Customize your JAMA Network experience by selecting one or more topics from the list below.
Alexander GC, Sehgal AR. Barriers to Cadaveric Renal Transplantation Among Blacks, Women, and the Poor. JAMA. 1998;280(13):1148–1152. doi:10.1001/jama.280.13.1148
Context.— Cadaveric renal transplantation rates differ greatly by race, sex, and
income. Previous efforts to lessen these differences have focused on the transplant
waiting list. However, the transplantation process involves a series of steps
related to medical suitability, interest in transplantation, pretransplant
workup, and movement up a waiting list to eventual transplantation.
Objective.— To determine the relative importance of each step in explaining differences
in cadaveric renal transplantation rates.
Design.— Prospective cohort study.
Setting and Patients.— A total of 7125 patients beginning long-term dialysis between January
1993 and December 1996 in Indiana, Kentucky, and Ohio.
Main Outcome Measures.— Completion of 4 separate steps during each patient-year of follow-up:
(A) being medically suitable and possibly interested in transplantation; (B)
being definitely interested in transplantation; (C) completing the pretransplant
workup; and (D) moving up a waiting list and receiving a transplant.
Results.— Compared with whites, blacks were less likely to complete steps B (odds
ratio [OR], 0.68; 95% confidence interval [CI], 0.61-0.76), C (OR, 0.56; 95%
CI, 0.48-0.65), and D (OR, 0.50; 95% CI, 0.40-0.62) after adjustment for age,
sex, cause of renal failure, years receiving dialysis, and median income of
patient ZIP code. Compared with men, women were less likely to complete each
of the 4 steps, with ORs of 0.90, 0.89, 0.80, and 0.82, respectively. Poor
individuals were less likely than wealthy individuals to complete steps A,
B, and C, with ORs of 0.67, 0.78, and 0.77, respectively.
Conclusions.— Barriers at several steps are responsible for sociodemographic differences
in access to cadaveric renal transplantation. Efforts to allocate kidneys
equitably must address each step of the transplant process.
COMPARED WITH long-term dialysis, cadaveric renal transplantation generally
offers a longer life span, a better quality of life, and lower health care
the scarcity of donor organs means that only a small fraction of patients
receive transplants.1 Despite Medicare financing
of kidney transplantation, blacks, women, and poor individuals are less likely
to receive transplants than whites, men, and wealthy individuals.4,5
The transplantation process involves a series of steps related to medical
suitability, interest in transplantation, pretransplant workup, and movement
up the waiting list to eventual transplantation. Some of these steps have
been examined individually. For example, blacks may have less interest in
transplantation than whites.6 Blacks also move
up the waiting list at a slower rate than whites. Movement on the waiting
list has been studied extensively and appears to reflect both biological factors
(eg, HLA-based tissue typing) and nonbiological factors (eg, transplant center
characteristics).7 Other steps, such as the
pretransplant workup, have not been examined closely. Moreover, the relative
importance of each step in explaining sociodemographic differences in transplantation
rates is unknown. Knowing which steps are most important may help target interventions
to allocate kidneys more equitably.
Eligible subjects were all patients aged 18 to 65 years beginning long-term
dialysis between January 1, 1993, and December 31, 1996, in Indiana, Kentucky,
and Ohio. We excluded patients with (1) human immunodeficiency virus or cancer
as the cause of renal failure, (2) a cadaveric renal transplantation prior
to 1993, or (3) a renal transplant from a living relative at any time. To
simplify the examination of racial differences, we also excluded the 2% of
patients who were neither black nor white. Patients were followed up until
transplant, death, or December 31, 1996.
Dialysis providers are required to discuss treatment options with patients
at least annually, to record the results of this discussion on a long-term
program form signed by the patient, and to report this information to The
Renal Network Inc, Indianapolis, Ind, a regional agency that monitors the
care of patients with renal failure. Each patient's transplant status is classified
using 1 of 5 possible codes: not a transplant candidate (ie, not medically
suitable or not interested in transplant), medically suitable but undecided,
pretransplant workup in progress, on waiting list, and transplant received.
