Abstract
Background and objectives Many factors have been shown to be associated with ESRD patient placement on the waiting list and receipt of kidney transplantation. Our study aim was to evaluate factors and assess the interplay of patient characteristics associated with progression to transplantation in a large cohort of referred patients from a single institution.
Design, setting, participants, & measurements We examined 3029 consecutive adult patients referred for transplantation from 2003 to 2008. Uni- and multivariable logistic models were used to assess factors associated with progress to transplantation including receipt of evaluations, waiting list placement, and receipt of a transplant.
Results A total of 56%, 27%, and 17% of referred patients were evaluated, were placed on the waiting list, and received a transplant over the study period, respectively. Older age, lower median income, and noncommercial insurance were associated with decreased likelihood to ascend steps to receive a transplant. There was no difference in the proportion of evaluations between African Americans (57%) and Caucasians (56%). Age-adjusted differences in waiting list placement by race were attenuated with further adjustment for income and insurance. There was no difference in the likelihood of waiting list placement between African Americans and Caucasians with commercial insurance.
Conclusions Race/ethnicity, age, insurance status, and income are predominant factors associated with patient progress to transplantation. Disparities by race/ethnicity may be largely explained by insurance status and income, potentially suggesting that variable insurance coverage exacerbates disparities in access to transplantation in the ESRD population, despite Medicare entitlement.
Introduction
Timely access to care is critical for ESRD patients considering kidney transplantation as a treatment modality. There is a significant survival benefit associated with kidney transplantation relative to the alternative treatment of dialysis, and most patients receive an estimated doubling of life expectancy across age, race, gender, and primary diagnosis (1). In addition, among patients who receive a kidney transplant, longer waiting times and elevated exposure to dialysis have a deleterious effect on outcomes after the procedure (2). However, only a minority of ESRD patients are considered appropriate candidates for kidney transplantation.
To receive a transplant, patients must first be evaluated to determine whether they are appropriate for the procedure based on physical, psychologic, and social support availability, and financial viability. After an evaluation, patients may be placed on a waiting list to receive a deceased donor transplant or seek living donors. Particularly in recent years, expanding waiting lists and increased mortality for candidates awaiting transplantation enhance the need for patients to be evaluated and placed on waiting lists rapidly (3).
There have been several studies identifying disparities in access to kidney transplantation. Based on national data, Wolfe et al. (4) reported that wait listing rates were significantly reduced among African Americans, women, and patients with diabetes as a primary diagnosis. Moreover, there has been minimal change in delays in listing among African Americans over the past decade (5). Other population-based studies have identified lower income/socioeconomic status, geographical region, and elevated body mass index as factors that are independently associated with delayed access to kidney transplantation and transplantation with lower-quality donor organs (6–9).
In addition to population-based studies, there have been several studies describing factors associated with patient progression to receive a kidney transplant. Ayanian et al. (10) identified disparities for African Americans to be referred, placed on waiting lists, and ultimately receive a transplant. Importantly, this study indicated that a portion of these disparities could be explained by patient preferences and lack of certainty about wanting to undergo the procedure. Alexander and Sehgal (11) identified differences in progression for ESRD patients by gender, race, primary diagnoses, income, and age toward receiving a transplant at several different intervals in the process. Epstein et al. (12) concluded that racial disparities among ESRD patients interested in receiving a kidney transplant were a product of both inappropriate classifications of candidacy among sicker Caucasians and inappropriate exclusions among relatively healthy African Americans.
The aim of this study was to evaluate outcomes for adult patients referred to our center for kidney transplantation to (1) determine factors that are associated with progression toward receiving a transplant in a single-center population and whether these factors are consistent with observations of prior studies, (2) identify specific stages at which differences in progression toward receiving a transplant are most evident and potential causes of these differences, and (3) determine whether patient characteristics that have been previously associated with progression to transplantation were explained or modified by other factors. The potential use of this study is to identify specific causes of access disparities in the transplantation process that may suggest interventions in which disparities can be ameliorated.
