Abstract
Rationale & Objective:
The national kidney allocation system (KAS) implemented in December 2014 redefined the start of waiting time from the time of waitlisting to the time of kidney failure. Waitlisting has declined post-KAS but it is unknown if this is due to transplant center practices or changes in dialysis facility referral and evaluation. The purpose of this study was to assess the impact of the 2014 KAS policy change on referral and evaluation for transplantation among a population of incident and prevalent patients with kidney failure.
Study Design:
Cohort study.
Setting & Participants:
37,676 incident (2012–2016) patients in Georgia, North Carolina, and South Carolina identified within the United States Renal Data System at nine transplant centers and followed through December 2017. A prevalent population of 6,079 patients from the same centers receiving maintenance dialysis in 2012 but not referred for transplantation in 2012.
Exposure:
KAS era (pre-KAS vs. post-KAS).
Outcomes:
Referral for transplantation, start of transplant evaluation, and waitlisting
Analytical Approach:
Multivariable time-dependent Cox models for the incident and prevalent population.
Results:
Among incident patients, KAS was associated with increased referrals (aHR: 1.16; 95% CI: 1.12–1.20) and evaluation starts among those referred (aHR: 1.16; 95% CI: 1.10–1.21), decreased overall waitlisting (aHR: 0.70; 95% CI: 0.65–0.76), and lower rates of active waitlisting among those evaluated compared to the pre-KAS era (aHR: 0.81; 95% CI: 0.74–0.90). Among the prevalent population, KAS was associated with increases in overall waitlisting (aHR: 1.74; 95% CI: 1.15–2.63) and active waitlisting among those evaluated (aHR: 2.01; 95% CI: 1.16–3.49), but had no significant impact on referral or evaluation starts among those referred.
Limitations:
Limited to three states, residual confounding.
Conclusions:
In the Southeastern US, the impact of KAS on steps to transplantation were different among incident and prevalent patients with kidney failure. Dialysis facilities referred more incident patients and transplant centers evaluated more incident patients following implementation of KAS but fewer evaluated patients were placed onto the waitlist. Changes in dialysis facility and transplant center behaviors following KAS implementation may have influenced the observed changes in access to transplantation.
Keywords: kidney transplantation, kidney allocation policy
Introduction
The public health burden of kidney failure in the United States (US) is substantial, with >500,000 US patients with kidney failure on dialysis.1 Kidney transplantation is the preferred treatment for kidney failure, offering longer survival, better quality of life, and fewer hospitalizations than dialysis.2,3 Despite national policy directives aimed at maximizing transplant access and incentivizing waitlisting for dialysis facilities, only ~18.5% of patients with kidney failure are waitlisted,4 and significant variability in access to the waitlist and subsequent transplant exists.5–7
The United Network for Organ Sharing Organ Procurement and Transplantation Network (OPTN) implemented a new policy8 in December 2014 that substantially impacted the way kidneys are allocated. One major purpose of the new kidney allocation system (KAS) was to improve equity in access to transplantation, particularly among sensitized patients and racial/ethnic minorities.9 The most significant change was how allocation time, the primary determinant of priority for organ allocation, was calculated. In KAS, the start of the allocation time changed from the date of waitlisting to the date of dialysis initiation, except for preemptively waitlisted patients, to offset the disadvantages resulting from delayed referral and evaluation that disproportionately affected racial minorities.
While data suggest that waitlisting has substantially declined following the implementation of KAS -- the adjusted rate of waitlisting declined 9% among incident patients with kidney failure and there were nearly 150 fewer waitlisting events per month among prevalent dialysis patients10 -- it is unknown whether the decline in waitlisting was due to a decline in referrals for transplant, or changes in practice by transplant centers because no national data exist to examine those early steps in the transplant process. In addition, it is expected that KAS may have led to reduced urgency to waitlist incident (new) dialysis patients compared to those who have been on dialysis longer (i.e. prevalent patients) because of how KAS reassigned allocation time to include time since kidney failure, but no prior research has examined the impact of KAS separately among both the incident and prevalent dialysis populations. The kidney allocation system continues to evolve, with the recent removal of donation service areas and OPTN region from allocation and replacement with a measure of distance between the donor hospital and the transplant hospital for each candidate. Understanding the impact of allocation policy changes on provider behavior and access to transplantation is essential and should inform any proposed subsequent changes in the allocation such as continuous distribution.
The purpose of this study was to assess the impact of the 2014 KAS policy change on referral and evaluation for transplantation among a population of both incident and prevalent patients. Because national surveillance data on early transplant steps are unavailable, we sought to describe referral and evaluation patterns using a unique dataset encompassing Georgia (GA), North Carolina (NC), and South Carolina (SC). Furthermore, these findings may have important implications for current efforts by the Centers for Medicare and Medicaid Services to identify new dialysis practitioner-level quality measures to increase access to transplantation for ESKD patients.
Methods
Study Population and Data Sources
All patient data were from the United States Renal Data System (USRDS) database (Figure 1), and then linked with the novel Early Transplant Access Registry database. Data on characteristics of patients’ 5-digit ZIP code were obtained from the 2014 American Community Survey and linked by patient ZIP code at start of dialysis to USRDS. As in our prior work,10 we examined the impact of KAS separately among an incident (new) dialysis population and a prevalent (existing) dialysis population because of how KAS assigns more allocation points to patients the longer they are on dialysis.
Figure 1.
Inclusion/Exclusion criteria of the incident study population of patients with kidney failure, 2012–2016, with follow-up through 2017
To construct the incident patient cohort, all new patients who initiated maintenance dialysis between January 1, 2012 and December 31, 2016 in GA, NC, or SC dialysis facilities were included and followed through December of 2017. Patients who were preemptively transplanted (n=4,167), <18 or >80 years (n=3,553), or preemptively waitlisted (n=1,672) were excluded. Referral data derived from the Southeastern Kidney Transplant Coalition Early Transplant Access Registry, a surveillance registry from all nine adult kidney transplant centers in the tristate area,11 were merged with USRDS. The final study cohort consisted of 37,676 patients after excluding those referred prior to kidney failure (n=4,132) (Figure 1).
The prevalent patient cohort included the same inclusion/exclusion criteria as the incident population, but was limited to only patients who initiated dialysis in 2012 (the earliest reporting of referral, evaluation, and waitlisting in the registry) but who were not referred for transplant in 2012. The total number of patients in the prevalent cohort was 6,079. For patients referred more than once, we captured only the first referral.
