A, Incidence of ESKD divided into quintiles and shown per 1 million state population. B, Ratio of waiting-list incidence to ESKD incidence for each state in 2017 divided into quintiles and shown per 100 state population. C, Ratio of kidney transplant (living and deceased donor) incidence to ESKD incidence for each state in 2017, divided into quintiles and shown per 100 state population.
A, State rankings for percentage of kidney transplant incidence/ESKD incidence and for median waiting times for transplant. Both living donor kidney transplant (LDKT) and deceased donor kidney transplant (DDKT) rates are shown for each state. States are rank ordered in the probability of receiving a kidney transplant as a function of ESKD incidence. The solid black line illustrates the median waiting time of patients with ESKD who received a deceased donor transplant. The dashed blue line represents the (lack of) correlation between median DDTK waiting time and receiving a DDKT; state rates of transplant decrease as waiting times increase. B, Proposed changes from the Organ Procurement and Transplantation Network in kidney allocation volume by geographic location compared with the current allocation model. Donation service areas (DSAs) are renamed for the geographic location of the state. States are rank ordered from the largest decrease to the largest increase in deceased donor kidney allocation.
eFigure 1. Heat Map of the US Showing ESKD Prevalence, Waiting List Incidence/ESKD Prevalence, and Kidney Transplant Incidence/ESKD Prevalence
eFigure 2. Transplant/ESKD Incidence Rate Comparison (Primary vs Sensitivity)
Customize your JAMA Network experience by selecting one or more topics from the list below.
DuBay DA, Morinelli TA, Su Z, et al. Association of High Burden of End-stage Kidney Disease With Decreased Kidney Transplant Rates With the Updated US Kidney Allocation Policy. JAMA Surg. 2021;156(7):639–645. doi:10.1001/jamasurg.2021.1489
Will the introduction of the new kidney allocation policy initiated by the Organ Procurement and Transplantation Network (OPTN) be equally beneficial to all areas of the US when considering the regional burden of end-stage kidney disease (ESKD)?
This economic evaluation of 122 659 patients with ESKD found that this policy change on kidney allocation using transplant rates normalized to the population with ESKD will result in disproportionate distribution of organs across the US.
These findings suggest that states with lowest transplant rates among the population with ESKD will not benefit from these changes by the OPTN, and some may experience a decrease in allocated organs.
The Organ Procurement and Transplantation Network (OPTN) approved changes to the US kidney allocation system in 2019. The potential effects of this policy change using transplant rates normalized to end-stage kidney disease (ESKD) incidence have not been investigated.
To estimate how the OPTN kidney allocation policy will affect areas of the US currently demonstrating low rates of kidney transplant, when accounting for the regional burden of ESKD.
Design, Setting, and Participants
This cross-sectional population-based economic evaluation analyzed access of patients with ESKD to kidney transplant in the US. Participants included patients with incident ESKD, those on the kidney transplant wait list, and those who received a kidney transplant. Data were collected from January 1 to December 31, 2017, and were analyzed in 2019.
Main Outcomes and Measures
The probability of a patient with ESKD being placed on the transplant wait list or receiving a deceased donor kidney transplant. States and donor service areas (DSAs) were compared for gains and losses in rates of transplanted kidneys under the new allocation system. Transplant rates were normalized for ESKD burden.
A total of 122 659 patients had incident ESKD in the US in 2017 (58.2% men; mean [SD] age, 62.8 [15.1] years). The probability of a patient with ESKD receiving a deceased donor kidney transplant varied 3-fold across the US (from 6.36% in West Virginia to 18.68% in the District of Columbia). Modeling of the OPTN demonstrates that DSAs from New York (124%), Georgia (65%), and Illinois (56%) are estimated to experience the largest increases in deceased donor kidney allocation. Other than Georgia, these states have kidney transplant rates per incident ESKD cases above the mean (of 50 states plus the District of Columbia, New York is 16th and Illinois is 24th). In contrast, DSAs from Nevada (−74%), Ohio (−67%), and North Carolina (−61%)—each of which has a transplant rate per incident ESKD cases significantly below the mean—are estimated to experience the largest decreases in deceased donor allocation (of 50 states plus the District of Columbia, North Carolina is 34th, Ohio is 38th, and Nevada is 47th).
