PEDIATRICS Vol. 106 No. 4 October 2000, pp. 756-761

Racial Differences in Access to the Kidney Transplant Waiting List for Children and Adolescents With End-Stage Renal Disease

Susan L. Furth, MD, PhD^*, Pushkal P. Garg, MD^{, §}, Alicia M. Neu, MD^*, Wenke Hwang, PhD, Barbara A. Fivush, MD^*, and Neil R. Powe, MD, MPH, MBA^{, §, , ¶}

From the Departments of ^* Pediatrics and  Medicine; ^§ Robert Wood Johnson Clinical Scholars Program;  Health Policy and Management; and the ^¶ Welch Center for Prevention and Epidemiology and Clinical Research, Johns Hopkins Medical Institutions, Baltimore, Maryland.

	ABSTRACT

Top Abstract Methods Results Discussion References

Context.  Renal transplantation is the treatment of choice for pediatric patients with end-stage renal disease (ESRD). Black patients wait longer for kidney transplants than do white patients.

Objective.  To determine whether the increased time to transplantation for black pediatric patients is attributable not only to a shortage of suitable donor organs, but also to racial differences in the time from a child's first treatment for ESRD until activation on the cadaveric kidney transplant waitlist.

Design.  National longitudinal cohort study.

Setting.  US Medicare-eligible, pediatric ESRD population.

Patients.  Children and adolescents 19 years old at the time of their first dialysis for ESRD between 1988 and 1993, followed through 1996. Patients who received living donor renal transplants were excluded from study.

Main Outcome Measures.  Time from first dialysis for ESRD until activation on the kidney transplant waiting list, relative hazard of activation on the waiting list for black compared with white pediatric patients.

Results.  Comparisons of the time from first dialysis for ESRD to waitlisting among the 2162 white (60.7%) and 1122 black (31.5%) patients studied using survival analysis revealed that blacks were less likely to be waitlisted at any given time in follow-up. In multivariate analysis, even after controlling for patient age, gender, socioeconomic status, geographic region, incident year of renal failure, and cause of ESRD, blacks were 12% less likely to be waitlisted than were whites at any point in time (relative hazard: .88: 95% confidence interval: .79-.97).

Conclusions.  Racial disparities in access to the renal transplant waiting list exist in pediatrics. Whether these disparities are attributable to differences in time of presentation to a nephrologist, physician bias in identification of transplant candidates, or patient preferences warrants further study.  Key words:  children, race, renal transplantation, access.

Among adults with end-stage renal disease (ESRD), multiple studies have shown that access to kidney transplantation is not completely independent of patient income, gender, and race. Since 1988, reports have consistently shown that white, male, young, high-income patients are more likely to receive cadaveric kidney transplants than are patients without these characteristics.^1-5 In examining potential reasons for racial differences in access to transplantation among adult patients with ESRD, investigators have separated the time a patient waits for a kidney transplant into the time between the date of first dialysis treatment for ESRD and the date of first activation on the cadaveric kidney transplant waitlist, and the time after activation on the waitlist to receipt of a cadaveric kidney graft.⁶ Although black/white discrepancies in waiting times after activation on the waitlist can primarily be explained by immunologic factors,⁴ among adults with ESRD, black patients also wait longer than do white patients to be activated on the transplant waitlist.^6-9

Although data from the US Renal Data System (USRDS)¹⁰ show rates of pediatric renal transplantation do vary according to race, with black patients less frequently transplanted than other groups, racial differences in the time from onset of ESRD until activation on the kidney transplant waiting list for children and adolescents with ESRD has not previously been examined. Because, in pediatrics, transplantation is associated with improved patient survival, better growth, and lower cost than chronic dialysis, and additionally, pediatric patients seldom have comorbid conditions contraindicating transplantation, we explored whether racial disparities in access to the renal transplant waiting list exist in children and adolescents with ESRD after controlling for patient age, gender, cause of ESRD, geographic region, and socioeconomic status (SES).