Materials and Methods
The study evaluated consecutive adult patients referred for kidney transplantation to the University of Florida from January 2003 to July 2008, with follow-up information available through December 2008. Data were derived from an internal database in which events and patient characteristics are routinely captured. Primary outcomes of the study were patient likelihood to receive an evaluation, to be placed on the waiting list, and to receive a transplant from the time of referral. Importantly, the policy of this center is to place patients on the active waiting list while potential living donors are evaluated. As such, failure to place patients on the national waiting list for a deceased donor transplant is generally not reflective of patient desire/opportunity to acquire a living donor. This program does not have any explicit upper age limit or other policies excluding transplant evaluations other than standard contraindications such as active malignancies. Evaluation and listing practices were relatively consistent over the study period and were conducted by a multidisciplinary transplant team.
Primary reasons for not receiving an evaluation or being placed on the waiting list were collected from the database. Patient information available at the time of referral included demographic characteristics, primary diagnoses, history of a prior transplant, primary insurance coverage, and zip code of patient residence. Race/ethnicity was classified based on medical records initially accompanied by patient referrals, as indicated by the patient's primary physician or caregiver. Race/ethnicity was considered a primary variable of interest based on literature identifying an association with access to care. Based on zip code information, we supplemented the median county income for patient residence based on 2000 U.S. census data (13). Using the SAS function zipcitydistance, we estimated the miles between patient residence and the transplant center.
Multivariable logistic models were used to assess factors associated with referred patient likelihood to be evaluated and placed on the transplant waiting list. Variables were selected for the models based on the availability of information and past research indicating an association with the primary outcomes along with year of referral to account for differences in time to progress through transplant processes. For the outcome of wait listing, separate nested logistic models were constructed using different sets of covariates. This was conducted to evaluate the relative contribution of related factors and the influence of the set of factors on the likelihood to be placed on the waiting list. Models were estimated for goodness of fit using the Hosmer-Lemeshow test, and concordance indices were reported. Because some patients received multiple referrals, only the first referral was considered for the analysis. All analyses were conducted with SAS (v.9.1.; SAS Institute, Cary, NC), and a type I error probability of 0.05 was used to construct confidence intervals and as the threshold for statistical significance. This study was approved by the Institutional Review Board.
Results
There were a total of 3029 patients included in the study population. Table 1 shows the distribution of the population characteristics. Patients were most commonly in the 50- to 59-year age range, male, Caucasian, and with commercial insurance at the time of referral. Overall, 56% of referred patients received an evaluation, 27% were placed on the waiting list for transplantation, and 16% received a transplant either from a deceased or living donor. Among patients who received an evaluation, 95% occurred within 1 year of referral, and patients who had >1 year to evaluation were more likely to be African American, reside in a lower income county, and have noncommercial insurance. Seventy-seven percent of transplants derived from deceased donors, 15% from living related donors, and 8% from nonrelated living donors. As indicated in Table 1, younger patients, Caucasian, patients with commercial insurance, patients residing in counties with the highest median income, and patients with polycystic kidney disease were more likely to receive evaluations, be placed on the waiting list, and receive a transplant.
Proportion of referred patients evaluated, wait listed, and receiving a transplant (n = 3029)
Among documented causes for not receiving an evaluation, the most common were denials for medical reasons (31%), patient refusals (23%), insurance or financial problems (16%), and death before evaluation (12%). Refusals for an evaluation were more common among older patients (27% among patients older than 70 years versus 7% among patients 18 to 29 years of age, P < 0.01) and patients without a prior transplant (18% among patients without a prior transplant versus 5% among patients with a prior transplant, P < 0.01). Patients with commercial insurance (21%) and Medicare (23%) were more likely to refuse evaluations relative to patients with Medicaid as insurance (8%, P = 0.01).
