Disparities in Provision of Transplant Information Affect Access to Kidney Transplantation

Authors


Dorry Segev, dorry@jhmi.edu

Abstract

Recently Centers for Medicare and Medicaid Services (CMS) began asking providers on Form-2728 whether they informed patients about transplantation, and if not, to select a reason. The goals of this study were to describe national transplant education practices and analyze associations between practices and access to transplantation (ATT), based on United States Renal Data System (USRDS) data from 2005 to 2007. Multinomial logistic regression was used to examine factors associated with not being informed about transplantation, and modified Poisson regression to examine associations between not being informed and ATT (all models adjusted for demographics/comorbidities). Of 236 079 incident end-stage renal disease (ESRD) patients, 30.1% were not informed at time of 2728 filing, for reasons reported by providers as follows: 42.1% unassessed, 30.4% medically unfit, 16.9% unsuitable due to age, 3.1% psychologically unfit and 1.5% declined counsel. Older, obese, uninsured, Medicaid-insured and patients at for-profit centers were more likely to be unassessed. Women were more likely to be reported as unsuitable due to age, medically unfit and declined, and African Americans as psychologically unfit. Uninformed patients had a 53% lower rate of ATT, a disparity persisting in the subgroup of uninformed patients who were unassessed. Disparities in ATT may be partially explained by disparities in provision of transplant information; dialysis centers should ensure this critical intervention is offered equitably.

Abbreviations: 
ATT, access to transplant; BMI, body mass index; CI, confidence interval; CMS, Centers for Medicare and Medicaid Services; ESRD, end-stage renal disease; KT, kidney transplantation; RR, relative rate; USRDS, United States Renal Data System.

 

Introduction

For many of the half million persons with end-stage renal disease (ESRD) in the United States, kidney transplantation (KT) is the preferred form of renal replacement, offering improved survival and quality of life (1–4). Each patient must weigh these benefits against the potential risks and costs and decide whether to pursue this modality. Previous studies have shown that patient education is an important component of informed decision making regarding ESRD treatment (5–10); however, in a recent multicenter survey, over 30% of ESRD patients reported having no knowledge of KT (11).

The nephrologist plays a key role in KT education (12). For most patients, the nephrologist is the first to discuss KT as a potential treatment and is responsible for referring interested patients to a transplant center for evaluation (13). The decision of whether to inform a patient about KT is left to the individual nephrologist, and there may be significant barriers to providing this information in a timely and comprehensive manner (14). While early referral is optimal (15), the nephrologist must provide counseling and management for myriad other issues relating to ESRD onset; as such, KT education may not be a priority. Furthermore, reimbursement for education is limited. For example, Medicare reimburses up to six sessions of ESRD education, which can include KT; however, these educational sessions must also be used for counseling on vascular access, comorbidity management, home and center dialysis options and prevention of uremic complications, leaving very little formally reimbursed time for KT education (16–18). Some patients are not even eligible for Medicare coverage until 1–4 months after ESRD onset, and as such their providers might not be reimbursed for any KT counseling during these months (19). Finally, formal protocols detailing the optimal timing, content and frequency of KT education, and which patients should receive it, are lacking.

Arguably, most patients should at least receive information about the option of KT in order to make an informed decision about whether to pursue further evaluation. While not every ESRD patient is an appropriate KT candidate, determining eligibility is not straightforward (13) as many patients fall into a gray area where they are neither definitively eligible nor ineligible, and these are ideally resolved through shared decision making between the patient, nephrologist and transplant center. Race, gender, age and income disparities in access to transplant (ATT) are well documented (13,20–28), and one potential mechanism for these disparities is inequitable provision of transplant information. The goals of this study were to (1) describe the provision of transplant information by nephrologists nationally, (2) identify patient characteristics associated with not being informed about transplant and (3) determine whether not being informed is associated with decreased ATT.

Methods

Study design

The study cohort included 236 079 adults who developed ESRD between January 1, 2005 (the date after which the new version of Form-2728 that required nephrologists to report whether each patient was informed about KT began distribution) and September 24, 2007 (the coverage date of the most recently available analytical file), as captured in the United States Renal Data System (USRDS). Patients who received a preemptive transplant (i.e. KT prior to ESRD-onset) were excluded from analysis. Data on comorbidities and provision of KT information were drawn from the 2005 version of the Centers for Medicare and Medicaid Services (CMS) 2728 Medical Evidence form which is filed by the patients’ nephrologist within 45 days of ESRD onset. Patients were followed until receipt of a transplant, death or end of study.