	METHODS

Top Abstract Methods Results Discussion References

Study Design

We performed a national historical cohort study of patients who were 19 years old at the time of their first dialysis for ESRD, using several standard analytic files of the USRDS. Patients were included if they underwent their first dialysis between January 1, 1988 and December 31, 1993 and were followed until October 31, 1996. Patients who received a living donor kidney transplant during the follow-up were excluded from analysis, because patients with identifiable living donors frequently are not activated on the cadaveric transplant waitlist.

Data Sources and Variable Definition

Data from the USRDS Standard Analysis Files were used for this study. Our analysis used the patient file, from which we obtained information on patient age (4, 5-9, 10-14, and 15-19 years), race (white, black, Asian/Pacific Islander, Native American/Alaska Native, and other), sex, assigned cause of ESRD, date of first dialysis, date and donor source of first transplant, and ESRD network. The ESRD network is 1 of 18 contiguous geographic regions in the United States that collects and provides data to the Health Care Financing Administration on patients with ESRD. ESRD networks in which patients received care were grouped geographically into Northeast (networks 1-5), Southeast, (networks 6-8 and 13-14), Midwest (networks 9-12), and West (networks 15-18).

The USRDS designates assigned causes of ESRD according to International Classification of Diseases, Ninth Revision-Clinical Modification codes.¹¹ These codes were grouped into categories that were thought likely to influence whether a patient would be a potential transplant candidate. These groups consisted of: 1) genitourinary anomalies (589, 591, 599, 753); 2) focal and segmental glomerulosclerosis (FSGS; 581, 582, 587); 3) lupus nephritis (695, 710); 4) malignancy (189, 202); and 5) human immunodeficiency virus (0429, 0439, 0449); and 6) other, which encompassed multiple International Classification of Diseases, Ninth Revision categories including non-FSGS causes of glomerulonephritis, as well as hereditary and metabolic causes of renal failure, which were not likely to affect whether a patient was a transplant candidate.

These demographic and clinical data were linked with the date of the patients' first activation on the United Network for Organ Sharing waitlist that was obtained from the USRDS. We also utilized the patient treatment history file and the facility file to identify characteristics of the facility, which provided the majority of ESRD care during the first year of ESRD. Facilities providing the majority of care during the first year of ESRD were categorized according to USRDS facility certification codes as hospital-based transplant centers (certification codes 1 and 5), free-standing dialysis units (certification code 4), and hospital-based dialysis centers (certification codes 2 and 3).

The beneficiary's zip code of residence at first dialysis for ESRD was used to obtain zip code-specific data on median household income from the 1990 census. Zip code-specific median household incomes were grouped into quartiles: $20 943; >$20 943 and $26 674; >$26 674 and $33 617; and >$33 617. A total of 303 of patients (9.2%) were missing data on zip code-specific median household income. Patients missing data on SES were coded as such with a dummy variable for missing in the analysis.

Statistical Analysis

For analyses examining differences in clinical and demographic characteristics according to patient race, comparisons of proportions were performed using ² statistics. We used survival analysis (bivariate) and Cox proportional hazard analyses (multivariate) to examine the relationship between patient race and time from first dialysis for ESRD until first activation on the cadaveric transplant waitlist for an index transplant. If the date of a patients' first activation on the transplant waitlist preceded the date of his/her first dialysis, for statistical purposes the date of waitlisting was reset to 1 day after the initiation of dialysis. We used Kaplan-Meier survival (time-to-event) analyses to examine how age, gender, SES, and assigned cause of ESRD each affected time from first dialysis to date of first waitlisting. For initial cadaveric transplantation, survival curves were compared using the log-rank test. Cox proportional hazards analysis was then performed to examine the independent effect of race on the time from first dialysis for ESRD to activation on the kidney transplant waitlist, while controlling for multiple potential confounding factors. These included patient age, gender, assigned cause of ESRD, SES, incident year of ESRD, ESRD network, and facility characteristics. The final model included covariates that were significantly associated with activation on the waiting list in bivariate analyses or variables that were deemed clinically relevant for the analysis. The proportionality of hazards assumption was verified graphically. For both the Kaplan-Meier and Cox proportional hazards analyses, patients not activated on the waitlist were censored at death or at the end of the study.