The Renal Network's patient database is updated throughout the year as dialysis
providers submit long-term program forms. However, archived data are maintained
only for the end of each calendar year. We obtained these year-end status
codes as well as each patient's race, sex, ZIP code, age, cause of renal failure,
dialysis start date, and date of any transplant from The Renal Network. We
determined the median income of each patient's ZIP code from census data.
We used the year-end transplant status codes to examine specific steps
in the transplant process. The last 3 status codes (pretransplant workup in
progress, on waiting list, and transplant received) represent completion of
a preceding step in the transplant process (Figure 1). For example, a status code of "pretransplant workup in
progress" indicates the patient is interested in transplantation (step B complete).
Similarly, a status code of "on waiting list" indicates that the pretransplant
workup (step C) has been completed, whereas a status code of "transplant received"
means the patient has moved up the waiting list and received a transplant
(step D complete). The first 2 status codes (not a transplant candidate, medically
suitable but undecided) are less distinct because they relate to both medical
suitability and interest in transplantation. For example, the status code
"not a transplant candidate" applies both to patients who are not medically
suitable and to patients who are definitely not interested. Despite this limitation,
each succeeding transplant status is closer to eventual transplantation. For
example, patients with a status code of "undecided" are medically suitable
and possibly interested in transplantation (step A complete) and are more
likely to eventually receive a transplant than patients categorized as "not
a transplant candidate."
Because of the progression represented by the transplant status codes,
each status indicates completion of specific steps in the transplantation
process. By contrast, failure to move beyond a particular status or moving
backward indicates a corresponding step has not been completed. For example,
a patient who remains on the waiting list for an extended period or who goes
from the waiting list to "not a transplant candidate" has not completed the
final step of moving up the waiting list to transplantation.
We examined patients during each year of follow-up. Each step (A through
D) was categorized as completed, not completed, or not applicable for a particular
patient year. Table 1 illustrates
this method for 3 hypothetical patients. Patient X began dialysis in 1994
and had a transplant status in December 1994 of "undecided." For this time
interval, step A is categorized as completed because the patient is medically
suitable and possibly interested in transplantation. Completion of this step
makes it possible for step B to be completed in the same year. However, the
patient is undecided about transplantation and step B is categorized as not
completed. Failure to complete step B makes it impossible for steps C and
D to be completed in the same year. As a result, steps C and D are categorized
as not applicable for that particular year. Patient X's transplant status
in December 1995 was still "undecided." For this year, step A is categorized
as not applicable because the patient has already completed this step. As
before, completion of step A makes it possible for step B to be completed,
but this did not occur. As a result, step B is categorized as not completed
and steps C and D as not applicable. Patient X received a transplant in 1996
and had a transplant status in December 1996 of "transplant received." For
this year, step A was categorized as not applicable and steps B, C, and D
were categorized as completed.
Three points about this method are worth noting. First, a subject may
be eligible to complete several steps in the same year as indicated by patient
X's completion of steps B, C, and D from December 1995 through December 1996.
Second, a patient who dies during a particular year will not have a year-end
transplant status code. As a result, that year cannot contribute to the analyses,
although earlier years will contribute as illustrated by patients Y and Z
in Table 1. Third, it is possible
that a patient will receive a transplant, reject the organ, and return to
dialysis in the same year. Because our goal was to examine access to transplantation,
we considered the year-end status code of such patients as "transplant received"
for our analyses.
Using patient-years as the unit of analysis, we calculated the proportion
of patients completing each step. Only steps categorized as completed or not
completed were used for these analyses. For example, in Table 1, a total of 4 patient-years provided information about step
A. In 3 (75%) of these, the first step was successfully completed. We used
either the χ2 test or the t test to
examine the univariate relationship between completion of each step and patient
demographic and medical variables (race, sex, ZIP code income, age, cause
of renal failure, and years on dialysis). Finally, we used logistic regression
analysis to examine the multivariate relationship between successful completion
of each step and patient demographic and medical variables.