Study Variables
The primary outcome variable was referral for kidney transplantation from a tristate dialysis facility to a tristate transplant center, which was defined using the gold standard definition: the date a transplant center received a referral form for a patient (including referrals from all providers and self-referrals). Secondary outcomes included start of the transplant evaluation (at any of the 9 transplant centers), measured as the date the patient first attended either a medical or educational appointment at the transplant center, overall waitlisting (includes both active, inactive, and missing active listing events), and active waitlisting (defined as UNOS listing status 1) among those that started the transplant evaluation. We examined events using the denominators of those that had achieved the prior transplant step, rather than all patients, in order to provide insight into the potential mechanisms of these outcomes. For example, examining waitlisting rates among those who started the evaluation may suggest factors related to transplant center processes such as listing practices, rather than dialysis facility referrals.
The primary exposure variable was KAS era. For both incident and prevalent cohorts, a time-varying variable was used in Cox models, where an indicator variable was used to denote post 12/4/2014 intervention date vs. time prior to 12/4/2014.
Other variables examined included age, sex, race/ethnicity, cause of kidney disease, dialysis modality, body mass index (BMI), chronic obstructive pulmonary disease (COPD), tobacco usage, diabetes, cancer, hypertension, insurance coverage (i.e., Medicare, Medicaid, employer, uninsured, other), whether the patient received nephrology care prior to kidney failure, and cardiovascular conditions. All variables are captured at time of dialysis start as reported on the Centers for Medicare and Medicaid Services (CMS) 2728 form, which is only completed at baseline. Neighborhood level characteristics included poor households (living below the federal poverty threshold).
Statistical Analysis
Standardized differences were used to compare demographic and clinical variables for pre-KAS and post-KAS groups. We defined patients in the post-KAS era as those who were impacted by KAS, i.e. patients who either started dialysis prior to KAS and were still on dialysis (and not yet referred) on the implementation date of the new KAS policy, or patients who started dialysis post KAS.
In the incident cohort, we created Kaplan-Meier plots using the Simon and Makuch method12 and performed Mantel-Byar tests to examine the impact of KAS13. The association of KAS on each outcome was also examined using bivariable and multivariable Cox regression analyses, adjusting for demographic and clinical variables. In these KM plots, Mantel-Byar tests, and time-dependent Cox regression models, KAS status was considered a time-dependent variable. In analyses examining transplant evaluation start among those referred and waitlisting among those started evaluation, we modeled transplant center as a random effect since the variance contributed by the center accounted for less than 10% of the total variance14. Patients were censored at death (n=10,147; 27%) or end of study follow-up (12/31/17). We also included death as a competing risk in a sensitivity analysis. The results for the KAS effect from those models are similar to the ones reported. We did not report competing risk models as the main analysis due to concerns of estimate bias from competing risk analysis with time-dependent covariates.15
For the prevalent cohort, bivariable and multivariable Cox regression analyses were also performed, and dialysis vintage was included in Cox regression models of all outcomes except for referral. Dialysis vintage was defined as time since dialysis initiation to start of follow-up for each outcome.
We also created a histogram to show the dialysis facility-level variation of change in the proportion of patients referred from post-KAS to pre-KAS among incident patients in the 790 dialysis facilities in ESRD Network 6. The proportion of patients referred in the pre-KAS period was calculated as the number of patients referred before KAS divided by the number of patients who started dialysis prior to KAS implementation; the proportion of patients referred in the post-KAS period was calculated as the number of patients referred post-KAS implementation divided by patients who were still on dialysis and not yet referred or who started dialysis after Dec. 4, 2014.
Sensitivity Analyses
We also examined the proportion of patients preemptively referred (referred prior to dialysis start) among all referred patients and waitlisted within a year of starting dialysis, and examined active and inactive listing status, both pre- and post-KAS implementation. For these crude analyses, we considered incident patients in 2012 and 2013 with follow-up through 2014 as pre-KAS, and incident patients in 2015 and 2016 with follow-up through 2017 as post-KAS patients.
Finally, because the KAS policy change was approved by the OPTN in June of 2013, we examined the impact of both the announcement of KAS and the date of KAS implementation in sensitivity analyses for both the incident and prevalent populations.
SAS version 9.4 was used for data management and survival analysis,; KM plots and Mantel-Byar tests were conducted using R 4.1.216 with packages survminer, ggplot2, Rcmdr and RcmdrPlugin.EZR17–20. The Emory IRB approved this study (IRB#81303) and waived written, informed consent.
Results
Association of KAS with Transplant Access among the Incident Population
We compared demographic, clinical, and socioeconomic characteristics between incident patients before and after KAS (Table 1). Post-KAS patients were slightly more likely to be non-Hispanic white, diabetic, informed of kidney transplantation, have a BMI >35 kg/m2, and have a cause of kidney failure due to diabetes vs. pre-KAS patients, but the differences were minimal between pre-KAS and post-KAS patients, as all standardized differences were less than 0.1 (Table 1).
Table 1.
Demographic and clinical characteristics of incident patients with kidney failure in Georgia, North Carolina, and South Carolina before (1/1/2012–12/03/2014) and after (12/04/2014–12/31/2016) the Kidney Allocation System (KAS) policy change.
