, its capacity and the oversight it answers to. Evaluation requires appointments, tests and consultations that take staff time. A center with limited capacity may sequence evaluations in ways that delay referred patients. A center may also be cautious about candidates it judges to be at higher risk. None of those decisions were priced by ETC. A dialysis provider pressing for faster evaluation had only persuasion and relationships to work with.
That is why the move from measuring transplants to measuring waitlisting, in the final rule, helped only partly. Waitlisting is earlier in the chain, but it is still the transplant center’s decision. A measure closer to the dialysis provider’s own action would have been the referral itself, or the completion of a referral packet. The tradeoff is that such a measure rewards activity rather than outcome, and a payer may judge that too weak a signal.
Coordination across organizations
For health systems that run both nephrology services and a transplant center, the coordination problem is internal and solvable with shared protocols, data and accountability. For independent dialysis organizations, it is a relationship with one or more transplant centers they do not control. Referral tracking, shared status updates and clear points of contact can reduce delays. Each of those takes work on both sides, and only one side was paid under ETC.
The supply of deceased donor organs
A separate system with separate rules
The fourth link is organ supply. Deceased donor kidneys are recovered by organ procurement organizations in their designated service areas and allocated through the national system. No dialysis provider controls how many kidneys are recovered or how many are discarded. The executive order addressed this link directly, by directing new OPO performance metrics and faster matching to reduce discards, and the order’s living donor directive addressed another part. ETC did not.
That division of labor made sense as policy. Different instruments suit different links. But it meant ETC’s transplant measures depended on progress in a separate program, on a separate timetable. If organ supply rose, ETC participants’ transplant numbers could rise without any change in their own behavior. If it stalled, their efforts on referral could stall too.
Living donors
Living donation is a separate supply, and the order addressed it too. It directed a proposed rule to expand the costs that can be reimbursed to living donors, raise the income limit for eligibility, and allow reimbursement for lost wages and for child and elder care. Those are barriers that sit with the donor, not the recipient’s dialysis provider. ETC’s final design measured living donor transplants among participants’ patients. The evaluation found no impact on living donor transplant rates in the first three years. Finding, evaluating and supporting a donor is work spread across the patient’s family, the transplant center and the donor’s own circumstances.
Matching and discards
The order also told the Secretary to streamline kidney matching and delivery to reduce the discard rate, noting that delayed acceptance by transplant centers harms organ quality. That directive aimed at the logistics between an OPO recovering a kidney and a transplant center accepting it. No ETC participant had a role in those steps. Yet any improvement in them would show up as more transplants among ETC participants’ patients, and any failure would show up as fewer. The model’s transplant measure was, in that respect, partly a measure of a system ETC did not touch.
The OPO metrics
CMS answered the order’s OPO directive with a final rule in December 2020. According to CMS’s later account, that rule replaced heavily risk-adjusted metrics with measures of OPO donation and transplantation volume. It moved from a pass/fail system to three tiers with dynamic thresholds, and used competition among OPOs to drive improvement. After a delay, it took effect on March 30, 2021, and OPOs were assessed over a 2022 to 2026 recertification cycle. For a dialysis provider, those changes were outside its control but inside its results: more organs recovered means more transplants for its patients.
Attribution in payment design
The general lesson for payment design is attribution. When an outcome depends on several parties, paying one of them for it creates two risks. The paid party may be rewarded or penalized for others’ performance. And the incentive may be too weak to change what the paid party can change, because the outcome is diluted by what it cannot. ETC’s benchmarks and improvement scoring were meant to manage this. The termination rule suggests they also allowed participants to earn net increases when home dialysis grew for reasons outside the model.
Payers have two ways out of the attribution problem. One is to pay each party on measures it controls and accept that no one is paid for the whole outcome. The other is to make a group of parties jointly accountable for the outcome, so that the dialysis provider, the nephrologist and, ideally, the transplant center share the reward and the risk. The first is simpler to run; the second aligns better with how the outcome is produced. ETC chose neither in full: it paid two of the parties on outcomes that needed several.
Objections
Mandatory, randomized participation was the point
One objection holds that ETC’s purpose was to test, not to succeed, and that a mandatory, randomized design gave Medicare a clean answer. CMS said it required participation to minimize selection effect, which occurs when only those who would benefit financially choose to join. The objection is strong. ETC produced a credible finding precisely because participation was assigned. Our argument is not that the test was badly designed as a test. It is that the payment lever it tested was aimed at parties who controlled only part of the outcome, and the test confirmed the limit.