Reported reasons for not informing patients about transplant

In addition to reporting whether each patient was informed about KT, nephrologists were required to select a reason for not informing a patient. For a minority of patients (10.9%), more than one reason was selected; for statistical analyses, patients were assigned one reason only using the following prioritization: (1) psychologically unfit, (2) declined counseling, (3) medically unfit, (4) unsuitable due to age and (5) unassessed at time of 2728 filing. In other words, a patient reported as both medically unfit and unsuitable due to age would be classified as medically unfit; a patient reported as both having declined counseling and unsuitable due to age would be classified as having declined counseling; and only patients with none of the other reasons reported were classified as unassessed in our analyses. Nephrologists could also select “other” as a reason for not informing a patient; patients with only “other” selected comprised only 6.0% of those who were not informed and were excluded from analysis.

Associations between being informed and patient/center characteristics

A multinomial logistic regression model was constructed to identify center and patient characteristics associated with each of the above reasons given by a provider for not informing a patient about KT, compared to the reference outcome of informing the patient. Center characteristics of interest included profit status of the dialysis center and center volume; patient characteristics of interest included age, gender, African American race, underweight (BMI < 18.5), severe obesity (BMI > 35), insurance type and nephrologist care prior to ESRD onset. The model was also adjusted for the following patient characteristics: time on dialysis at the time Form 2728 was filed, congestive heart failure, atherosclerotic heart disease, other cardiac disease, cerebrovascular disease, peripheral vascular disease, amputation, smoking, alcohol dependence, drug dependence, inability to ambulate, diabetes, being institutionalized (in assisted living, a nursing home or other institution), needing assistance with daily living and cause of renal failure. Covariates were chosen using an integrative approach; we included those that were either statistically significant in univariate analysis or have been previously shown to be associated with access to transplantation or posttransplant outcomes (23). We compared the Akaike information criterion (AIC) for the full model to a model with each covariate systematically deleted; however, the AIC was minimized when all covariates were included.

Access to transplantation

ATT was defined as either joining the deceased donor waiting list or receiving a live donor KT, and modified Poisson regression was used as previously described (29) to analyze associations between receiving KT information and ATT. Separate analyses for the two components of ATT (joining the waitlist only and receiving a live donor KT only) were also performed to explore whether the association varied by type of ATT. All models were adjusted for all covariates listed earlier.

Unassessed patients

Reporting a reason for not informing a patient about KT at the time of Form 2728 filing (e.g. that the patient is unsuitable due to age) likely implied that further efforts to inform the patient would not be made. However, those patients who were reported as unassessed at the time of Form 2728 filing might be informed later in their ESRD course, and as such might represent a different type of KT information disparity. In order to more closely examine ATT for this subgroup, patients who were unassessed at the time of Form 2728 filing were matched to patients who were informed at that time, based on the following factors: time on dialysis at Form 2728 filing, age, cause of renal failure, diabetes (insulin dependent, oral medication dependent or not medication dependent as separate categories), insurance status, nephrologist care prior to ESRD onset, chronic obstructive pulmonary disease, cerebrovascular disease, artherosclerotic heart disease, peripheral vascular disease and malignant neoplasm/cancer. Patients with strong contraindications to KT (institutionalized, drug dependent or alcohol dependent) were excluded from this analysis. Rates of ATT for unassessed patients were compared to the matched informed patients using Kaplan–Meier methods, censoring for death.

Statistical analyses

All analyses were performed by using multiprocessor Stata version 11.0/MP for Linux (Stata-Corp, College Station, TX, USA). All hypothesis tests were two sided.