| Characteristics at the time of kidney failure | Study Population of incident patients N=37,676 | Pre-KAS N=21660 (58%) | Post-KAS N=16016 (43%) | Standardized Difference between Pre -and Post-KAS |
|---|---|---|---|---|
| Age group, N (%) | 0.0763 | |||
| 18–39 | 3239 (8.60) | 1927 (8.9) | 1312 (8.19) | |
| 40–49 | 4820 (12.79) | 2827 (13.05) | 1993 (12.45) | |
| 50–59 | 8447 (22.42) | 4888 (22.57) | 3559 (22.22) | |
| 60–69 | 11425 (30.32) | 6577 (30.36) | 4848 (30.27) | |
| >=70 | 9745 (25.87) | 5441 (25.12) | 4304 (26.87) | |
| Male, N (%) | 20904 (55.48) | 12005 (55.42) | 8899 (55.56) | 0.0028 |
| Race/ethnicity | 0.0623 | |||
| White, non-Hispanic | 15248 (40.52) | 8544 (39.47) | 6704 (41.94) | |
| Black, non-Hispanic | 20611 (54.77) | 12094 (55.87) | 8517 (53.28) | |
| Hispanic | 1018 (2.70) | 593 (2.74) | 425 (2.66) | |
| Other | 756 (2.01) | 417 (1.92) | 339 (2.12) | |
| Assigned cause of kidney failure, N (%) | 0.0573 | |||
| Diabetes | 16949 (45.78) | 9646 (45.22) | 7303 (46.54) | |
| Hypertension | 13400 (36.19) | 7737 (36.27) | 5663 (36.09) | |
| Glomerulonephritis | 2432 (6.57) | 1456 (6.82) | 976 (6.22) | |
| Other | 4243 (11.46) | 2493 (11.69) | 1750 (11.15) | |
| Type of Dialysis | −0.0212 | |||
| Hemodialysis | 34388 (91.29) | 19713 (91.04) | 14675 (91.63) | |
| Peritoneal Dialysis | 3281 (8.71) | 1941 (8.96) | 1340 (8.37) | |
| BMI ≥ 35 kg/m2, N (%) | 9347 (25.42) | 5292 (25) | 4055 (26) | 0.023 |
| History of comorbid conditions | ||||
| Congestive heart failure, N (%) | 10352 (27.48) | 5899 (27.23) | 4453 (27.8) | 0.0127 |
| Atherosclerotic heart disease, N (%) | 3640 (9.66) | 2269 (10.48) | 1371 (8.56) | −0.0653 |
| Other cardiac disease, N (%) | 6525 (17.32) | 3751 (17.32) | 2774 (17.32) | 0.0001 |
| Chronic obstructive pulmonary disease, N (%) | 3426 (9.09) | 1884 (8.7) | 1542 (9.63) | 0.0322 |
| Peripheral vascular disease, N (%) | 3300 (8.76) | 1958 (9.04) | 1342 (8.38) | −0.0234 |
| Smoker, N (%) | 3432 (9.11) | 2013 (9.29) | 1419 (8.86) | −0.0151 |
| Diabetes, N (%) | 22241 (59.03) | 12683 (58.55) | 9558 (59.68) | 0.0228 |
| Cancer, N (%) | 2278 (6.05) | 1322 (6.1) | 956 (5.97) | −0.0056 |
| Hypertension, N (%) | 33081 (87.80) | 19046 (87.93) | 14035 (87.63) | −0.0092 |
| Geographic Region | 0.0226 | |||
| Georgia | 16223 (43.06) | 9256 (42.73) | 6967 (43.5) | |
| North Carolina | 13316 (35.34) | 7734 (35.71) | 5582 (34.85) | |
| South Carolina | 8137 (21.60) | 4670 (21.56) | 3467 (21.65) | |
| Patient informed of kidney transplant option, N (%) | 32517 (86.31) | 18441 (85.14) | 14076 (87.89) | 0.0805 |
| Socioeconomic Status Indicators | ||||
| Insurance, N (%) | 0.0735 | |||
| Medicaid | 9022 (24.36) | 5382 (25.23) | 3640 (23.18) | |
| Medicare | 13519 (36.50) | 7636 (35.79) | 5883 (37.47) | |
| Employer | 4821 (13.02) | 2782 (13.04) | 2039 (12.99) | |
| Other | 5847 (15.79) | 3191 (14.96) | 2656 (16.92) | |
| No insurance | 3826 (10.33) | 2343 (10.98) | 1483 (9.44) | |
| Nephrology care prior to kidney failure, N (%) | 23377 (62.05) | 13442 (62.06) | 9935 (62.03) | −0.0006 |
| Patient Neighborhood (ZIP Code) poverty category, N (%) | 0.0268 | |||
| 0–4% | 1547 (4.16) | 881 (4.12) | 666 (4.22) | |
| 5%-9% | 5712 (15.38) | 3273 (15.33) | 2439 (15.44) | |
| 10%-14% | 9106 (24.52) | 5212 (24.41) | 3894 (24.66) | |
| 15%-19% | 8765 (23.60) | 5031 (23.56) | 3734 (23.65) | |
| >=20% | 12014 (32.34) | 6956 (32.58) | 5058 (32.03) |
Abbreviations: KAS, kidney allocation system; BMI, body mass index
- Race/ethnicity: 0.1%
- Assigned cause of kidney failure: 2%
- BMI: 2%
- Insurance: 2%
- Patient Neighborhood (ZIP Code) poverty category: 1%
A total of 43.4% of patients were referred, 52.4% of those referred started the evaluation, and 35.2% of evaluated patients were waitlisted during the study period (Table 2). There were minimal sociodemographic and clinical characteristics differences noted in the proportions of patients who had each outcome pre- vs. post-KAS. For example, among referred patients, a slightly higher proportion were > age 70 years in the post-KAS vs. pre-KAS period (11.46% vs. 10.25%), had a BMI ≥ 35 kg/m2 (27.04% vs. 25.74%), and had comorbidities (e.g., CHF, 23.00% vs. 21.08%); waitlisting declined or remained the same in these subgroups after KAS. In addition, Black patients represented 55.8% of all ESKD patients pre-KAS (vs. 53.3% post-KAS) and Black vs. white patients represented a larger proportion of those referred but differences pre- and post-KAS were minimal (62.4% vs. 61.9%, respectively). Finally, among patients who were waitlisted after completing their evaluation, Black patients represented 56.8% of waitlisted patients pre-KAS but 62.7% post-KAS.
Table 2.