We also performed stratified subgroup analyses by quartile of SES to examine whether the effect of race on access to the waitlist varies according to patient's SES. All analyses were performed using SAS statistical software (SAS, Cary, NC).¹²

	RESULTS

Top Abstract Methods Results Discussion References

Characteristics of Patients

There were 5448 patients 19 years old who had their first dialysis for ESRD between January 1, 1988 and December 31, 1993. Of these patients, 1889 who received kidney transplants from living donors during the follow-up period were excluded from analysis. The majority (81%) of the patients who received living donor transplants and were excluded were white. Because of small numbers, 275 patients of neither white nor black race: 54, Native American/Alaska Natives; 157, Asian American/Pacific Islanders; 25, patients of unknown race; and 39, other patients, were also excluded from study.

Table 1 shows the demographic characteristics of the 1122 black (31.5%) and 2162 white (60.7%) patients remaining in the study cohort of 3284 patients. Black children and adolescents with ESRD were older at presentation than were whites. Median age at presentation was 16.9 years for black patients and 14.8 years for white patients (P = .001, Table 1). Assigned causes of ESRD also differed by race, with blacks more frequently presenting with FSGS and systemic lupus erythematosis (SLE), and whites over-represented in the category of genitourinary anomalies. Median household incomes in the zip codes in which black children lived were lower than those of their white counterparts. A higher percentage of black patients received care in ESRD networks in the Southeastern United States.

Characteristics of Providers

Of the 3284 patients, 2488 (76%) had identifiable facilities providing the majority of dialysis care during the first year of ESRD. The remaining 796 (24%) had no identifiable dialysis provider as they were transplanted shortly after the initiation of dialysis (median: 100 days from first dialysis for ESRD until transplantation). Because we relied on dialysis treatment histories to identify providers and many children and adolescents with ESRD are transplanted within a few months of initiating dialysis, many have gaps in their dialysis histories that impaired our ability to identify a dialysis provider. The most common causes for gaps in dialysis histories leading to a missing dialysis provider in the first few months after the date of first ESRD service include Medicare eligibility delays, and a patient having other insurance, with Medicare as a secondary payer. Of those patients with identifiable providers, 1118 (34%) received care in hospital-based transplant centers, 976 (30%) received care in free-standing dialysis facilities, and 394 (12%) received care in hospital-based dialysis centers. There were 852 identifiable facilities providing care for the 3284 patients included in the analysis. The number of patients per facility ranged from 1 to 35.

Relationship Between Race and Time to Waitlisting: Kaplan-Meier Analysis

Over the follow-up, crude rates of waitlisting and transplantation differed by race. Among black patients, 28.6 patients were waitlisted per 100 person-years of follow-up, compared with 33.9 whites. Two hundred fifty-two patients (7.7%) were activated on the waiting list before their first dialysis, 76% of these patients were white. The cumulative incidence of patient waitlisting according to race is shown in Fig 1. Kaplan-Meier analyses demonstrated that at any point in time, black patients were less likely than were white patients to be activated on the kidney transplant waiting list (P = .007, log-rank). The disparity was present early on and remained relatively constant with time. At 2 years after first dialysis for ESRD, 56% of white patients and 50% of black patients were activated on the kidney transplant waitlist. At 5 years, 66% of white patients and 63% of black patients were waitlisted. The median days from first dialysis for ESRD to activation on the waitlist were 215 days for white patients and 275 days for black patients.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Cumulative incidence of wait listing for black and white patients after first dialysis for ESRD.