Results

Characteristics of informed and not informed patients

Among 236 079 incident ESRD patients, 70 945 (30.1%) were not informed about KT at the time of Form 2728 filing. Patients who were not informed were older (mean age 68.6 vs. 61.3), less likely to have private insurance (19.9% vs. 29.1%) and less likely to have nephrology care prior to ESRD onset (51.0% vs. 60.4%, Table 1, p < 0.001 for all comparisons). Uninformed patients also had higher rates of comorbidities such as artherosclerotic heart disease (27.2% vs. 20.2%), cardiac failure (40.6% vs. 30.9%), peripheral vascular disease (17.7% vs. 13.6%), chronic obstructive pulmonary disease (12.0% vs. 7.8%) and cancer (10.4% vs. 6.3%) (p < 0.001 for all). The most common reason reported for not informing a patient about KT was that the patient was unassessed at the time of Form 2728 filing (n = 29 866, 42.1% of uninformed patients). Only 3.3% of uninformed patients and 6.2% of unassessed patients gained ATT during the study period, compared to 14.1% of informed patients (Figure 1).

Table 1.  Characteristics of patients who were informed (left column) and not informed (right column) about kidney transplantation at the time of Form 2728 filing
 Informed, N = 165 234Not informed, N = 70 945p-Value
  1. All cells are percentage unless otherwise specified; SD = standard deviation.

Age, Mean (SD)61.3 (15.2)68.6 (14.3)<0.001
Female43.745.8<0.001
African American30.226.0<0.001
BMI, Mean (SD)28.7 (7.4)27.9 (7.4)<0.001
Nephrology care prior to ESRD onset60.451.0<0.001
Insurance type  <0.001
 Private29.119.9 
 Medicare38.448.2 
 Medicaid24.425.8 
 No insurance8.36.1 
Cause of ESRD  <0.001
 Diabetes46.243.7 
 Hypertension26.929.5 
 Glomerulonephritis7.74.9 
 Polycystic kidney disease2.60.8 
 Other16.621.1 
Diabetes   
 Insulin34.334.30.9
 Oral medication13.713.70.9
 Retinopathy9.77.8<0.001
Artherosclerotic heart disease20.227.2<0.001
Other cardiac problem14.219.2<0.001
Cardiac failure30.940.6<0.001
Peripheral vascular disease13.617.7<0.001
Cerebrovascular disease8.812.6<0.001
Cancer6.310.4<0.001
Chronic obstructive pulmonary disease7.812.0<0.001
Institutionalized4.412.1<0.001
Needs assistance8.416.8<0.001
Amputation3.03.8<0.001
Immobile5.111.4<0.001
Smoker6.56.20.025
Alcohol dependence1.51.9<0.001
Drug dependence1.41.60.035
Figure 1.

Reported reasons for not informing incident ESRD patients about kidney transplantation. ATT = access to transplantation, defined as joining the waitlist or receiving a kidney transplant from a live donor during the study period.

Reasons for not informing patients

After adjusting for multiple patient characteristics, patients at for-profit centers were 39% less likely reported as medically unfit and 34% less likely as psychologically unfit, but were 20% more likely reported as unassessed (Table 2). Older patients were more likely uninformed in general; conversely, patients who were under the care of a nephrologist prior to ESRD onset, and those with private insurance, were less likely uninformed in general. Patients with Medicaid or without insurance were 11% and 14% more likely to be unassessed, respectively. Medicaid patients were more likely to be reported as medically unfit and psychologically unfit. Women were 21% more likely reported as unsuitable due to age, 5% more likely reported as medically unfit (even after adjusting for all captured comorbidities) and 23% more likely to have declined information. African Americans were 27% more likely reported as psychologically unfit. Obese patients (BMI > 35) were 10% more likely unassessed and 14% more likely reported as medically unfit, but less likely reported as unsuitable due to age, declined information or psychologically unfit.

Table 2.  Relative rate of reported reasons for not informing patients about kidney transplantation; multivariate multinomial logistic regression accounting for dialysis center clustering, reference group is patients who were informed
 UnassessedUnsuitable due to ageMedically unfitDeclined counselPsychologically unfit
  1. Each cell shows relative rate (RR) on upper row and 95% CI below. RR signifies the relative rate that patients with a given characteristic were reported as uninformed for each reason, compared to being informed. For example, women were 23% more likely than men to be categorized as unsuitable due to age compared to being informed. Italics indicates that factor is associated ``with significantly increased rate of not being informed'' for reason of interest, compared to base outcome of being informed (significance defined as a p value <0.05); bold indicates that factor is associated ``with significantly decreased rate of not being informed'' for reason of interest, compared to base outcome of being informed (significance defined as a p value <0.05).