Table 2 shows the adjusted likelihood of an evaluation from referral. Progressively older age was associated with the likelihood not to receive an evaluation; race/ethnicity was not associated with likelihood of an evaluation including similar adjusted odds ratios (AORs) for African Americans relative to Caucasians (AOR = 0.95, 95% confidence interval [CI] = 0.79 to 1.13). Patients with diabetes and hypertension as primary diagnoses were less likely to receive an evaluation relative to other causes. Patients with noncommercial insurance were also significantly less likely to receive an evaluation relative to patients with commercial insurance.
Adjusted odds ratios for likelihood of an evaluation from referral
Three models are shown assessing the likelihood not to be placed on the waiting list (Table 3). As indicated by the initial model adjusted only for demographic factors and year of referral, progressively older age and African Americans (AOR = 1.47, 95% CI = 1.22 to 1.75) were less likely to be wait listed. The second model, additionally adjusted for certain clinical factors, indicated that, with additional adjustment, age, African American race/ethnicity (AOR = 1.30, 95% CI = 1.08 to 1.57), primary diagnoses, and prior transplant recipients were less likely to be listed. The final model indicates that, with additional adjustment for median income, insurance type, and distance to center, African American race/ethnicity (AOR = 1.14, 95% CI = 0.94 to 1.40) was no longer statistically significantly less likely to be listed. Significant factors in the final model were older age, primary diagnosis, lower median income, and noncommercial insurance type. The predictive value of the models significantly increased with each model as reflected by the concordance indices.
Adjusted odd ratios and 95% confidence intervals of referred patients not being placed on the waiting list for transplantation
Among patients with commercial insurance, there was no difference in the adjusted likelihood for failure to be placed on the waiting list (AOR = 0.93, 95% CI = 0.72 to 1.19) between African Americans and Caucasians (Table 4). However, the likelihood for failure to be placed on the waiting list was higher for African Americans with noncommercial insurance relative to Caucasians with noncommercial insurance (AOR = 1.47, 95% CI = 1.07 to 2.01). This association was generally consistently limited to patients who had received an evaluation, with similar likelihood of listing between African Americans and Caucasians with commercial insurance and numerically lower likelihood of listing for African Americans without commercial insurance (Table 5). Among patients with commercial insurance, the majority (84%) also had Medicare listed as secondary insurance. In general, patients with both forms of insurance (rather than commercial insurance alone) had a numerically higher likelihood of receiving an evaluation and being placed on the waiting list.
Likelihood for referred patients to not be placed on the waiting list based on interaction of race/ethnicity and primary insurancea
Likelihood for referred patients to not be placed on the waiting list based on interaction of race/ethnicity and primary insurance conditional on receiving a transplant evaluationa
Factors significantly associated with the adjusted likelihood of receipt of a transplant over the study period included age, race/ethnicity, prior transplants, income, insurance, and year of referral (Table 6). Noncommercial insurance, older age, and repeat transplant patients were significantly less like to receive a transplant after referral. In addition, despite a lack of significant difference of African Americans to be placed on the waiting list for transplant, this group had a significantly reduced likelihood to receive a transplant.
Likelihood for referred patients to receive a kidney transplanta
Discussion
The primary findings of this study indicate that there are several significant factors associated with patient progression toward receiving a kidney transplant in our referral population. Consistent with prior research findings, older age, minority race/ethnicity, diabetes and hypertension as primary diagnoses, and lower income were associated with placement on the kidney transplant waiting list. Interestingly, socioeconomic factors seemed to be significant modifiers of the differences between African Americans and Caucasians for the propensity to be placed on the waiting list in our population. In particular, differences in the likelihood of listing for African Americans were significantly reduced, and no longer statistically significant, with adjustment for socioeconomic factors. Furthermore, the primary differences in placement on the waiting list by race/ethnicity were not evident among patients with commercial insurance, but rather only among patients with noncommercial forms of insurance. These results may suggest that variations in health insurance coverage is a primary modifying factor explaining previously documented racial/ethnic disparities in access to care among patients referred for transplantation despite the presence of Medicare entitlement for this population.