Other Factors Affecting Time to Activation on the Waiting List

Because factors other than race, including age, gender, assigned cause of ESRD, and SES, may affect access to the waitlist, we also examined these factors using Kaplan-Meier analyses. At any point in time, patients 4 years old were less likely to be activated on the transplant waiting list compared with patients in the other age groups (5-9, 10-14, and 15-19 years old; P = .0001, log-rank). Females were also less likely than males to be activated on the transplant waiting list over time (P = .0001, log-rank). At 2 years after first dialysis for ESRD, 64% of males and 52% of females were listed; at 5 years, 73% of males and 67% of females were listed. Similarly, patients in the highest quartile of SES were more likely to be listed than were those in the lowest quartile of SES at any time after first dialysis for ESRD (P = .05, log-rank). Median days to activation on the waitlist were 44 days for patients in the highest quartile, and 89 days for patients in the lowest quartile. Assigned cause of ESRD was also associated with time to waitlisting (P = .0001, log-rank). At any point in time patients with genitourinary anomalies causing renal failure were more likely than those with focal and segmental FSGS or SLE to be activated on the cadaveric transplant wait list.

Multivariate Analysis: Cox Proportional Hazards Analysis

To assess the independent effect of race on time from first dialysis until activation on the kidney transplant waiting list, we performed a Cox proportional hazards analysis (Table 2). Even after controlling for important confounders identified in the Kaplan-Meier analyses (including patient age, gender, SES, assigned cause of ESRD, as well as geographic region, and incident year of dialysis), black patients were 12% less likely than were white patients to be activated on the kidney transplant waiting list (relative hazard: .88; 95% confidence interval [CI]: .79,.97). Other factors that remained important predictors of activation on the transplant waiting list included male sex, age >4 years at presentation with ESRD, living in the northeastern compared with the southeastern United States, and genitourinary anomalies as a cause of renal failure compared with the reference group (Table 2). Additionally, patients with SLE were less likely than were patients with other causes of renal failure to be activated on the transplant waiting list. The inclusion of facility characteristics did not change the estimate of the effect of race on access to the waiting list.

Race and SES: Stratified Analyses

To determine whether the effect of race on access to the transplant waitlist differed according to SES, we performed a stratified analysis. We repeated the Kaplan-Meier analyses examining the effect of race on time to activation on the waitlist for patients in each quartile of median household income. Black/white differences in time to activation on the kidney transplant waitlist were statistically significant for patients in the lowest quartile of SES (P = .047, log rank). In the highest quartile, however, black/white differences were no longer demonstrable (P > .5, log-rank). Results of Cox proportional hazards multivariate stratified analyses demonstrated that after adjusting for age, sex, assigned cause of ESRD, network, and incident year of dialysis, the relative hazard for a black patient in the lowest quartile of SES being activated on the transplant waitlist was .84 (95% CI: .70,1.01), compared with relative hazard of 1.0 (95% CI: .8, 1.3) for a black patient in the highest quartile of SES.

	DISCUSSION

Top Abstract Methods Results Discussion References

In 1972 Congress passed legislation extending Medicare benefits to virtually all patients in the United States with end-stage renal failure.¹³ This funding was meant to achieve equal access to the lifesaving but expensive therapies of dialysis and transplantation for all patients with ESRD by removing financial barriers to care. Studies after this legislation was passed suggested that universal access to dialysis had been achieved. However, differences in access to transplantation continued to exist according to patient race.^1-9

To our knowledge, this is the first report to document that black/white differences in access to the renal transplant waiting list exist in the pediatric ESRD population. For children and adolescents with ESRD, transplantation is clearly the treatment goal, and the vast majority of patients present with primary renal disease and few significant comorbid conditions that preclude transplantation.¹⁴ Our multivariate analysis shows that even if patient age, gender, cause of ESRD, geographic region, and SES are equivalent, in the pediatric ESRD population, black patients are 12% less likely than are white patients to be activated on the cadaveric transplant waitlist at any time after presentation with renal failure. Stratified analyses suggested that racial differences in access to the transplant waitlist may vary according to SES. The black/white differences that may exist in the lowest quartile of SES are erased in the highest quartile of median household income, suggesting that high income may overcome the disparities in access to the transplant waitlist that seem to exist by race.