Three years is too short
A second objection holds that three years of data, through December 2023, was too short for transplant measures, which move slowly. That may be right for waitlisting and living donation, where evaluation and donor work take time. But CMS judged the evidence enough to end the model, citing both the lack of effect on key measures and the net increase in spending. Section 1115A of the Social Security Act requires the Secretary to terminate or modify a model unless it is expected to improve quality without increasing spending, reduce spending without reducing quality, or both. CMS concluded that ETC had raised Medicare spending without moving home dialysis use, waitlisting or living donation.
Home dialysis grew anyway
A third objection holds that home dialysis grew during these years, so the policy direction succeeded even if ETC did not cause the growth. Home dialysis did keep growing, in ETC areas and comparison areas alike, according to the evaluation. CMS noted that stronger than expected increases in home dialysis, caused by factors other than the model, contributed to participants performing better than projected. For the patients involved, growth is good whatever its cause. For a payer, paying more for growth that would have happened anyway is a cost without a return, and CMS named that growth, together with the improvement scoring method, as a reason for the $99 million net increase.
Voluntary models did better
A fourth objection would point to the voluntary Kidney Care Choices Model, which CMS reported improved home dialysis, training, optimal starts and pre-emptive and living donor transplants. Perhaps the lesson is that voluntary, accountable care arrangements work where mandatory payment adjustments do not. The quality results support that reading. The cost results complicate it: CMS reported a net loss to Medicare of about $304 million in the model’s second performance year. Voluntary participants may also be those most able to succeed, which is the selection effect ETC was designed to avoid. KCC’s broader accountability, across the total cost of care for some options, fits our argument that payment works better when it spans more of the chain, but its cost shows that spanning the chain is expensive.
Where the argument is weakest
Our argument is weakest on the first link. If education and choice had been the main constraint, ETC should have moved home dialysis, and it raised training without raising overall use. That suggests the support link, or patient preference, mattered more than education alone. We cannot say from the evaluation which factor dominated. We would only say the evidence fits a chain whose weakest links lay outside the paid parties, and does not fit a story in which a payment adjustment to dialysis providers was enough. A test that paid transplant centers for waitlisting, or paid a joint entity for the whole path, would have told us more about whether the missing links were the binding ones. ETC was not built to answer that question, and nothing we cite answers it directly.
How it played out by type of organization
Large dialysis organizations
A dialysis organization with facilities both inside and outside the selected regions faced a portfolio effect: some facilities had adjustments and others did not. They could spread investments in home programs and education across the whole network, which may have contributed to growth in comparison areas. That spillover is one reason a randomized test can understate an effect, and one reason the evaluation’s comparison of selected and non-selected regions needs care in interpretation.
Smaller and independent facilities
Independent and smaller dialysis facilities in selected regions faced the same adjustments with fewer resources to build home programs, analytics or transplant relationships. For them, the Home Dialysis Payment Adjustment in the early years mattered relatively more, and the later downside risk more so. A facility without an established home program had to decide whether to start one, partner with another provider or accept the adjustment. Each choice carried costs the model did not cover, and each depended on links in the chain beyond the facility.
Nephrology practices
Nephrology practices billing the monthly capitation payment were Managing Clinicians under ETC. For them, the adjustment applied to a professional payment and depended on patients attributed to them. Practices that wanted to influence outcomes needed data on their attributed patients, referral pathways to transplant centers and partnerships with home dialysis programs. Nephrology practices later had a voluntary option as well. The Kidney Care Choices Model, launched in 2022, brought nephrologists and others together to take accountability for patients with late-stage chronic kidney disease and ESRD.
Health systems with transplant programs
Health systems that run transplant programs sit at the third link. ETC did not pay them for waitlisting, but their decisions shaped its results. For a system that also employs nephrologists or runs dialysis units, the model created an internal reason to align referral and evaluation. For a system with an independent dialysis community around it, the useful step was to make referral and evaluation status visible to referring nephrologists and facilities, since delays and poor communication were among the barriers the evaluation reported.