One of the challenges for understanding the causes of racial/ethnic outcomes in the ESRD population involves identifying whether findings are a reflection of differences or disparities in care. African Americans have cited high levels of discrimination during the transplant process, and it is possible that provider behaviors significantly impact lower listing rates among this population (14). Household poverty has also been shown to partially explain disparities between prevalence of chronic kidney disease between Caucasians and African Americans (15). However, African Americans also have significantly reduced graft survival compared with other race/ethnic groups and as a group may have a higher prevalence of underlying clinical contraindications for the procedure (16). In this case, these issues are salient in determining whether the likelihood to be listed reflects appropriate clinical decision-making, patient preferences, interactions with providers, or other socioeconomic or logistical factors.
One important novel contribution of this study is to suggest that rather than attributing differences in access to care uniformly between African Americans and Caucasians, the disproportionate placement on the waiting list between racial groups appears heightened among patients with noncommercial forms of insurance coverage. It is possible that the financial burden or logistical challenges such as travel or acquiring community or social support have a disproportionate effect by race/ethnicity for dissuading patients who are medically viable to seek care (17). In a recent study, Axelrod et al. (18) nicely represented the many complex geographic and logistical barriers for access to care in the kidney transplant population. Insurance coverage is a well-documented factor independently shown to be associated with progression of ESRD patients to receive care (7,19,20). A recent study suggested that time to listing for transplantation largely explains the association of pretransplant dialysis time and diminished transplant outcomes and that race and insurance are primary factors explaining delays in listing (21). This study potentially extends these associations by suggesting that insurance coverage may have a differential effect by racial/ethnic group.
Despite many efforts to improve access to care and identify disparities in the ESRD population, lower socioeconomic status and African-American race/ethnicity remain predominant factors explaining delays and failure to complete steps toward receiving a transplant in this population. As identification of these differences/disparities alone does not seem to have resulted in profound changes in the United States, enhanced efforts to develop more potent interventional programs are needed. These may include educational programs and culturally appropriate dissemination of the potential benefits of transplantation. Rodrigue et al. (22) determined that home-based educational programs were more effective at increasing living donor rates than clinic-based educational programs, and as such, these programs may also be important across race/ethnic groups. Contrary to previous studies, our study did not indicate that differences in placement on the waiting list by race/ethnicity were explained by differences in evaluations (12,23). This study suggests that other factors following evaluation are primary drivers of differences in placement on the waiting list. Despite equivocal likelihood of listing, African Americans did have a reduced likelihood to receive a transplant, which may be, at least in part, explained by factors in organ allocation algorithms such as HLA matching and regional distribution. Shands at the University of Florida was one of 11 kidneys transplant centers in the state of Florida and one of the top four centers for kidney transplant volume in the state during this study. Waiting times are generally shorter in the region relative to other United Network for Organ Sharing regions, and on average, there is a higher representation of African Americans at this center relative to the country. The degree to which the results of this study are generalizable to other populations will require further study.
There have been several studies in cancer research indicating that levels of care and outcomes between African Americans and Caucasians are no longer evident when controlling for socioeconomic status (24–26). In the context of cardiovascular procedures, one study determined that disparities by race/ethnicity were evident among patients except those who were privately insured (27). Haider et al. (28) also indicated that insurance status was the primary explanatory factor associated with outcomes after trauma and a significantly larger driver of outcomes compared with race. Among transplant recipients, Woodward et al. (29) concluded that uniform insurance coverage significantly reduces disparities associated with socioeconomic status after the transplant procedure. One potential interpretation of the culmination of past studies is that uniform forms of insurance may have the potential to ameliorate disparities in care between racial/ethnic groups in certain healthcare settings.