Black/white disparities in treatment decisions have been previously documented in internal medicine and in a variety of adult subspecialties.^15-18 In pediatrics, racial differences in access to specific medical therapies have less frequently been examined. Black/white differences in dialysis modality choice for children with ESRD have previously been described.¹⁹ We now demonstrate racial differences in time to activation on the transplant waitlist for pediatric patients with ESRD. These racial discrepancies cannot be explained by differences in insurance status, because Medicare coverage exists for virtually all US patients with renal failure.

Precise reasons for racial differences in time to activation on the transplant waiting list are not clear, although several possible explanations exist. Among adults, racial differences in medical suitability for transplantation have been proposed to explain this difference; however, our study suggests that variability in comorbid conditions between black and white patients do not explain racial discrepancies, because black/white differences exist even among the youngest and healthiest patients with ESRD. Black patients with renal failure may present to a nephrologist later in the course of renal disease, allowing less time for completion of a transplant workup. In general pediatrics, black children and adolescents less frequently have a usual source of health care,²⁰ and this may contribute to the timing of presentation with advanced kidney disease. Compliance with therapy has been suggested as a possible explanation for racial differences in the identification of potential transplant candidates; however, at least 1 study of compliance issues posttransplant suggests that noncompliance is a by-product of SES, rather than race,²¹ and this was adjusted for in our analysis. The possibility of physician bias as a cause of racial differences in identifying potential transplant candidates cannot be ruled out.^7-9,18 Multiple studies suggest that graft survival is worse for black patients posttransplant, and this may on some level influence providers to ration the scarce resource of cadaveric kidneys to those patients who they believe are likely to do well.²² Additionally, as reported by the North American Pediatric Renal Transplant Cooperative Study, a substantial percentage of patients on dialysis are not transplant candidates either for medical reasons or because of family/patient preference.²³ It is not clear whether these factors differ by race. Differences in patient knowledge and preferences for transplantation among black and white adults with ESRD have been described.²⁴ Patient preferences were not measured in our analysis.

Our study demonstrates that among pediatric patients with ESRD, black patients are 12% less likely than are white patients to be activated on the transplant waiting list even after controlling for multiple important confounding factors. A similar analysis, carefully controlling for comorbidity, using several special studies of adult ESRD in the USRDS demonstrated that among adults with renal failure, blacks were 32% less likely than were whites to be activated on the waiting list.⁹ The difference between the adult and pediatric population may in part be related to the increased emphasis on transplantation in the pediatric age group,¹³ or differences in how ESRD care is delivered to this population. In contrast to adult patients with ESRD, 77% of whom received dialysis care in free-standing dialysis facilities,¹⁰ pediatric patients with ESRD primarily receive care in hospital-based facilities, frequently associated with transplant centers. This may lessen the barrier to completing a pretransplant workup, which has been identified as a particular step in the transplant process where differences exist according to patient race.⁸

Our study has several strengths. The population was national, including all pediatric patients enrolled in the Medicare ESRD program between 1988 and 1993, with follow-up for a minimum of 3 and a maximum of 8 years. Although there is an 18-month waiting period for Medicare eligibility for patients <65 years of age with other primary insurance, and during this time patient data may be incomplete in the Medicare files, the lag time between when these patients entered the Medicare files and when the data were analyzed should ensure that the data on these patients were complete. Therefore, our study population should include all incident Medicare-eligible pediatric patients in the years 1988-1993, which the USRDS has estimated as 92% of the US ESRD population.¹⁰

Several limitations of our study deserve mention. This is an analysis of administrative data, which are not collected with as much care as would be data in a prospective cohort study. However, any errors in the recorded dates of first dialysis and first waitlisting would be expected to be random, and not associated with patient race. This random misclassification would tend to minimize rather than to exaggerate the racial differences in time from first dialysis for ESRD to activation on the waitlist that we saw.

