More than 720,000 adults in the United States were referred for a kidney transplant over the past decade, yet roughly half of them never took the first step toward evaluation. Only about one in ten ultimately received a transplant. Those numbers, drawn from the largest electronic health record analysis of its kind, expose a bottleneck that sits well before the national organ waitlist and that no federal data system currently tracks.
Why the pre-evaluation dropout rate demands attention now
A peer-reviewed cohort study led by Donnelly et al. and published in the Journal of the American Society of Nephrology used Epic Cosmos data covering 720,348 adults referred for kidney transplant between 2014 and 2025. The findings are stark: approximately 48% of those referred initiated evaluation, roughly 19% reached the waitlist, and about 10% were transplanted. The steepest drop occurred at the very first gate, between referral and evaluation start, where more than half of all patients fell away.
That gap matters because the federal transplant infrastructure only begins counting patients once they appear on the waitlist. The national statistics compiled in the OPTN/SRTR annual report track waitlist additions, transplant volumes, and organ donation trends, but they do not capture referral volumes or measure how many patients stall before evaluation. The result is a blind spot: policymakers can see how many people are waiting for kidneys but not how many were lost upstream.
Federal payment models now tie dialysis facility revenue to transplant access rates. The CMS Innovation Center’s ESRD Treatment Choices model and the newer Increasing Organ Transplant Access model both create financial incentives for facilities to move patients toward transplant. Yet if attrition is highest before evaluation even begins, those incentives may be pushing against a wall they cannot see. Facilities that built structured referral protocols before these models launched could, in theory, show lower pre-evaluation dropout in future OPTN-linked data compared with matched controls. That hypothesis remains untested because no national dataset currently links referral records to waitlist outcomes at scale.
What the Donnelly cohort study and regional data reveal
The Donnelly study is the first to quantify the full referral-to-transplant pipeline using a nationwide EHR network. Epic Cosmos aggregates de-identified records from hundreds of health systems, giving researchers a view of clinical activity that sits outside traditional claims-based surveillance. The 720,348 adults in the study represent a population large enough to expose patterns that smaller registries miss. The finding that only 48% initiated evaluation confirms what regional research had suggested for years but had never been measured at this scale.
Earlier work by Patzer et al. in the American Journal of Transplantation documented steep early dropout between referral and first evaluation visit among incident end-stage renal disease patients in the Southeastern United States. That study linked USRDS patient records to referral and evaluation data from multiple transplant centers and found that evaluation initiation was a major barrier well before listing. The Donnelly findings now suggest that the Southeastern pattern is not an outlier but a national norm, with similar early losses likely occurring in other regions.
The federal transplant data system, maintained by HRSA through the Organ Procurement and Transplantation Network, collects candidate waiting list, organ donation, matching, and transplant outcomes information. Public dashboards on the OPTN data portal allow users to explore waitlist counts, transplant rates, and center-level performance. However, the scope of these data begins at the point a patient is formally listed. The United States Renal Data System, a surveillance program run through the National Institute of Diabetes and Digestive and Kidney Diseases in collaboration with CMS and OPTN, tracks chronic kidney disease and ESRD populations. Neither system captures the referral-to-evaluation transition that the Donnelly study identifies as the single largest source of patient loss.
Gaps in tracking and what patients should watch for
Several questions remain open. No national dataset records how many patients are referred for transplant evaluation in a given year, which means the 48% initiation rate from the Donnelly study cannot be independently verified through government sources. Longitudinal follow-up of patients who never initiate evaluation is missing from both the Cosmos dataset and OPTN records, so researchers cannot yet determine whether those patients died, remained on dialysis without further transplant consideration, or simply moved to a different health system where records are not captured.
Facility-level referral rates linked to individual patient outcomes exist only in regional studies like the Patzer analysis. Scaling that approach nationally would require linking EHR referral records to OPTN waitlist data, a step that neither CMS nor HRSA has publicly announced. Without that linkage, the ESRD Treatment Choices model and the Increasing Organ Transplant Access model are measuring downstream outcomes while the upstream problem goes unmonitored.
Direct patient and provider statements about barriers to evaluation are absent from the aggregate data reports. The Donnelly study quantifies the dropout but does not explain it. Transportation, insurance verification delays, incomplete medical workups, and communication failures between dialysis centers and transplant programs are all plausible contributors, but none has been isolated as a primary driver at the national level. Understanding which of these factors dominates would require mixed-methods research that combines quantitative tracking with interviews and surveys.
For patients with advanced chronic kidney disease or those already on dialysis, the lack of visibility into this early bottleneck means they must be especially proactive. Patients should confirm with their nephrologist and dialysis team that a referral has been sent, ask which transplant centers received it, and request written confirmation or portal messages documenting the referral date. Because evaluation often requires multiple tests and visits, patients can also ask for a clear checklist of required steps so that missing lab work or imaging does not silently delay scheduling.
Families and caregivers can play a critical role by helping track appointments, arranging transportation, and ensuring that medical records from other hospitals or clinics are forwarded to the transplant center. When language barriers, low health literacy, or limited internet access are present, social workers and patient navigators may be able to bridge gaps that would otherwise lead to missed evaluation opportunities. In the absence of systematic federal tracking, these individual actions become a frontline defense against falling into the unmeasured half of referred patients who never start evaluation.
Policy implications and next steps
The disconnect between what national systems measure and where patients are lost has clear policy implications. If nearly half of referred candidates never begin evaluation, quality metrics that focus solely on waitlist additions and transplants will underestimate disparities and overestimate system performance. Dialysis facilities and transplant centers may appear to be meeting benchmarks even while large numbers of patients never cross the first threshold.
One option would be to require standardized reporting of transplant referrals and evaluation starts from dialysis facilities and transplant programs, with those data linked to existing OPTN and USRDS records. Even a limited pilot in a subset of states could clarify whether the national patterns seen in Epic Cosmos hold across payer types and regions. Another approach would be to incorporate referral-to-evaluation metrics into future iterations of payment models, creating explicit incentives to reduce early attrition rather than assuming it is unavoidable.
Researchers, meanwhile, can build on the Donnelly findings by examining subgroups within the EHR data to identify which patients are least likely to initiate evaluation. While the current study focuses on overall rates, future analyses could explore patterns by age, race and ethnicity, insurance status, and comorbid conditions. That level of detail would help target interventions, such as transportation support or streamlined pre-evaluation testing, to the patients who need them most.
The Donnelly cohort analysis does not offer a simple fix for the transplant pipeline, but it does clarify where attention should turn. The largest loss of potential kidney recipients occurs not on the waitlist, but in the quiet space between a referral being sent and an evaluation actually beginning. Until national data systems and payment models are reoriented to illuminate and address that gap, hundreds of thousands of patients will continue to disappear from the transplant pathway long before anyone starts counting them.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