Demographic and clinical characteristics of incident patients referred, evaluated, and waitlisted in Georgia, North Carolina, and South Carolina pre- and post-KAS, 2012–2016, with follow-up through 2017.
| Referred (among those who started dialysis) | Started Evaluation (among those referred) | Waitlisting (among those who started evaluation) | |||||||
|---|---|---|---|---|---|---|---|---|---|
| N=16352/37676 (43.4%) | N=8572/16352 (52.4%) | N=3014/8572 (35.2%) | |||||||
| Pre-KAS | Post-KAS | Standardized Difference | Pre-KAS | Post-KAS | Standardized Difference | Pre-KAS | Post-KAS | Standardized Difference | |
| N = 7201 Referred | N = 9151 Referred | N = 3198 Evaluated | N = 5374 Evaluated | N = 992 Waitlisted | N = 2022 Waitlisted | ||||
| Age group, N (%) | 0.0696 | 0.0685 | 0.0972 | ||||||
| 18–39 | 1095 (15.21) | 1271 (13.89) | 580 (18.14) | 873 (16.24) | 243 (24.5) | 421 (20.82) | |||
| 40–49 | 1443 (20.04) | 1747 (19.09) | 662 (20.7) | 1100 (20.47) | 228 (22.98) | 498 (24.63) | |||
| 50–59 | 1924 (26.72) | 2477 (27.07) | 872 (27.27) | 1485 (27.63) | 246 (24.8) | 555 (27.45) | |||
| 60–69 | 2001 (27.79) | 2607 (28.49) | 841 (26.3) | 1482 (27.58) | 231 (23.29) | 468 (23.15) | |||
| >=70 | 738 (10.25) | 1049 (11.46) | 243 (7.6) | 434 (8.08) | 44 (4.44) | 80 (3.96) | |||
| Sex, N (%) | −0.0195 | −0.0202 | 0.0254 | ||||||
| Male, N (%) | 4279 (59.42) | 5350 (58.46) | 1933 (60.44) | 3195 (59.45) | 600 (60.48) | 1248 (61.72) | |||
| Female, N (%) | 2922 (40.58) | 3801 (41.54) | 1265 (39.56) | 2179 (40.55) | 392 (39.52) | 774 (38.28) | |||
| Race/ethnicity | 0.0000 | 0.0657 | 0.1533 | ||||||
| White, non-Hispanic | 2314 (32.14) | 2944 (32.19) | 1021 (31.93) | 1679 (31.27) | 347 (34.98) | 568 (28.09) | |||
| Black, non-Hispanic | 4496 (62.44) | 5670 (61.99) | 1973 (61.69) | 3326 (61.94) | 564 (56.85) | 1268 (62.71) | |||
| Hispanic | 240 (3.33) | 314 (3.43) | 130 (4.07) | 216 (4.02) | 53 (5.34) | 115 (5.69) | |||
| Other | 150 (2.08) | 218 (2.38) | 74 (2.31) | 149 (2.77) | 28 (2.82) | 71 (3.51) | |||
| Assigned cause of kidney failure, N (%) | 0.0485 | 0.0466 | 0.1125 | ||||||
| Diabetes | 3087 (43.25) | 4097 (45.37) | 1302 (41.12) | 2303 (43.38) | 344 (35.03) | 701 (35.14) | |||
| Hypertension | 2724 (38.17) | 3358 (37.19) | 1209 (38.19) | 1956 (36.84) | 362 (36.86) | 810 (40.6) | |||
| Glomerulonephritis | 664 (9.3) | 732 (8.11) | 329 (10.39) | 526 (9.91) | 146 (14.87) | 279 (13.98) | |||
| Other | 662 (9.28) | 843 (9.33) | 326 (10.3) | 524 (9.87) | 130 (13.24) | 205 (10.28) | |||
| Type of Dialysis | −0.0227 | −0.0326 | −0.0458 | ||||||
| Hemodialysis | 6304 (87.56) | 8079 (88.30) | 2715 (84.9) | 4624 (86.04) | 788 (79.44) | 1643 (81.26) | |||
| Peritoneal Dialysis | 896 (12.44) | 1071 (11.70) | 483 (15.1) | 750 (13.96) | 204 (20.56) | 379 (18.74) | |||
| BMI ≥ 35 kg/m2, N (%) | 1826 (25.74) | 2431 (27.04) | 0.0297 | 770 (24.45) | 1279 (24.17) | −0.0065 | 190 (19.45) | 389 (19.6) | 0.0038 |
| History of comorbid conditions | |||||||||
| Congestive heart failure, N (%) | 1518 (21.08) | 2105 (23.00) | 0.0464 | 584 (18.26) | 1070 (19.91) | 0.0420 | 132 (13.31) | 265 (13.11) | −0.0059 |
| Atherosclerotic heart disease, N (%) | 539 (7.49) | 585 (6.39) | −0.043 | 209 (6.54) | 296 (5.51) | −0.0432 | 48 (4.84) | 78 (3.86) | −0.0481 |
| Other cardiac disease, N (%) | 894 (12.41) | 1201 (13.12) | 0.0213 | 356 (11.13) | 618 (11.5) | 0.0116 | 90 (9.07) | 191 (9.45) | 0.0129 |
| Chronic obstructive pulmonary disease, N (%) | 336 (4.67) | 528 (5.77) | 0.0497 | 92 (2.88) | 241 (4.48) | 0.0855 | 12 (1.21) | 24 (1.19) | −0.0021 |
| Peripheral vascular disease, N (%) | 443 (6.15) | 544 (5.94) | −0.0087 | 153 (4.78) | 257 (4.78) | −0.0001 | 34 (3.43) | 66 (3.26) | −0.0091 |
| Smoker, N (%) | 646 (8.97) | 768 (8.39) | −0.0205 | 223 (6.97) | 420 (7.82) | 0.0322 | 45 (4.54) | 78 (3.86) | −0.0339 |
| Diabetes, N (%) | 4013 (55.73) | 5291 (57.82) | 0.0422 | 1721 (53.81) | 2999 (55.81) | 0.0400 | 469 (47.28) | 947 (46.83) | −0.0089 |
| Cancer, N (%) | 227 (3.15) | 328 (3.58) | 0.0239 | 90 (2.81) | 185 (3.44) | 0.0361 | 25 (2.52) | 50 (2.47) | −0.0030 |
| Hypertension, N (%) | 6460 (89.71) | 8181 (89.4) | −0.0101 | 2848 (89.06) | 4822 (89.73) | 0.0218 | 880 (88.71) | 1798 (88.92) | 0.0067 |
| Geographic Region | 0.0273 | 0.2306 | 0.1740 | ||||||
| Georgia | 3197 (44.4) | 4001 (43.72) | 1763 (55.13) | 2573 (47.88) | 555 (55.95) | 1099 (54.35) | |||
| North Carolina | 2406 (33.41) | 3143 (34.35) | 1081 (33.8) | 1759 (32.73) | 319 (32.16) | 567 (28.04) | |||
| South Carolina | 1598 (22.19) | 2007 (21.93) | 354 (11.07) | 1042 (19.39) | 118 (11.9) | 356 (17.61) | |||
| Patient informed of kidney transplant option, N (%) | 6366 (88.4) | 8286 (90.55) | 0.0699 | 2850 (89.12) | 4889 (90.98) | 0.0621 | 885 (32.24) | 1860 (67.76) | 0.0952 |
| Socioeconomic Status Indicators | |||||||||
| Insurance, N (%) | 0.1087 | 0.1303 | 0.0875 | ||||||
| Medicaid | 1676 (23.49) | 2080 (23.04) | 658 (20.78) | 1133 (21.34) | 157 (15.99) | 316 (15.86) | |||
| Medicare | 2060 (28.87) | 2702 (29.93) | 851 (26.87) | 1434 (27.01) | 231 (23.52) | 428 (21.48) | |||
| Employer | 1402 (19.65) | 1644 (18.21) | 761 (24.03) | 1133 (21.34) | 309 (31.47) | 612 (30.71) | |||