  2. 1Categorized as 18–39, 40–49, 50–59, 60–69, 70–79 and ≥80; RR reported as per-category (effectively, per decade).

  3. 2Patient was under the care of a nephrologist prior to ESRD onset.

  4. Model also adjusted for time on dialysis when 2728 form was filled out and comorbidities delineated in Table 1.

Age11.094.471.521.561.30
 (1.07–1.11)(4.31–4.63)(1.49–1.55)(1.46–1.66)(1.24–1.36)
Female1.011.211.051.231.08
 (0.99–1.04)(1.16–1.27)(1.02–1.09)(1.08–1.40)(0.99–1.19)
African American0.980.980.970.881.27
 (0.90–1.07)(0.90–1.08)(0.90–1.05)(0.74–1.05)(1.12–1.43)
BMI     
 18.5–35ReferenceReferenceReferenceReferenceReference
 <18.51.001.101.290.971.36
 (0.93–1.08)(1.00–1.21)(1.20–1.39)(0.71–1.32)(1.14–1.62)
 >351.100.851.140.890.73
 (1.06–1.15)(0.78–0.92)(1.09–1.20)(0.74–1.08)(0.63–0.84)
Nephrologist20.680.910.730.840.66
 (0.64–0.72)(0.84–0.98)(0.69–0.77)(0.72–0.97)(0.60–0.73)
Insurance     
 MedicareReferenceReferenceReferenceReferenceReference
 Private0.820.950.830.780.63
 (0.77–0.87)(0.87–1.05)(0.78–0.89)(0.64–0.95)(0.53–0.75)
 Medicaid1.111.021.111.172.13
 (1.05–1.18)(0.94–1.11)(1.05–1.18)(0.99–1.37)(1.89–2.40)
 None1.140.550.850.791.07
 (1.06–1.25)(0.44–0.69)(0.71–1.02)(0.54–1.15)(0.84–1.36)
Center     
 For-profit1.201.030.610.950.66
 (1.01–1.42)(0.87–1.21)(0.53–0.71)(0.75–1.21)(0.56–0.79)
 Center volume1.021.050.981.020.95
 (0.95–1.11)(0.95–1.15)(0.90–1.06)(0.90–1.16)(0.88–1.03)

Access to transplant

Similar to previous reports, patients at for-profit centers (30), older patients (23,26), women (20,23,28), African Americans (20,31), those with severe obesity (32) and those without insurance (20) had decreased ATT, while those with private insurance and those under the care of a nephrologist prior to ESRD onset had increased ATT (9). However, even after adjusting for all of these factors, not being informed about KT was independently associated with a 53% lower rate of ATT (51% less likely to join the waitlist and 65% less likely to find a live donor, Table 3). Furthermore, unassessed patients had a 46% lower rate of ATT compared to a matched cohort of patients who were informed (Figure 2).

Table 3.  Relative rate of access to transplant (left column), access to the waitlist (middle column), and receipt of a live donor kidney transplant (right column); multivariate modified Poisson regression
 Any accessWaitlistLive donor
  1. Each cell shows relative rate (RR) and 95% CI. RR signifies rate at which patients with each characteristic have access to transplant, adjusting for other characteristics. For example, patients at for-profit centers have a 20% lower rate of access to transplant than patients at nonprofit centers.

  2. 1Categorized as 18–39, 40–49, 50–59, 60–69, 70–79 and ≥80.

  3. 2Patient was under the care of a nephrologist prior to ESRD onset.