Also, as the USRDS contains no direct data on SES, we adjusted for SES based on zip code-based census data. This may not be as desirable as using patient-specific income data. However, recent studies have provided evidence that zip code-based measures of SES are robust and reliable proxies for patient socioeconomic conditions.²⁵

One final consideration in our study of race and access to transplantation in the pediatric ESRD population is the issue of living donor transplantation. Because living donor transplants in the pediatric population make up such a large proportion of pediatric transplant recipients, some have argued that access to the cadaveric waitlist is a misleading measure of access to transplantation for pediatric ESRD patients.⁶ Because many patients with identifiable living donors are never activated on the cadaveric transplant waitlist, we excluded these 1889 patients from analysis. Because the rate of pediatric living donor transplantation is lower in black that in white patients,¹⁰ this exclusion would be likely to leave black patients overrepresented in our analysis but would not be expected to introduce bias in the time from first dialysis for ESRD until activation on the cadaveric waitlist.

Before Medicare was extended to all patients with ESRD, social considerations were used to determine who would receive treatment.¹³ Comparisons of the dialysis patient population before and after the extension of Medicare benefits reflect this practice. Pre-Medicare extension to ESRD, in 1967, the dialysis population was 25% female, 41.7% employed, and only 7% black. Post-Medicare extension, in 1978, the dialysis population had shifted to 51% female, 18.4% employed, and 35% black. Although characteristics of the dialysis population have shifted to reflect the demographics of incident patients with renal failure in the United States, characteristics of the population with ESRD receiving kidney transplants have not changed as dramatically. Our study shows that racial differences in access to renal transplantation in pediatrics occur at the early step of activation on the cadaveric transplant waiting list. Exact causes for this discrepancy cannot be defined by our analysis; however, racial differences in the time of presentation with renal failure, patient preferences for transplantation, or differences in physician's identification of transplant candidates remain possible explanations. Further investigation into causes for these persistent racial differences is necessary, so targeted interventions that will ensure equal access to transplantation for all pediatric patients with renal failure, regardless of race, can be instituted.

	ACKNOWLEDGMENTS

Dr Furth is supported by Grant K08 DK02586-01A1 from the National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Maryland. Dr Garg is a Robert Wood Johnson Clinical Scholar. Dr. Powe is supported by Grant K24 DK02643 from the National Institute of Diabetes and Digestive and Kidney Diseases.

The data reported here have been supplied by the USRDS. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the US government.

	FOOTNOTES

Received for publication Aug 3, 1999; accepted Feb 25, 2000.

This work was presented in part at the Pediatric Academic Societies' 1999 Annual Meeting; May 1-4, 1999; San Francisco, CA; and the American Society of Transplantation 18th Annual Meeting; May 15-19, 1999; Chicago, IL.

Reprint requests (S.L.F.) Pediatrics, 600 N Wolfe St/Park 327, Baltimore, MD 21287. E-mail: sfurth{at}jhmi.edu