| Other | 860 (12.05) | 1321 (14.63) | 358 (11.3) | 774 (14.58) | 116 (11.81) | 273 (13.7) | |||
| No insurance | 1138 (15.95) | 1282 (14.2) | 539 (17.02) | 835 (15.73) | 169 (17.21) | 364 (18.26) | |||
| Nephrology care prior to kidney failure, N (%) | 4516 (62.71) | 5741 (62.74) | 0.0005 | 2013 (62.95) | 3379 (62.88) | −0.0014 | 625 (33.49) | 1241 (66.51) | −0.0336 |
| Patient Neighborhood (ZIP Code) poverty category, N (%) | 0.0000 | 0.0729 | 0.0761 | ||||||
| 0–4% | 313 (4.4) | 349 (3.87) | 160 (5.07) | 220 (4.15) | 58 (5.92) | 109 (5.47) | |||
| 5%-9% | 1097 (15.44) | 1350 (14.95) | 541 (17.14) | 843 (15.9) | 191 (19.51) | 355 (17.8) | |||
| 10%-14% | 1688 (23.75) | 2168 (24.01) | 707 (22.4) | 1278 (24.11) | 230 (23.49) | 480 (24.07) | |||
| 15%-19% | 1685 (23.71) | 2173 (24.07) | 738 (23.38) | 1286 (24.26) | 233 (23.8) | 475 (23.82) | |||
| >=20% | 2324 (32.7) | 2989 (33.10) | 1010 (32) | 1674 (31.58) | 267 (27.27) | 575 (28.84) | |||
KAS is a time-dependent variable, and therefore pre-KAS and post-KAS denominators may not represent unique patients. The pre-KAS group includes patients who had outcome events or were censored before KAS implementation on Dec. 4, 2014, and the post-KAS group includes (1) patients whose start of follow-up was after KAS implementation or (2) patients who were followed before KAS implementation but had outcome events or censored after KAS. Pre-KAS was defined as patients had outcomes (referral, evaluation or waitlisting) before KAS implementation, which indicated that patients were not influenced by KAS at all.Post-KAS was defined as patients who had outcomes after KAS implementation.
Inactive listings among patients who had started their evaluation (and had known active/inactive waitlist status data) declined following KAS implementation (38.1% pre-KAS vs. 26.4% post-KAS). In adjusted analyses, KAS had a positive association with referral (aHR: 1.16; 95% CI: 1.12–1.20) and evaluation (aHR: 1.16; 95% CI: 1.10–1.21), but a negative association with overall waitlisting (aHR: 0.70; 95% CI: 0.65–0.76). Evaluated patients had lower active waitlisting (aHR:0.81; 95% CI:0.74–0.90) post- vs. pre-KAS (Figure 2; full model Table S1). Among those evaluated, post-KAS was associated with lower inactive listing (HR: 0.49; 95% CI: 0.42–0.57) vs. pre-KAS (results not shown).
Figure 2.
Adjusted Hazard Ratios (95% Confidence Intervals) of the Impact of the new Kidney Allocation System (KAS) policy compared to pre-KAS population on Referral, Evaluation, and Waitlisting among Incident Patients (circle) and Prevalent Patients (triangle) with kidney failure 2012–2016, with at least 12 months on dialysis (2013–2016) with follow-up through 2017 in Georgia, North Carolina, and South Carolina.
Outcomes were defined as referral among all patients with ESKD; evaluation start among all patients referred, overall waitlisting among all patients who started the evaluation, and active waitlisting among all patients who started the evaluation. Models are adjusted for age, sex, race/ethnicity, cause of kidney failure, dialysis modality, BMI, comorbid conditions, informed of transplant options, insurance, pre-ESKD nephrology care, state, and neighborhood poverty.; Dialysis vintage is also included in Cox time-dependent models for evaluation start, overall waitlisting and active waitlisting.
The cumulative incidence plots for incident patients were consistent with Cox modeling results and showed that KAS was associated with a significant increase in referral (Figure S1 A), and evaluation start (Figure S1 B), but was associated with a decrease in overall waitlisting (Figure S1 C) and active waitlisting (Figure S1 D).
Association of KAS on Transplant Access among the Prevalent Population
In the prevalent population, KAS was associated with a non-significant increase in referral (aHR: 1.18; 95% CI: 0.86–1.61), no impact on evaluation start (aHR: 0.99; 95% CI: 0.75–1.30), higher waitlisting (aHR: 1.74; 95% CI: 1.15–2.63) and higher active waitlisting (aHR: 2.01; 95% CI: 1.16–3.49) (Figure 2).
Among 790 dialysis facilities, the mean increase in referral post-KAS was 6.93% (median: 2.12%) (Figure 3). A total of 422 facilities (53.4%) increased their referrals post-KAS (median increase:14.16%). A total of 314 facilities (39.7%) decreased referral post-KAS (median:−10.10%), and 54 facilities (6.8%) stayed the same. The median time from referral to waitlisting increased from pre- to post-KAS implementation (4.6 months pre-KAS vs. 6.0 months post-KAS; p<0.001). Additionally, the proportion of patients who were preemptively referred increased from 18.1% pre-KAS to 19.6% following KAS (p<0.001). Waitlisted patients who were preemptively waitlisted represented 9.7% of the waitlisted population (of which 21.1% were actively waitlisted) pre-KAS and increased to 18.8% (of which 34.2% were actively waitlisted) post-KAS (p<0.001). In sensitivity analyses examining the date of KAS approval, rather than date of implementation, overall results were similar to main results reported for incident and prevalent populations (results not shown).