Conclusions and Relevance
The new OPTN-approved kidney allocation policy may result in worsening geographic disparities in access to transplants when measured against the burden of ESKD within a particular region of the US.
The Organ Procurement and Transplantation Network (OPTN) approved changes to the kidney allocation system in December 2019 in an effort to improve parity across the US.1Quiz Ref ID The rationale for the allocation changes was that the calculated waiting time for a deceased donor kidney transplant differed by more than 5 years, depending on the geographic location of the transplant center for a patient on the wait list (hereinafter referred to as wait-listed patients).2,3 Making everyone wait in line an equal amount of time (central to the refined OPTN policy) is a generally accepted practice in the US with democratic roots. The OPTN argued that the current allocation system failed to minimize geographic location in prioritization of distribution of available organs.1-4 Implementation of the kidney allocation system in 2014 redefined the prioritization and distribution of available organs, which helped to alleviate racial and medical priority disparities.5,6 Analyses performed after implementation of the kidney allocation system report that geographical location remains the factor most associated with transplant access.7,8
In contrast, others have argued that the kidney allocation system changes will exacerbate existing geographic and racial disparities of access to kidney transplant by patients with ESKD.9 These differences in opinion are primarily related to how one defines access to deceased donor kidney transplants. The OPTN allocation modeling essentially uses transplant rates as a function of cumulative waiting time of transplant registrants. One problem with such an approach is that it does not consider the regional burden of ESKD (only wait-listed patients are considered). Thus, geographic regions with high wait list rates would be allocated more organs (and vice versa) regardless of regional burden of ESKD, the condition treated by kidney transplant.9 In 2018, the United Network for Organ Sharing (UNOS) board of directors published principles of geographic distribution, which included the principle that “deceased donor organs are a national resource.”10(p4) We posit that if deceased donor organs are a national resource, all patients with ESKD should be considered in the allocation policy. Accordingly, we analyzed the impact of this new allocation policy using transplant rates normalized to the population burden of ESKD across the US.
We used a merged data set containing both the United States Renal Disease System (USRDS)11 and Scientific Registry of Transplant Recipients (SRTR) databases for analyses. The USRDS divides the US into networks and gathers robust data on all patients with ESKD, which are publicly available. Similarly, SRTR acquires data on all patients wait-listed for solid organ transplant and all receiving a kidney transplant. The SRTR conducts analyses using these data and reports the proportion of wait-listed patients undergoing transplant. The SRTR also reports median waiting times for all transplant centers, a calculation derived from the center’s entire waiting list—including inactive and highly sensitized patients who are significantly less likely to receive a deceased donor transplant. Approval was obtained from the Medical University of South Carolina institutional review board; informed consent was waived because the study was a retrospective review of medical records and complied with Exempt Research Category 4 requirements. This study met the majority of criteria for the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline.
We examined the incidence of ESKD by state provided by the USRDS and estimated the likelihood of patients with ESKD receiving a kidney transplant based on state of residence. Basic demographic information was summarized by patient count and proportion or mean (SD). Heat maps of ESKD incidence, waiting list rates, and transplant rates by state were created using SAS PROC GMAP, version 9.4 (SAS Institute, Inc). Waiting time was computed for patients who received a deceased donor kidney transplant, calculated either from the first dialysis date or first wait list date, whichever was earlier, to first transplant date. The Pearson correlation coefficient was calculated for the median waiting time and transplant rate to determine whether a linear association existed between these 2 variables. Statistical significance was determined at 2-sided P < .05. SAS, version 9.4, was used for the statistical analysis.