Because subjects generally had more than 1 year of follow-up, it may
be argued that patient-years are not independent observations. To address
this, we performed separate hierarchical logistic regressions accounting for
the nesting of years within patients.8 These
analyses yielded virtually identical findings and are not presented herein.
Of the 9860 patients who met the eligibility criteria, 914 had incomplete
demographic data, 685 had no year-end transplant status code because they
died or were lost to follow-up within the first calendar year, and 1136 had
missing transplant status codes. The remaining 7125 patients formed our study
sample and are described in Table 2.
Compared with study subjects, the 1136 patients with missing transplant status
codes were more likely to be male (58% vs 55%, P
= .04) but did not otherwise differ from the study subjects by age, race,
cause of renal failure, or ZIP code income.
A total of 882 cadaveric transplantations occurred among all subjects
for a transplantation rate of 9.0 per 100 patient-years of follow-up. Transplantation
rates among whites and blacks were 11.6 and 5.1, respectively, and rates among
men and women were 10.4 and 7.3, respectively. Transplantation rates for patients
with ZIP code incomes of less than $11000, $11000 to $14000, and more than
$14000 were 6.7, 9.4, and 12.4, respectively. As mentioned herein, these calculations
exclude 18 transplantations that occurred in the same year as the patient's
Using patient-years as the unit of analysis, the overall completion
rate was approximately 50% for steps A and D and about 35% for steps B and
C (Table 3). On univariate analysis,
blacks, women, and poor individuals were less likely to complete most steps
in the transplant process compared with whites, men, and wealthy individuals.
Racial differences were most pronounced at steps B, C, and D. For example,
step B (definite interest in transplantation) was completed in 776 (30%) of
2555 eligible patient-years by black subjects compared with 1667 (39%) of
4290 eligible patient-years by white subjects (P<.001).
Modest sex differences existed at each step. Income differences were most
pronounced at steps A, B, and C. For example, step C (pretransplant workup)
was completed in 387 (31%) of 1268 patient-years by individuals with incomes
of less than $11000 compared with 381 (41%) of 927 patient-years by individuals
with incomes of more than $14000 (P<.001). Age,
diabetes, and length of time on dialysis were also associated with completion
of several steps.
After multivariate adjustment (Table
4) for sex, ZIP code income, age, cause of renal failure, and years
on dialysis, blacks were less likely than whites to complete steps B (odds
ratio [OR], 0.68), C (OR, 0.56), and D (OR, 0.50). Women were less likely
than men to complete steps A (OR, 0.90), B (OR, 0.89), C (OR, 0.80), and D
(OR, 0.82). Poor individuals were less likely to complete steps A (OR, 0.67),
B (OR, 0.78), and C (OR, 0.77). Age, diabetes, and years of dialysis were
also independently associated with completion of several steps.
Patients unable to complete a step generally remained at a specific
transplant status rather than moving backward. For example, as noted in Table 3, there were 2420 patient-years
(3739 − 1319) during which step C (pretransplant workup) was not completed.
Of these patient-years, 2342 (97%) had a year-end status of "pretransplant
workup in progress" whereas only 78 (3%) were categorized as not a transplant
candidate or as "undecided."
We also examined our data for possible interactions among race, sex,
and income. A significant interaction between race and income (P = .01) occurred at step C; the effect of low income as a barrier
was more pronounced among whites than blacks. A significant interaction between
sex and income (P = .002) occurred at step B; the
effect of low income as a barrier was more pronounced among men than women.
Our results confirm the existence of substantial differences in access
to cadaveric renal transplantation by race, sex, and income. More important,
this is the first study to determine the relative importance of 4 key steps
in the transplantation process: (A) being medically suitable and possibly
interested in transplantation, (B) being definitely interested in transplantation,
(C) completing the pretransplant workup, and (D) moving up a waiting list
and receiving a transplant. By focusing on subjects eligible to complete each
step annually over a 4-year interval, we obtained detailed information about
the individual effect of each step. We found that all 4 steps play an important
role in explaining sociodemographic differences in transplantation rates.
Steps B through D are the most important impediments for blacks, all 4 steps
are impediments for women, and steps A through C are the most important impediments
for poor individuals.