	ABBREVIATIONS

ESRD, end-stage renal disease; USRDS, US Renal Data System; SES, socioeconomic status; FSGS, focal and segmental glomerulosclerosis; SLE, systemic lupus erythematosis; CI, confidence interval.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Held PJ, Pauly MV, Bovberg RR, Newman J, Salvaterra O Access to kidney transplantation: has the US eliminated income and racial differences? Arch Intern Med 1988; 148:2594-2600 [Abstract/Free Full Text]
2. Kjellstrand CM Age, sex and race inequality in renal transplantation. Arch Intern Med 1988; 148:1305-1309 [Abstract/Free Full Text]
3. Kasiske BL, Neylan JF, Riggio RR, The effect of race on access and outcome in transplantation. N Engl J Med 1991; 324:302-307 [Medline]
4. Sanfilippo FP, Vaughn WK, Peters TG, Factors affecting the waiting time of cadaveric kidney transplant candidates in the United States. JAMA 1992; 267:247-252 [Abstract/Free Full Text]
5. Gaylin DS, Held PJ, Port FK, The impact of comorbid and sociodemographic factors on access to renal transplantation. JAMA 1993; 269:603-608 [Abstract/Free Full Text]
6. Eggers PW Racial differences in access to kidney transplantation. Health Care Financing Rev 1995; 17:89-103 [Medline]
7. Soucie JM, Neylan JF, McClellan W. Race and sex differences in the identification of candidates for renal transplantation. Am J Kid Dis. 1992;19;414-419
8. Alexander GC, Sehgal AR Barriers to cadaveric renal transplantation among blacks, women and the poor. JAMA 1998; 280:1148-1152 [Abstract/Free Full Text]
9. Garg PP, Diener-West M, Powe NR Access to the kidney transplant waiting list: are African-Americans held to a higher standard? J Am Soc Nephrol 1998; 9:674A
10. US Renal Data System. USRDS 1998 Annual Data Report. Bethesda, MD: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 1998
11. Baierschmidt C. International Classification of Diseases, Ninth Revision-Clinical Modification. 5th ed. Salt Lake City, UT: Medicode Publications;1998
12. SAS Institute Inc. SAS Procedures Guide, Version 6.03. Cary, NC: SAS Institute Inc; 1988
13. Evans RW, Blagg CR, Bryan FA Implications for health care policy: a social and demographic profile of hemodialysis patients in the United States. JAMA 1981; 245:487-491 [Abstract/Free Full Text]
14. Warady BA, Hebert D, Sullivan EK, Alexander SR, Tejani A Renal transplantation, chronic dialysis and chronic renal insufficiency in children and adolescents: the 1995;annual report of the NAPRTCS. Pediatr Nephrol 1997; 11:49-64 [CrossRef][Medline]
15. Whittle J, Conigliaro J, Good CB, Lofgren RP Racial differences in the use of invasive cardiovascular procedures in the Department of Veterans Affairs medical system. N Engl J Med 1993; 329:621-627 [Abstract/Free Full Text]
16. Goldberg KC, Hartz AJ, Jacobsen SJ, Krakauer H, Rim AA Racial and community factors influencing coronary artery bypass graft surgery rates for all 1986 Medicare patients. JAMA 1992; 267:1473-1477 [Abstract/Free Full Text]
17. Barker-Cummings C, McClellan W, Soucie JM, Krisher J Ethnic differences in the use of peritoneal dialysis as initial treatment for end stage renal disesase. JAMA 1995; 274:1858-1862 [Abstract/Free Full Text]
18. Schulman KA, Berlin JA, Harless W, The effect of race and sex on physicians' recommendations for cardiac catheterization. N Engl J Med 1999; 340:618-626 [Abstract/Free Full Text]
19. Furth SL, Powe NR, Hwang W, Neu AM, Fivush BA. Racial differences in choice of dialysis modality for children with end-stage renal disease. Pediatrics. 1997;99(4). URL: http://www.pediatrics.org/cgi/content/full/99/4/e6
20. Newacheck PW, Hughes DC, Stoddard JJ Children's access to primary care: differences by race, income, and insurance status. Pediatrics 1996; 97:26-32 [Abstract/Free Full Text]
21. Rovelli M, Palmeri D, Vossler E, Bartus S, Hull C, Schweizer R Noncompliance in renal transplant recipients: evaluation by socioeconomic groups. Transplant Proc 1989; 21:3979-3981 [Medline]
22. Lowe M, Kerridge IH, Mitchell KR These sorts of people don't do very well: race and allocation of health care resources. J Med Ethics 1995; 21:356-360 [Abstract/Free Full Text]
23. Lerner GR, Warady BA, Sullivan EK, Alexander SR Chronic dialysis in children and adolescents. Pediatr Nephrol 1999; 13:404-417 [CrossRef][Medline]
24. Ayanian JZ, Cleary PD, Weissman JS, Epstein AM The effect of patients' preferences on racial differences in access to renal transplantation. N Engl J Med 1999; 341:1661-1669 [Abstract/Free Full Text]
25. Krieger N Overcoming the absence of socioeconomic data in medical records: validation of a census-based methodology. Am J Public Health 1992; 82:703-710 [Abstract/Free Full Text]