Figure 3.
Dialysis facility-level variation in change in the proportion of patients referred from post-KAS to pre-KAS among incident patients starting dialysis 2012–2016, with follow-up through 2017, in 790 dialysis facilities in ESRD Network 6. Pre-KAS (Dec. 4, 2014), the proportion of patients referred was calculated as the number of patients referred before KAS divided by the number of patients who started dialysis prior to KAS implementation; Post-KAS, the proportion of patients referred was calculated as the number of patients referred post-KAS implementation divided by patients who were still on dialysis or who started dialysis after Dec. 4, 2014.
Discussion
Nationally, waitlisting has declined following implementation of the 2014 change in the KAS.10,21 Our study sought to examine whether this decrease in waitlisting was driven by lower rates of referral, evaluation, or waitlisting in the Southeastern US, the only region with systematic data collection on early transplant steps that would facilitate this analysis. Results from this region suggest that among the incident population that started the evaluation process, KAS led to a decline in overall waitlisting. While many have hypothesized that these declines in waitlisting may have been due to the result of the removal of the loss of allocation time that results from delayed referral in KAS, our results suggest that the lower rates of waitlisting among the incident population following KAS are not explained by lower rates of referral or evaluation among incident patients, but lower waitlisting among evaluated patients at transplant centers. Among the prevalent population, the group most likely to benefit from the allocation time changes in KAS, we observed higher rates of waitlisting among those who were referred and started the evaluation, but no significant impact of KAS on referral and evaluation start. These results emphasize the importance of understanding how policy changes can lead to health system and provider-level behavior changes among both dialysis facilities and transplant centers.
There are several potential explanations for these findings. First, among the incident population, transplant center listing practices have changed following KAS due to lower sense of urgency to waitlist newly diagnosed patients who, on average, will wait longer to receive a transplant. Prior research has examined how transplant center behavior is influenced by performance oversight, such as increases in waitlist removal and declines in transplant rates among centers that have had low performance evaluations.22 However, little is known about these listing behaviors because national data on referral, transplant evaluation, and transplant candidacy decisions are lacking. Our results suggest that while transplant centers may evaluate referred patients at a higher rate following KAS, rates of transplant center waitlisting have decreased. We do not have detailed data to attribute the decline in incident waitlisting to differences in medical eligibility, lower rates of evaluation completion, or less urgency to waitlist patients. However, our data suggest that increased referrals of older and obese patients with lower waitlisting rates for these patients may reflect center reluctance to waitlist individuals with risk factors, or that centers are concerned patients would not be healthy enough to be transplanted by the time they accrue enough waiting time. Because waitlisting patients as inactive while awaiting completion of their medical evaluation is no longer incentivized under KAS, patients with incomplete work ups may not be listed anymore. While it is not possible to identify within the data what proportion are still undergoing evaluation workup, KAS was associated with greater declines in inactive (vs. active) waitlisting. Future research should explore why incident patients are not being listed inactively as frequently post-KAS, and whether this is due to medical eligibility or longer expected waiting times for patients and the attenuation of the advantage of early listing for patients already on dialysis. Our results complement recent findings that transplant center behavior influences access to transplantation. Schold et al found that from 2001–2015, mortality among waitlisted candidates declined, but with it, a doubling of waitlist removals from the waiting list.23 Transplant center selection criteria and concerns about regulatory oversight and organ availability all may influence these transplant center practices.
It is possible that referrals and evaluations have increased because dialysis facilities are referring more patients with a higher burden of disease, as evidenced by higher rates of referral and evaluation start but lower rates of waitlisting among incident patients evaluated. We found some support for this hypothesis in our study, where we saw small increases in the proportion of patients referred who were older than 70 years of age, had BMI ≥ 35 kg/m2, and had more comorbidities in the pre- vs. post-KAS era, although these are not absolute contraindications and this hypothesis warrants further exploration. Since 2012, the CMS Statement of Work for ESRD Networks has emphasized quality improvement to increase dialysis facility referrals.24 Facilities selected to participate in Network Quality Improvement Program (QIP) activities face increased pressure to refer patients. In a study of 124 dialysis facilities in GA that participated in a QIP in 2014, facilities in the QIP intervention had increased rates of both referral and waitlisting.25 For that QIP, results suggest that more eligible patients were referred and evaluated due to the higher waitlisting rates observed. However, it is unknown if these QIP activities influenced referral patterns and may explain lower waitlisting rates among the incident population in this study. The relatively large number of seemingly transplant eligible (i.e. no obvious contraindications) incident patients makes this somewhat unlikely.
One concern about the national decline in waitlisting is whether the decreased urgency to waitlist patients may also adversely impact transplant center education about living donor transplant -- which has been on the decline in the United States -- or eliminate the possibility of early offers26 where the survival benefit associated with reduced dialysis time would offset the potential disadvantage of a higher KDPI kidney.4 Our results suggest that, on average, transplant centers in the Southeast region continue to evaluate incident patients but waitlist them at a lower rate – a pattern that likely represents a national trend. This suggests that referred patients are likely still receiving education about living donor transplantation, although future research should examine how KAS has impacted living donor transplantation.
Among the prevalent population most likely to benefit from the KAS waiting time change, KAS resulted in no change in dialysis facility referral and evaluation start rates, but higher rates of waitlisting among those who started the evaluation compared to pre-KAS. These results imply that dialysis facilities may be under-referring prevalent patients who could potentially benefit the most from the change in KAS by delaying the referral of those with limited dialysis vintage. A 2018 survey of 653 U.S. dialysis facilities with low waitlisting found that only 57.9% of dialysis facility staff were aware of the KAS change.27 Interventions such as the Allocation Changes for Equity in kidNey Transplantation (ASCENT) that consist of educational materials and a benchmarking feedback report about how KAS impacts a particular dialysis facility may help inform dialysis facility leadership and staff about the importance of referring eligible patients already on dialysis.28 While interim results suggest provider knowledge of KAS impact increased following the intervention,29 more research is needed to determine whether increased referrals and evaluations lead to increased rates of waitlisting and transplant.