Organ procurement organizations
OPOs were outside ETC but inside its outcomes. The 2020 rule on OPO performance, prompted by the same executive order, put them under their own measures and recertification process. Their results flowed into ETC participants’ transplant numbers. For a policy maker, the pairing illustrates a choice: pay each link for what it controls, or pay one link for the whole chain. The order did the first across its several actions. ETC, on its own, did something closer to the second.
Health IT and analytics teams
For technology leaders in health systems and dialysis organizations, ETC created a data problem. Attribution, rates and adjustments were calculated by CMS from claims and registry data, months after the fact. Participants who wanted to manage performance needed their own estimates: which patients were likely attributed to them, what their home dialysis and transplant rates looked like in near real time, and which patients were waiting at which step. That required joining clinical, scheduling and referral data, and sometimes data from transplant centers outside the organization. The analytics built for ETC remain useful for any kidney value-based arrangement, which is the better reason to build them.
Payers and accountable care organizations
Payers and accountable care organizations managing populations with kidney disease face the same attribution question in their own contracts. A value-based arrangement with a dialysis provider that pays on transplant rates inherits ETC’s problem. Arrangements that pay each party on measures it controls, such as referral within a set time, completion of education, home dialysis retention, and timely waitlisting decisions, may give clearer signals. We think the evidence from ETC favors those narrower measures, combined with shared accountability across the chain, over a single outcome measure placed on one party.
With hindsight
The final rule for the ETC Model was published in the Federal Register on September 29, 2020. It began on January 1, 2021 rather than in 2020, and was scheduled to run until June 30, 2027. The final rule changed the measures to home dialysis, transplant waitlisting and living donor transplants, and cut the model’s scale from a random 50 percent of Hospital Referral Regions to 30 percent: CMS later described it as mandatory for approximately 30 percent of ESRD facilities and Managing Clinicians. The Radiation Oncology Model proposed in the same rule was also finalized at a reduced scale.
The CMS Innovation Center’s third annual evaluation covered 2021 to 2023. It counted 95 selected Hospital Referral Regions, 31 percent of the national total, with 2,591 participating facilities, 6,379 Managing Clinicians and 141,890 Medicare fee-for-service beneficiaries, 34 percent of those with ESRD. It found more home dialysis training but no impact on overall home dialysis use, no impact on waitlisting or living donor transplants, faster growth in deceased donor transplants, no increase in infections, complications or mortality, preserved patient quality of life and experience of care, and a net increase in Medicare payments of $99 million.
On March 12, 2025, CMS announced its intention to end the model early. Its calendar year 2026 ESRD payment rule, published on November 24, 2025, terminated ETC as of December 31, 2025. CMS cited its statutory duty to terminate or modify models that do not meet the quality and spending tests, the lack of effect on the key measures, and the net spending increase. It said stronger than expected growth in home dialysis from other causes, combined with the improvement scoring method, had let participants earn more than projected. CMS estimated the termination would save about $1 million over the 18 months, January 2026 through June 2027, that the model would otherwise have run.
The voluntary Kidney Care Choices Model, launched in 2022 and extended through 2027, showed a different pattern. CMS reported that its second performance year brought increases in home dialysis, home dialysis training, optimal starts and preemptive and living donor transplants, together with a net loss to Medicare of about $304 million in 2023. The OPO rule that the order prompted stood. On November 3, 2026, CMS finalized procedural revisions for OPOs, effective January 4, 2027, while keeping the outcome measures and tier structure from 2020.
CMS also signaled where it saw promise. In its calendar year 2025 ESRD payment rule it asked for ideas on increasing home dialysis. It said many of the suggestions it received, such as a waiver for the kidney disease education benefit, home dialysis quality measures focused on retention and optimal starts, and an emphasis on peritoneal dialysis, were already being tested in the Kidney Care Choices Model.
When it proposed ETC, CMS said a mandatory design would test payment incentives across a diverse group of providers. When it ended ETC, it judged that test against the statute’s quality and spending criteria, using evidence from the model’s first three years.
The record after 2019 is consistent with our argument. The measure closest to what dialysis providers controlled, home dialysis training, moved under ETC. Overall home dialysis use, waitlisting and living donation did not, and the one transplant gain came through deceased donor transplants, which depend on organ supply.
Spotted an error? Tell us and we will check it and correct the article where needed.