There are several notable limitations that we must acknowledge for this study. Primarily, there are many potential factors that may be associated with the likelihood to be evaluated and placed on the waiting list for transplantation that were not available for this study. In particular, certain patient comorbidities, employment status, patient preferences, educational attainment, cultural/linguistic barriers, and individual income levels (rather than median county incomes used in the study) have all been shown to impact outcomes for this population and were not available for these analyses (7,9–11,19). As such, the magnitude of the effects shown in this study may be affected without adjustment for these factors. For example, patient age may be a proxy for comorbid conditions that were not captured at the time of referral and do not completely represent the independent effect of age alone. In addition, we also did not have systematic collection of patient blood type, which is a strong explanatory factor for time to transplantation. We are also limited in determining the degree to which differences in placement on the waiting list are explained by provider decisions as opposed to patient preferences. Furthermore, results from this single-center study may not be evident in other settings, and further validation of these findings is needed. Center-level differences may include attitudes of physicians responsible with placing patients on the waiting list, the specific processes involved in scheduling and identifying candidates, specific characteristics of the referral population in this region, teaching and for-profit institutional approaches, and the status of the program with respect to quality oversight.
Conclusions
This study identified significant factors associated with patient progression toward receiving a transplant after referral at a single institution. These factors included older age, African American race/ethnicity, lower income, and noncommercial insurance coverage. Lower rates of placement on the waiting list for African Americans were largely explained by income and insurance coverage. Disparities in placement on the waiting list between African Americans and Caucasians were not evident among patients with commercial insurance. Further understanding of the sources of impaired access to transplant among lower socioeconomic patients, identification of effective interventional programs and the applicability of these findings in novel settings are needed.
Disclosures
None.
Acknowledgments
We thank Suzanne Johnson for valuable contributions to the production of this study, Dr. Richard Howard for review of the manuscript, and David Tomlin for significant assistance with data management.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
- Received September 28, 2010.
- Accepted February 28, 2011.
- Copyright © 2011 by the American Society of Nephrology
References
In this issue
Jump to section
More in this TOC Section
Original Articles
Renal Transplantation
Cited By...
- Association of Socioeconomic Status and Comorbidities with Racial Disparities during Kidney Transplant Evaluation
- Access to Kidney Transplantation after a Failed First Kidney Transplant and Associations with Patient and Allograft Survival: An Analysis of National Data to Inform Allocation Policy
- Subclinical Cognitive Impairment and Listing for Kidney Transplantation
- Frailty and Access to Kidney Transplantation
- Native American patients perception and attitude about kidney transplant: a qualitative assessment of patients presenting for kidney transplant evaluation
- Impact of Navigators on First Visit to a Transplant Center, Waitlisting, and Kidney Transplantation: A Randomized, Controlled Trial
- How important is social support in determining patients suitability for transplantation? Results from a National Survey of Transplant Clinicians
- The Affordable Care Act, Kidney Transplant Access, and Kidney Disease Care in the United States
- Transplant Center Patient Navigator and Access to Transplantation among High-Risk Population: A Randomized, Controlled Trial
- Standardized Transplantation Referral Ratio to Assess Performance of Transplant Referral among Dialysis Facilities
- Early Changes in Kidney Distribution under the New Allocation System
- Nonadherence to therapy after adult solid organ transplantation: A focus on risks and mitigation strategies
- Assessing Transplant Education Practices in Dialysis Centers: Comparing Educator Reported and Medicare Data
- Assessing pharmacologic and nonpharmacologic risks in candidates for kidney transplantation
- Illinois Law Opens Door To Kidney Transplants For Undocumented Immigrants
- Health-Related and Psychosocial Concerns about Transplantation among Patients Initiating Dialysis
- Modifiable Patient Characteristics and Racial Disparities in Evaluation Completion and Living Donor Transplant
- Impact of Navigators on Completion of Steps in the Kidney Transplant Process: A Randomized, Controlled Trial
- Impact of a Patient Education Program on Disparities in Kidney Transplant Evaluation