Given the limited impact of KAS on referral and evaluation among the prevalent population, and the variability we observed in referral across dialysis facilities, especially among the 39.7% of facilities that decreased referrals post-KAS, there are more opportunities for improvement. Other interventions at the dialysis facility level could offer information to dialysis staff about KAS to emphasize the importance of transplant access for prevalent patients who are prioritized in KAS.30Additionally, quality measures may also be a way to incentivize appropriate behaviors, but current quality metrics are not well aligned between dialysis facilities and transplant centers. In the Spring of 2021, a CMS Technical Expert Panel identified practitioner-level measures to improve access to kidney transplantation among ESKD patients.31 While dialysis facility metrics may help, it will be important to ensure that measures are aligned with transplant center metrics which tend to drive an inappropriate emphasis on greater patient selectivity in an attempt to improve post-transplant outcomes.32 For example, recent proposals that focus on waitlist mortality rates as a potential quality metric may inadvertently disincentivize waitlisting for patients until they have accrued enough allocation priority to be transplanted soon after waitlisting. Increased attention to kidney transplantation from CMS in the form of new value-based kidney care payment models are likely to have an additional impact as well, making it imperative that we have the data to improve our understanding and ability to monitor the multi-step pre-transplant process. It is also unclear if there is enough system capacity to adequately consider the increased referrals at transplant centers – nor can we understand how the current pretransplant reimbursement models may be contributing to these processes
This study has several weaknesses. First, results may not be generalizable outside the Southeast, where obesity, diabetes, and hypertension are more prevalent and transplant rates are among the lowest in the nation33. Second, results do not capture patient referral and evaluation events that occur outside of GA, NC, and SC, and in our study sample, ~11% of patients were waitlisted outside of this region. Our estimates may underrepresent referrals and evaluations; however, for this study we assumed that if a patient was waitlisted, they were successfully referred and evaluated and were counted as such in analyses. In addition, our surveillance data are limited by the data lag in USRDS and only available through December 2017; it is possible that dialysis facility and transplant center program clinician behavior has changed, and that these results may not capture the most updated patterns with respect to transplant referral, evaluation, and waitlisting. Additionally, our analyses do not include referrals and evaluations for Chronic Kidney Disease (CKD) patients who had not yet started dialysis, and thus our results are only generalizable to dialysis patients. In sensitivity analyses, we saw a slight increase in preemptive referrals and an increase in preemptive listing consistent with national data,34 which suggests that referral patterns among CKD patients have not been negatively influenced by KAS, but further research is needed to confirm these findings. Our prevalent population analyses are limited to a small subset of the population for whom we had outcome data on, since we did not have information on referral and evaluation prior to 2012. Moreover, ESRD Networks implement QIPs to a select group of dialysis facilities annually, which could include interventions and resources to improve access to kidney transplant, but we do not have data on this participation and therefore cannot account for the co-occurring impact these QIPs may have on patterns of referral, evaluation, and waitlisting. Finally, this study was observational, and unmeasured confounding could influence effect estimates of the association between kidney failure and transplant access endpoints examined in our study. However, a major strength of this study is the examination of transplant referral and evaluation start, process measures that are unavailable in national data, in addition to waitlisting.
Overall, the effect of KAS differed by transplant step and by incident vs. prevalent dialysis patients, and overall, declines in waitlisting observed in the post-KAS era are largely due to decreased transplant center waitlisting of referred patients. These findings suggest that the change in KAS policy likely influenced provider behavior both at dialysis facilities (with respect to transplant referrals) as well as transplant programs (e.g., evaluation and waitlisting practices). This study offers context for why overall waitlisting rates have declined nationally since the implementation of the 2014 kidney allocation system change and underscores the need to collect surveillance data on these important pre-waitlisting process measures nationally, particularly as new changes in kidney allocation and new quality measures in dialysis facilities and transplant programs are under development.
Supplementary Material
Table S1. Crude and Adjusted Hazard Ratios of the Impact of the new Kidney Allocation System (KAS) policy compared to pre-KAS population on Referral, Evaluation, and Waitlisting among Incident patients with kidney failure 2012–2016 with follow-up through 2017 in Georgia, North Carolina, and South Carolina.
Figure S1. Cumulative incidence plots with Mantel Byar test for referral (panel a), evaluation start (panel b), overall waitlisting (panel c), and active waitlisting (panel d) for incident patients, 2012–2016 with follow-up through 12/31/2017.
Support:
This work was supported by the National Institute on Minority Health and Health Disparities grants R01MD010290 and U01MD010611. The funder did not have a role in study design, data collection, analysis, reporting, or the decision to submit for publication.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Financial Disclosure: SM is a national faculty chair for the ESRD treatment choices learning collaborative (ETCLC). The other authors declare that they have no relevant financial interests.
Publisher's Disclaimer: Disclaimer: The data reported here have been supplied in part by the USRDS. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as official policy or interpretation of the US government.
Data Sharing:
De-identified data can be shared with researchers with appropriate human subjects approvals and data use agreements.