The study population included patients with incident ESKD and incident kidney transplant procedures from January 1 to December 31, 2017 (Table 1). A total of 122 659 new patients with ESKD were included (71 397 men [58.2%] and 51 262 women [41.8%]; mean [SD] age, 62.8 [15.1] years), translating into an incidence of 377 per million US population (Figure 1A). A total of 35 447 patients with ESKD were added to the kidney transplant waiting list and 19 694 total kidney transplants were performed in 2017; of these, 5728 were from living donors, 13 956 from deceased donors, and 10 from unknown donor type. A total of 15 365 wait-listed patients either died or were delisted. Of note, the kidney allocation policy only pertains to deceased kidney donors.
Quiz Ref IDMarked differences were found in ESKD incidence across the US; the highest rates are in the District of Columbia, the Southeast, West Virginia, and New Jersey. In contrast, the lowest incidence of ESKD was in the Mountain West states, Minnesota, and some New England states (Figure 1A). The ESKD prevalence rates are depicted in eFigure 1A in the Supplement.
Quiz Ref IDThe probability of a new patient with ESKD being added to the wait list for kidney transplant also significantly varied across the US. The highest probabilities were noted in Wyoming, Colorado, Minnesota, and several New England states. In contrast, the lowest rates were observed in Hawaii, Oregon, Nevada, Oklahoma, Arkansas, Louisiana, and Ohio River Valley states (Figure 1B). Wait list incidence and ESKD prevalence rates are depicted in eFigure 1B in the Supplement.
There were stark differences in the probability of a new patient with ESKD receiving a living or deceased donor kidney transplant across the US. The highest probability was in Minnesota, other Midwestern states, Mountain West states, Alaska, and several New England states. In contrast, the lowest probability of receiving a kidney transplant was in Nevada, West Virginia, and the Southeast (Figure 1C). Transplant incidence and ESKD prevalence are depicted in eFigure 1C in the Supplement.
State-by-state comparison of the probability of a patient with ESKD receiving a kidney transplant is illustrated in Figure 2A. Quiz Ref IDThe probability of a patient with ESKD in calendar year 2017 receiving a deceased donor kidney transplant varied 3-fold across the US, with the lowest probability being in West Virginia (6.36%) and the highest probability being in the District of Columbia (18.68%). Similarly, the probability of a patient with ESKD receiving a living donor kidney transplant varied more than 10-fold, with the lowest probability being in Arkansas (1.21%) and the highest being in Utah (12.87%). Collectively, the likelihood of a patient with ESKD receiving either a deceased or living donor kidney transplant varied more than 3-fold across the US, with the lowest probability being in Hawaii (9.16%) and the highest being in Utah (31.16%).
A sensitivity analysis was performed to exclude poor transplant candidates (instead of normalizing transplant probabilities to all incident ESKD). Deceased and living donor transplant probabilities in 2017 were calculated from ESKD incidence in 2016 among patients who were 70 years or younger or were older than 70 years but survived 12 months of dialysis by the end of 2017 or death (whichever came first). This approach reduced the sample size (ie, the denominator) from the primary analysis (N = 122 659) to the sensitivity analysis (n = 110 765). eFigure 2A in the Supplement compares the primary and sensitivity analyses. Some nominal changes occurred among the states, most notably New Hampshire, Wyoming, and Colorado, where the transplant proportions increased the most. These states also are in the highest quintile of wait-listing rates (Figure 1B).
Median transplant waiting time for a deceased donor kidney transplant varied from a low of 2.6 years in Nebraska to 7.2 years in Wyoming (Figure 2A). There was no correlation between median waiting time and deceased donor kidney transplant, living donor kidney transplant, and both donor transplants. The Pearson correlation coefficient between median deceased donor transplant waiting time and deceased donor kidney transplant was −0.2164 (P = .13) (Figure 2A).
The modeling method used by the OPTN for deceased donor allocation is based on donor service areas (DSAs), which are the nonprofit agencies that coordinate deceased donation. The OPTN reports12 include maps of US DSAs. In many cases, DSAs are state based, but in some heavily populated areas, there may be more than 1 DSA per state (eg, New York, California, and Florida) and in some sparsely populated areas, there may be more than 1 state in a DSA (eg, Washington State, Montana, and Alaska).
Based on kidney transplant frequency illustrated in Figure 2A, one might expect with the new OPTN allocation system that Hawaii, West Virginia, Arkansas, Mississippi, and Nevada should receive the largest increase in deceased donor kidneys. However, OPTN modeling13 demonstrates that DSAs from New York, Georgia, and Illinois are likely to experience the largest increases in deceased donor allocation. Other than Georgia, these states with increased kidney allocation have transplant rates above the mean (of 50 states plus the District of Columbia, New York is 16th and Illinois is 24th) (Figure 2B). Quiz Ref IDModeling use by the OPTN demonstrates that DSAs from Nevada, Ohio, and North Carolina are likely to experience the largest decreases in deceased donor allocation. These states with decreased kidney allocation have transplant rates below the mean (of 50 states plus the District of Columbia, North Carolina is 34th, Ohio is 38th, and Nevada is 47th) (Figure 2B).
New York City is estimated to experience the largest increase in deceased donor kidneys (124%), whereas the state of Nevada is estimated to experience the largest decrease (−74%). There are 10 kidney transplant centers in the greater New York City area with 6456 wait-listed patients, and 691 patients received deceased donor kidney transplants in 2017; this represents 10.7% of the New York City wait list, with an SRTR-calculated median center waiting time of greater than 72 months14 (Table 2). Conversely, there is 1 kidney transplant center in Nevada with 179 wait-listed patients, and 57 patients received deceased donor kidney transplants in 2017, representing 32.8% of the Nevada waiting list with an SRTR-calculated median waiting time of 11.8 months.14 When measured against the ESKD burden of disease, patients with ESKD in New York City had a 59% higher probability of receiving a deceased donor kidney transplant compared with ESKD patients in Nevada (14.6% vs 9.2%, respectively) (Figure 2A). Furthermore, although not considered in the new deceased donor allocation policy, ESKD patients in New York City had a nearly 3-fold higher probability of receiving a living donor kidney transplant compared with ESKD patients in Nevada (6.4% vs 2.4%, respectively) (Figure 2A).
Kidney allocation is complicated, with multiple stakeholders making arguments why their patient populations with ESKD are most deserving of deceased donor kidney allocation volume increases.15,16 We strongly believe the best method to assess for organ allocation equity and geographic disparities is to estimate the proportion of patients with ESKD who receive a transplant, which varies by more than 3-fold across the US (Figure 2A). Because kidney transplant is the criterion standard treatment for ESKD,17,18 the most appropriate measure to evaluate access to this sparse commodity is to measure it against the burden of the disease it is intended to treat.9 Such an approach is consistent with the UNOS’s vision statement to provide “a lifesaving transplant for everyone in need.”19,20 This is further exemplified by reviewing the OPTN’s vision statement to promote “long, healthy and productive lives for persons with organ failure.”19 Neither statement limits their scope to only wait-listed patients or those who are referred and evaluated for transplant. Regrettably, ESKD burden is completely ignored in the changes approved by the OPTN to the kidney allocation system in late 2019.9 This is an especially important principle for regions of the US with long-standing racial and socioeconomic disparities in transplant waiting list registrations,21,22 the starting point for the approved allocation changes. For example, African American individuals living in USRD Network 6 (Southeast US) are surprisingly referred for kidney transplant at a higher rate than all other racial and ethnic groups but are less likely to initiate an evaluation and even less likely to be placed on a kidney transplant waiting list.23 Our analysis demonstrates that states with the lowest transplant rates normalized for ESKD burden will not benefit from the changes by the OPTN, and several are projected to experience significant decreases in kidney organ allocation volume.