  4. Model also adjusted for comorbidities delineated in Table 1.

  5. All p-values <0.001.

Not informed0.470.490.35
 (0.45–0.49)(0.47–0.51)(0.31–0.41)
Age10.660.690.56
 (0.65–0.66)(0.68–0.69)(0.54–0.57)
Female0.870.870.84
 (0.85–0.89)(0.85–0.89)(0.79–0.90)
African American0.760.860.36
 (0.74–0.78)(0.84–0.88)(0.33–0.40)
BMI   
 18.5–35ReferenceReferenceReference
 <18.50.750.740.77
 (0.70–0.81)(0.69–0.80)(0.64–0.93)
 >350.710.780.54
 (0.69–0.74)(0.74–0.80)(0.48–0.59)
Nephrologist21.731.721.73
 (1.68–1.77)(1.67–1.77)(1.61–1.87)
Insurance type   
 MedicareReferenceReferenceReference
 Private1.671.602.07
 (1.62–1.72)(1.55–1.66)(1.89–2.27)
 Medicaid0.850.900.61
 (0.82–0.89)(0.86–0.93)(0.54–0.69)
 None0.750.790.54
 (0.71–0.79)(0.75–0.84)(0.46–0.64)
Center   
 For-profit0.800.820.72
 (0.78–0.82)(0.79–0.84)(0.68–0.78)
 Center volume1.051.080.94
 (1.04–1.06)(1.06–1.09)(0.91–0.97)
Figure 2.

Time to access to transplant, patients who were unassessed compared to a matched cohort of patients who were informed about transplant. Unassessed patients were matched to informed patients on time on dialysis when 2728 form was filled out, age, cause of renal failure, diabetes, insurance type, prior nephrologist care, needs assistance with daily living, chronic obstructive pulmonary disease, cerebrovascular disease, congestive heart failure, artherosclerotic heart disease, peripheral vascular disease and malignant neoplasm/cancer. Patients who were alcohol dependent, drug dependent or institutionalized for any reason were excluded.

Discussion

This national study identifies disparities in the provision of KT information that mirror previously documented disparities in ATT (13,20–28). Even after adjusting for many demographic factors and comorbidities, patients at for-profit centers, older patients, obese patients and patients with Medicaid or no insurance were significantly more likely to be unassessed. Women were more likely to be reported as unsuitable due to age, medically unfit and having declined information compared to men of the same age and comorbid conditions. Patients under the care of a nephrologist prior to ESRD onset, and those with private insurance were less likely to be uninformed about KT for any reason. Not being informed was associated with a 53% lower rate of ATT.

The most commonly reported reason for not informing a patient about transplantation was that the patient was unassessed at the time of Form 2728 filing, and even this reason for lack of KT information was associated with 46% lower rates of ATT. Given long waiting times and high death rates on dialysis (1), a delay in KT education for appropriate KT candidates leads to later referral and likely worse outcomes (15,33,34). Importantly, these unassessed patients never achieved ATT at rates similar to informed matched controls. That patients at for-profit centers and those with Medicaid or no insurance were more likely unassessed suggests that financial pressures at the center level may be driving this disparity. For-profit centers are less likely to be affiliated with transplant hospitals and as such may lack trained personnel dedicated to transplant education (30). Similarly, centers serving patients without insurance or with Medicaid may lack the resources for comprehensive and timely KT education (16). It is also possible that these patients are perceived to lack the financial resources necessary to complete the KT evaluation process, and as such provision of KT information is a lower priority.

Also concerning is that women were 23% more likely to be reported unsuitable due to age, consistent with previous findings that gender disparities in ATT are limited to older age groups (23). It is possible that providers are more likely to perceive older women as frail and unable to tolerate a major surgical procedure compared to men, and as such are less likely to provide KT information. It is important to note that women were also significantly more likely reported as having declined information; as such, it is possible that disparities in ATT for older women are partly patient driven.

Several limitations of this study are worth considering. CMS began collecting data on KT education in 2005 and as such patients have limited follow-up; however, a sample size of 236 079 incident ESRD patients was clearly sufficient to make inferences about characteristics associated with provision of KT information. Also, for most comorbidities, only presence or absence is captured on Form 2728, with little information about severity. Finally, what constitutes informing a patient about KT is not defined, and there may be significant variation in practices regarding provision of KT information that could not be captured.

Disparities in provision of kidney transplant information exist in the United States and correlate with disparities in access to transplantation, suggesting a possible mechanism for these disparities. National guidelines detailing the timing and content of transplant information for ESRD patients, as well as which patients should receive it, are needed to ensure that this important intervention is offered to all patients who might benefit.

Acknowledgments

The authors thank the Organ Procurement and Transplantation Network (OPTN) and the United States Renal Disease System (USRDS) for provision of the data. The OPTN is supported by Health Resources and Services Administration contract 234-2005-370011C. The analyses described here are the responsibility of the authors alone and do not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government.

Disclosure

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

Ancillary