References
- 1.Johansen KL, Chertow GM, Foley RN, et al. US Renal Data System 2020 Annual Data Report: Epidemiology of Kidney Disease in the United States. Am J Kidney Dis. 2021;77(4 Suppl 1):A7–A8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Danovitch WGGaGM. Options for Patients with Kidney Failure. In: Danovitch GM, ed. Handbook of Kidney Transplantation. 2nd ed. Philadelphia: Lippincott Williams & Wilkins; 2005:1–22. [Google Scholar]
- 3.Tonelli M, Wiebe N, Knoll G, et al. Systematic review: kidney transplantation compared with dialysis in clinically relevant outcomes. Am J Transplant. 2011;11(10):2093–2109. [DOI] [PubMed] [Google Scholar]
- 4.Hart A, Smith JM, Skeans MA, et al. OPTN/SRTR 2017 Annual Data Report: Kidney. Am J Transplant. 2019;19 Suppl 2:19–123. [DOI] [PubMed] [Google Scholar]
- 5.Patzer RE, McClellan WM. Influence of race, ethnicity and socioeconomic status on kidney disease. Nat Rev Nephrol. 2012;8(9):533–541. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Schold JD, Mohan S, Huml A, et al. Failure to Advance Access to Kidney Transplantation over Two Decades in the United States. J Am Soc Nephrol. 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.King KL, Husain SA, Schold JD, et al. Major Variation across Local Transplant Centers in Probability of Kidney Transplant for Wait-Listed Patients. J Am Soc Nephrol. 2020;31(12):2900–2911. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Network OPaT. Organ Procurement and Transplantation Network (OPTN) Policies In: Resourcs UDoHaH, ed. Richmond, VA2014. [Google Scholar]
- 9.Glasgow RE, Vogt TM, Boles SM. Evaluating the public health impact of health promotion interventions: the RE-AIM framework. American journal of public health. 1999;89(9):1322–1327. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zhang X, Melanson TA, Plantinga LC, et al. Racial/ethnic disparities in waitlisting for deceased donor kidney transplantation 1 year after implementation of the new national kidney allocation system. Am J Transplant. 2018;18(8):1936–1946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Patzer RE, McPherson L, Wang Z, et al. Dialysis facility referral and start of evaluation for kidney transplantation among patients treated with dialysis in the Southeastern United States. Am J Transplant. 2020;20(8):2113–2125. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Schultz LR, Peterson EL, Breslau N. Graphing survival curve estimates for time-dependent covariates. Int J Methods Psychiatr Res. 2002;11(2):68–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Cantor AB. A test of the association of a time-dependent state variable to survival. Comput Methods Programs Biomed. 1995;46(2):101–105. [DOI] [PubMed] [Google Scholar]
- 14.McPherson L, Basu M, Gander J, et al. Decisional conflict between treatment options among end-stage renal disease patients evaluated for kidney transplantation. Clin Transplant. 2017;31(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Poguntke I, Schumacher M, Beyersmann J, Wolkewitz M. Simulation shows undesirable results for competing risks analysis with time-dependent covariates for clinical outcomes. BMC Medical Research Methodology. 2018;18(1):79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.R: A Language and Environment for Statistical Computing [computer program]. R Foundation for Statistical Computing; 2021. [Google Scholar]
- 17.survminer: Drawing Survival Curves using ‘ggplot2’ [computer program]. 2021.
- 18.ggplot2: Elegant Graphics for Data Analysis [computer program]. Springer-Verlag; New York; 2016. [Google Scholar]
- 19.Fox J The R Commander: A Basic-Statistics Graphical User Interface to R. Journal of Statistical Software. 2005;14(9):1–42. [Google Scholar]
- 20.Kanda Y Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplantation. 2013;48(3):452–458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gander JC, Zhang X, Plantinga L, et al. Racial disparities in preemptive referral for kidney transplantation in Georgia. Clin Transplant. 2018;32(9):e13380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Schold JD, Buccini LD, Poggio ED, Flechner SM, Goldfarb DA. Association of Candidate Removals From the Kidney Transplant Waiting List and Center Performance Oversight. Am J Transplant. 2016;16(4):1276–1284. [DOI] [PubMed] [Google Scholar]
- 23.Schold JD, Arrigain S, Flechner SM, et al. Dramatic secular changes in prognosis for kidney transplant candidates in the United States. Am J Transplant. 2019;19(2):414–424. [DOI] [PubMed] [Google Scholar]
- 24.CMS. ESRD Network Organizations. 2019; https://www.cms.gov/Medicare/End-Stage-Renal-Disease/ESRDNetworkOrganizations/index.html. Accessed October 26, 2019.
- 25.Patzer RE, Paul S, Plantinga L, et al. A Randomized Trial to Reduce Disparities in Referral for Transplant Evaluation. J Am Soc Nephrol. 2017;28(3):935–942. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Husain SA, King KL, Pastan S, et al. Association Between Declined Offers of Deceased Donor Kidney Allograft and Outcomes in Kidney Transplant Candidates. JAMA Netw Open. 2019;2(8):e1910312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Patzer RE, Basu M, Smith KD, et al. Awareness of the New Kidney Allocation System among United States Dialysis Providers with Low Waitlisting. Am J Nephrol. 2018;47(2):115–119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Patzer RE, Smith K, Basu M, et al. The ASCENT (Allocation System Changes for Equity in Kidney Transplantation) Study: a Randomized Effectiveness-Implementation Study to Improve Kidney Transplant Waitlisting and Reduce Racial Disparity. Kidney Int Rep. 2017;2(3):433–441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Magua W, Basu M, Pastan SO, et al. Effect of the ASCENT Intervention to Increase Knowledge of Kidney Allocation Policy Changes Among Dialysis Providers. Kidney Int Rep. 2020;5(9):1422–1431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Waterman AD, Peipert JD. An Explore Transplant Group Randomized Controlled Education Trial to Increase Dialysis Patients’ Decision-Making and Pursuit of Transplantation. Prog Transplant. 2018;28(2):174–183. [DOI] [PubMed] [Google Scholar]
- 31.CMS. Technical Expert Panel Charter. Practitioner Level Measurement of Effective Access to Kidney Transplantation 2021; https://www.cms.gov/files/document/measurement-effective-access-kidney-transplantation-practitioner-level-tep-charter.pdf.
- 32.Chandraker A, Andreoni KA, Gaston RS, et al. Time for reform in transplant program-specific reporting: AST/ASTS transplant metrics taskforce. Am J Transplant. 2019;19(7):1888–1895. [DOI] [PubMed] [Google Scholar]
- 33.Patzer RE, Plantinga L, Krisher J, Pastan SO. Dialysis facility and network factors associated with low kidney transplantation rates among United States dialysis facilities. Am J Transplant. 2014;14(7):1562–1572. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.King KL, Husain SA, Jin Z, Brennan C, Mohan S. Trends in Disparities in Preemptive Kidney Transplantation in the United States. Clin J Am Soc Nephrol. 2019;14(10):1500–1511. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1. Crude and Adjusted Hazard Ratios of the Impact of the new Kidney Allocation System (KAS) policy compared to pre-KAS population on Referral, Evaluation, and Waitlisting among Incident patients with kidney failure 2012–2016 with follow-up through 2017 in Georgia, North Carolina, and South Carolina.
Figure S1. Cumulative incidence plots with Mantel Byar test for referral (panel a), evaluation start (panel b), overall waitlisting (panel c), and active waitlisting (panel d) for incident patients, 2012–2016 with follow-up through 12/31/2017.
Data Availability Statement
De-identified data can be shared with researchers with appropriate human subjects approvals and data use agreements.