The rationale for this dramatic shift in deceased donor kidneys produced by this new policy is based on SRTR-calculated transplant center waiting times. This measurement is imprecise and significantly overestimates actual waiting time for transplant centers with large waiting lists, including inactive patients and highly sensitized patients who are unlikely to ever find a suitable match.24 In general, centers based in heavily populated urban areas can generate very large waiting lists and correspondingly very long calculated waiting times. The converse is also true: transplant centers located in sparsely populated rural areas often have shorter waiting lists and shorter waiting times. Our analyses report the median waiting times, not of a transplant center waiting list but simply of patients who received a deceased donor transplant. The median waiting time of patients who received a deceased donor kidney transplant was 49 months in New York and 46 months in Nevada (Table 2). In contrast, the SRTR estimation of median waiting time is longer than 72 months in New York vs 11.8 months in Nevada in 2017.25 One can readily see how inflated waiting time estimates distort actual waiting time, and the net effect will be to move deceased donor kidneys to heavily urban areas of the US, thus inducing more disparities in access to care within rural vulnerable populations with ESKD. Estimating waiting times using patients who are inactive on the list is problematic because it can make the system gameable; centers can create long lists and waiting times by simply listing all referrals as inactive. Using waiting time estimates based on those who actually undergo transplant eliminates this issue.
The ideal solution to mitigate disparities in access to transplant is to improve health care infrastructure throughout the US. States with the highest wait list rates also have the highest transplant rates (Figure 1, B and C). Unfortunately, the converse is also true. Of note, the OPTN cannot refer or wait-list patients with ESKD; only local physicians and transplant centers can do so. In addition, the OPTN only allocates and distributes organs to patients on the waiting list. Although federal initiatives, such as the Advancing American Kidney Health Initiative, promote kidney transplant access, these initiatives are not specifically directed to vulnerable populations, and it is not clear that they will affect existing disparities. Until the playing field is level, it is important for the OPTN to not create policies that potentially worsen disparities in access to transplant.
There are several limitations to this analysis. We used registry data, which are known to have inaccuracies and missing data. Some calculations and inferences were based on modified definitions, such as waiting time, which differs from previous analyses. Most important, we used data gathered before the allocation change to infer future transplant rates and outcomes. History has demonstrated that these predictive models can often be inaccurate because they fail to account for the dynamic effects of organ acceptance behaviors by clinicians, which markedly influence transplant rates.
The new OPTN-approved kidney allocation policy may result in worsening geographic disparities in access to transplant when measured against the burden of ESKD within a particular region of the US. Paradoxically, the largely urban areas with much higher transplant rates gain from the new allocation policy, whereas rural areas with low transplant rates, vulnerable patient populations, and a much higher ESKD burden lose access to deceased donor organs. The OPTN allocation policy may further exacerbate the inequity in these regions of the country, where patients with ESKD have a much lower probability of being wait-listed for transplant. With the updated policy, those patients who are placed on the waiting list in rural areas may have a lower probability of getting a deceased donor kidney transplant.
Accepted for Publication: February 13, 2021.
Published Online: May 26, 2021. doi:10.1001/jamasurg.2021.1489
Corresponding Author: Derek A. DuBay, MD, Division of Transplant Surgery, Department of Surgery, Medical University of South Carolina, 96 Jonathan Lucas St, 409 Room C1 Clinical Science Building, Charleston, SC 29425 (email@example.com).
Author Contributions: Dr DuBay was the principal investigator. Dr DuBay had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: DuBay, Mauldin, Weeda, Baliga, Taber.Acquisition, analysis, or interpretation of data: DuBay, Morinelli, Su, Weeda, Casey, Taber.Drafting of the manuscript: DuBay, Morinelli, Su, Taber.Critical revision of the manuscript for important intellectual content: DuBay, Su, Mauldin, Weeda, Casey, Baliga, Taber.Statistical analysis: DuBay, Su.Administrative, technical, or material support: Morinelli, Taber.Supervision: DuBay, Casey, Baliga.
Conflict of Interest Disclosures: Dr Casey reported receiving grants from Dialysis Clinic, Inc, outside the submitted work. No other disclosures were reported.
Disclaimer: The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the US government.
Additional Information: The data reported herein are publicly available from the US Census Bureau (https://www.census.gov/newsroom/press-kits/2018/pop-estimates-national-state.html), the Scientific Registry of Transplant Recipients (https://www.srtr.org/reports/program-specific-reports/), and the United States Renal Data System (https://www.usrds.org).