Our large, representative patient sample makes these findings especially
noteworthy. We included all new patients in 3 states that collectively represent
8% of American dialysis patients. The proportion of blacks, women, and patients
with diabetes in our sample is comparable with patients nationally. For example,
44% of national patients aged 20 to 64 years have diabetes compared with 49%
of our subjects. In addition, the cadaveric renal transplantation rates we
observed are comparable with national rates for all patients and for race
and sex subgroups.1
Our findings indicate that efforts to allocate kidneys more equitably
must target each step in the transplantation process. Recent work has demonstrated
sociodemographic differences in access to the waiting list (equivalent to
the combined effect of steps A, B, and C in our analysis).6,9-11
However, interventions have focused largely on patients already on the waiting
list. For example, modifying the waiting list matching algorithm by adding
points for waiting time has increased transplantations among blacks.12 Other interventions have sought to increase organ
donation by minorities in an effort to increase the availability of histocompatible
kidneys and speed movement up the waiting list by minorities.13
Our findings also highlight the role of the pretransplant workup (step
C), which acts as a barrier among all 3 subgroups, blacks, women, and the
poor. The importance of this step has not previously been appreciated. Tasks
performed at this step may include referral to transplant surgeons, evaluation
and treatment of medical conditions, and laboratory studies such as tissue
typing. We found that only 3% of patients not completing this step in a given
year are categorized as "not a transplant candidate" or "undecided." Thus,
failure to complete this step is generally not due to medical unsuitability
or lack of interest in transplantation. Rather, it appears that many patients
remain at this step for an extended period.
Our study cannot determine why specific steps serve as barriers among
blacks, women, and the poor. Other investigators have speculated about a variety
of patient and provider factors that may be responsible. Possible patient
factors include biological and medical variables, lack of knowledge about
transplantation, and concerns about surgery, adverse effects of medication,
and health care costs.4-7,9,14-16
Possible provider factors include subconscious bias and financial disincentives.4,5,17-19
Transplant center size and proximity, as well as regional variations in matching
algorithms, may also play a role.4,5,7,11,20
Understanding how these factors affect specific steps in the transplant process
may help identify interventions to lessen sociodemographic differences in
kidney transplantation rates.
Several limitations must be considered in interpreting our findings.
First, as discussed herein, the status code "not a transplant candidate" applies
both to patients who are medically unsuitable and to patients who are uninterested
in transplantation. This makes it difficult to separate the effect of suitability
from the effect of interest, although the status code "undecided" does provide
additional information about interest in transplantation. Second, the transplant
status codes are archived only at the end of each year. Thus, a patient who
dies during a particular year will not contribute that year's experience to
the analyses. However, we found that only 18 transplantations occurred in
the same year as patient deaths compared with 882 transplantations among patients
who survived at least to the end of the transplant year. This suggests that
excluding death years from our analyses probably has little effect on our
conclusions. Finally, we examined patients in a 3-state area, and it is possible
that other regions have different barriers to cadaveric renal transplantation
among blacks, women, and poor individuals.
These limitations point out the need for a national database with uniform
transplant status codes that correspond directly to distinct steps in the
transplantation process. Currently, different coding systems exist among the
18 regional renal failure networks, and there is no national repository. Such
a resource would be invaluable for studying regional differences in barriers
to transplantation, identifying possible interventions, and determining the
impact of interventions. Even in the absence of a national effort, we encourage
individual transplant and dialysis providers to examine steps in the transplantation
process among their patients to identify areas for improvement. This does
not mean forcing all patients through each step in the transplantation process.
Rather, transplant and dialysis providers need to ensure that transplant candidates
are identified equitably and then assisted through the transplantation process
as expeditiously as possible. Our method for examining sequential steps may
also be applicable to other areas such as access to liver transplantation
or cancer treatment.
A generous public, motivated primarily by altruism, donates organs of
loved ones expecting they will be distributed equitably to needy patients.21 Efforts to fulfill this expectation must be pursued
with the same vigor with which we seek new immunosuppressive medications or
improved surgical techniques.
Create a personal account or sign in to: