Abstract
Racial/ethnic and gender disparities in living donor kidney transplantation are large and persistent but incompletely explained. One previously unexplored potential contributor to these disparities is differential willingness to donate to recipients in specific relationships such as children, parents, and friends. We collected and analyzed data from an online sample featuring an experimental vignette in which respondents were asked to rate their willingness to donate to a randomly chosen member of their family or social network. Results show very large differences in respondents’ willingness to donate to recipients with different relationships to them, favoring children, spouses/partners, siblings, and parents, and disfavoring friends, aunts/uncles, and coworkers. Evidence suggesting an interactive effect between relationship, respondent race/ethnicity, respondent or recipient gender, was limited to a few cases. At the p<0.05 level, the parent-recipient gender interaction was statistically significant, favoring mothers over fathers, as was other/multiracial respondents’ greater willingness to donate to friends compared to Whites. Additionally, other interactions were significant at the p<0.10 level, such as Hispanics’ and women’s higher willingness to donate to parents compared to Whites and men respectively, women’s lower willingness to donate to friends compared to men, and Blacks’ greater willingness to donate to coworkers than Whites. We also examined differences by age and found that older respondents were less willing to donate to recipients other than their parents. Together these results suggest that differential willingness to donate by relationship group may be a moderately important factor in understanding racial/ethnic and gender disparities in living donor kidney transplantation.
Keywords: Living kidney donation, social networks, relationships, race/ethnicity, gender
1. Introduction
Living donor kidney transplantation (LDKT) is frequently the optimal therapy for patients with end-stage kidney disease (ESKD), but it is significantly less common than dialysis or deceased donor kidney transplantation (DDKT).1 Furthermore, despite numerous amelioration efforts large racial/ethnic and moderate gender disparities in LDKT donation and receipt persist and in the case of race/ethnicity have continued to grow.2 One factor that may contribute to these disparities, which has heretofore received scant attention, is how willingness to donate varies by relationship to the recipient, race/ethnicity, gender, and age. Although the diversity of donor-recipient relationships has increased over time,1,3 LDKT donor-recipient relationships remain concentrated in a handful of relationships,4 including siblings, spouses/partners, children, and friends, while extended kin remain comparatively underutilized.
However, an underexplored dynamic of these trends is that trajectories in the frequency of these donor-recipient relationships vary meaningfully by transplant candidate race/ethnicity and gender.5–13 Despite experiencing an almost 30% increase in ESKD prevalence from 2000–2020 and having the highest prevalence of ESKD among all racial/ethnic groups, Black individuals are less likely to receive kidney transplants, particularly LDKTs.13,14 Research suggests that Black ESKD patients experience several barriers to LDKT.6,15 For example, one study showed Black transplant candidates perceived similar rates of access to healthy and willing donors compared to White candidates.10 Although the study found no significant Black-White differences in candidates’ perceptions of potential donor willingness, results may differ when willingness is examined from the perspective of the donors. Racial-ethnic differences in willingness to donate may arise from medical mistrust, fears of discrimination, and fears of economic and health risks associated with living donation.16
Furthermore, women are more likely to be living kidney donors compared to men, and both men and women are more likely to donate to men.8,17–20 However, men’s higher ESKD prevalence only partially explains their higher share of LDKT recipients.14,21,22 Previous research suggests the disproportionate number of female-to-male donations among living related, unrelated non-spousal, and spousal pairs may be driving the gender imbalance in LDKT.18,22,23 Gendered differences in socioeconomic profiles (such as employment and insurance coverage rates) are thought to be one contributing factor.18,24–27 Biological factors such as pregnancy-induced HLA presensitization and higher prevalence of comorbidities and substance use among men in the United States may also contribute to men’s low donation rate.19,22,28–31 Finally, research has found that women may be more willing to donate than men and are more likely to volunteer to donate.24,28,32–35
This article explores the potential role of racial/ethnic and gender differences in willingness to donate to members of their family and social network by their relationship to the recipient. It leverages original data employing randomly-assigned experimental vignettes nested within a web survey, in which participants were asked to rate and explain their level of willingness to be evaluated for donation in a hypothetical scenario in which one of their social network members needed a kidney transplant. Though they are hypothetical, vignette experiments are useful because they can shed light on the mechanisms that produce disparities. In this study, the relationship of the respondent to the hypothetical transplant candidate was randomized conditional on their distribution of reported social network relationships. Results find moderate evidence of racial/ethnic and gender interactions with potential recipient relationship in determining willingness in some, but not most, cases. Similarly, we find that older respondents are less willing to donate than younger respondents, unless the recipient is their parent.
2. Objective
Our objective is to answer the following research questions:
How does willingness to donate vary by relationship, respondent race/ethnicity, respondent gender, and potential recipient gender?
Do these factors interact to predict respondent willingness to donate?
3. Materials and Method
3.1. Data
Data for this study were collected by the authors in three waves using a convenience sample of online survey participants. The first wave was collected in 2016, and the 2nd and 3rd waves were collected in 2018. All survey participants were recruited on Amazon’s Mechanical Turk (mTurk) platform in which participants can be affordably hired to complete discrete tasks, such as taking a survey. As such, this sample does not directly represent any identifiable population. However, previous methodological research investigating potential differences between mTurk samples and population-based samples in the US have found that results of survey experiments and nonexperimental association studies are often quite similar once demographic compositional differences have been statistically accounted for.36–38
3.1.1. Wave 1 Data Collection
3.1.1.1. Overview
In 2016, we collected our first wave of data, implementing an experimental vignette study to test how different factors influence survey respondents’ reported willingness to be evaluated as living donors. We recruited 2,249 respondents on mTurk to participate in this study. A $1 participation incentive was offered.
Without appropriate quality control, web surveys can suffer from a preponderance of lowquality responses.39 Accordingly, restrictions were imposed to ensure that respondents were based in the U.S., could not participate more than once, and had completed 1,000 previous human intelligence tasks (HITs, or work assignments like completing surveys, finding online information, etc.) on mTurk with a 95% approval rate. Additionally, following the vignette, respondents were asked to complete an attention check question requiring them to answer a factual question about living kidney donation whose answer was previously provided in the vignette. Of the 2,249 respondents, 80 either did not correctly answer the attention check item or had been previously rejected in other mTurk studies conducted by the research team, leading to a final sample size of 2,169 (96.4% of cases were retained). All participants provided informed consent and confirmation of being 18 years or older of age.
3.1.1.2. Conditionally Experimental Vignette Factor: Potential Recipient Relationship
Prior to the vignette, respondents were asked to describe several demographic characteristics, including the number of living gender-specific relatives of different types, in response to this prompt: “For each of the following types of relative, please indicate the number of living relatives you have of that type. You may also include step-family, adopted family, and family by marriage. If you are unsure, submit your first estimate and move on.” The respondents were then prompted to enter their number of fathers, mothers, brothers, sisters, sons, daughters, uncles, and aunts.
Next, respondents were presented with a vignette in which they were asked to rate their willingness to be tested as a living donor for the assigned recipient. Respondents were assigned to one of six relationship types, choosing among relationships they reported having: parent, sibling, child, aunt/uncle, friend, and coworker. Siblings were only included in later data collection batches within wave 1. The presence of family members in the respondent’s network was determined using the kin availability measures just described; co-workers were inferred from non-mTurk employment status; and the presence of 1 or more friends was assumed.
3.1.1.3. Wave 1 Vignette
Respondents in wave 1 were asked to rate their willingness to be evaluated for living kidney donation to a member of their family or social network in response to a randomly-generated vignette describing the donation scenario. In addition to the relationship to the recipient, several other factors were varied when generating these vignettes, including how the respondent heard about their need for a living donor kidney transplant, the putative cause of their kidney failure, and how the effects of living kidney donation for both recipient and donor were described. Full details are presented in the supplement.
3.1.2. Waves 2–3 Data Collection
3.1.2.1. Differences versus Wave 1
Waves 2–3 of our data collection shared many commonalities with Wave 1, but also exhibited a few key differences. As described at greater length in the supplement, the vignettes we employed in these waves asked respondents to respond to a scenario in which the randomly chosen transplant candidate asked them directly to consider being evaluated for living kidney donation. Randomly-assigned variations in wave 2 included alternative framings of the risk/reward of living kidney donation, and whether respondents were asked to donate or be evaluated for donation. Randomly-assigned variations in wave 3 included varying details on opportunities for indirect cost reimbursement and the degree of emphasis on the medical/scientific basis of the information presented. The details of these scripts and randomly-assigned variations are provided in the supplement.
In addition to this key change, a few other changes were made compared to Wave 1. First, a few minor items in the previous survey were eliminated in order to minimize respondent burden, and respondent survey compensation was reduced proportionally, to $0.80. Second, the attention check item was changed to whether the respondent could correctly identify their relationship to the potential recipient in the script provided. Third, we changed the relationship options: we eliminated aunts/uncles and coworkers from the set of relationships tested and added spouses/partners.
3.1.2.2. Wave 2–3 Data Cleaning
Before data cleaning, 1,863 respondents participated in wave 2, and 1,758 participated in wave 3, for a total of 3,621. N=102 participants were rejected, 13 failed the attention check, 18 offered non-sensical explanations for their willingness ratings, and 22 were missing 1+ key variables, reducing the total sample size to N=3,466 (1,716 in wave 2, 1,750 in wave 3).
3.1.3. Variables Analyzed
All analyzed variables were measured identically in all three waves. To measure willingness to donate, after the vignette respondents were asked to indicate their willingness to donate in the presented scenario on a 1–9 scale, with 1 indicating least likely to donate and 9 most likely to donate. Specifically, the vignette was followed by the question: “After hearing this, how likely do you think it is that you would volunteer to be tested for compatibility as a living kidney donor for your [relationship type]?”
Aside from the experimental vignette items, all survey questions were based on existing items from governmental and academic surveys or previously validated measures used in medical transplantation research. For demographic variables of interest, respondents were asked to indicate their gender (male, female, or other), their race/ethnicity (Caucasian or White, Hispanic or Latino/a, Black or African-American, Asian or Pacific Islander, American Indian or Alaska Native, or Other; respondent could mark all that applied, and if more than one marked were recoded as Other/Multiracial), age (less than 20 years, options for 10 year age groups between the ages of 20–89, or 90+; this was recoded to ≤29 and ≥60 due to relatively few <20 and >70 respondents), as well as their education (high school, some college, four-year college, post-baccalaureate, or still in school). The survey also collected data on the respondents’ family structure, with a prompt for respondents to report the number of their living relatives and their relationship to them (parents, siblings, children, aunts/uncles). In addition to these relationships, friend and coworker relationship ties were included in the variable coded for recipient relationship to respondent in wave 1, and spouse/partner ties were included in waves 2–3. Recipient gender was also coded as a binary variable from the presented vignette; in the case of spouse/partner ties, the recipient is assumed to be a man if the respondent is a woman and vice versa.
3.2. Statistical Methods
Data from all three waves were pooled for this analysis, which proceeds in several steps. First, the distribution of key variables overall and by wave is described. Second, the two research questions are assessed descriptively. Mean willingness to be evaluated for living kidney donation by recipient relationship to the respondent was separately assessed in combination with respondent race/ethnicity, respondent gender, and recipient gender. Additionally, mean willingness to be evaluated for living kidney donation was assessed by recipient relationship to the respondent, respondent gender, and recipient gender, all together – this table is provided in the supplement. This last procedure does not include respondent race/ethnicity because the minimum cell size was too small for meaningful analysis.
In the multivariable analysis, ordinary least squares regression was used to model willingness to be evaluated for living kidney donation. For each of the seven relationships examined (Child, Spouse/Partner, Sibling, Parent, Friend, Aunt/Uncle, and Coworker), the dataset is subset to those who indicated they had one or more of this type of tie in their social network and who participated in a wave in which this relationship was included in the randomly-assigned vignettes.
Among those deemed to have 1+ tie of the relationship type in question, the effect of being conditionally-randomly-assigned the relationship was modeled dichotomously, =1 if this relationship was assigned and =0 if some other relationship was assigned. In this way, the relationship coefficients may be interpreted as the effect of having the relationship in question randomly selected among their eligible tie types, compared to having any other eligible tie type selected, among those with the tie type in question.
To adjust for the fact that the living kidney donation scenario presented differed substantially across waves (especially between wave 1 and waves 2–3), all models control for wave of data collection, using wave 1 as the reference category. To adjust for the fact that respondents who reported more tie types had a lower probability of each tie type being assigned (as well as differing in their social network structure), all models control for whether the respondent reported having 1+ of each other relationship type besides the relationship in question. Finally, all regression models specify recipient gender, respondent gender, respondent race/ethnicity, respondent age, and respondent education in the regression model.
To test the hypothesis that the effect of relationship differs based on race/ethnicity, gender, and age, interactive specifications are used, with the effect of each dichotomous relationship variable interacted with recipient gender, respondent gender, respondent race/ethnicity, and age, in four separate models. The coefficients and p-values for these regression terms are reported. Additionally, we report the results of Wald tests for the joint significance of the relationship-race/ethnicity and relationship-age interactive terms (since there are four individual interactive coefficients).
4. Results
4.1. Sample Characteristics
Table 1 describes the sample characteristics, overall and by data collection wave. Willingness to donate was significantly lower in wave 1 (5.97) than wave 2 (7.49) or 3 (7.72), which likely reflects a combination of cross-wave differences in scenario presentation, relationship options, and time. Overall, friends were the most commonly-assigned relationship (22.9%), followed by parent (21.3%), sibling (18.1%), and child (14.0%). Because they were not options in all waves, spouse/partner (10.2%), aunt/uncle (6.8%), and coworker (6.8%) were less common than the first four relationships.
Table 1:
Descriptive Statistics by Data Collection Wave
| Wave | ||||
|---|---|---|---|---|
| 1 | 2 | 3 | Total | |
| N | 2,169 | 1,716 | 1,750 | 5,635 |
| Willingness to Donate | 5.97 | 7.49 | 7.72 | 6.98 |
| Recipient Relationship to Respondent | ||||
| Child | 17.52 | 12.35 | 11.20 | 13.98 |
| Spouse | 0.00 | 16.38 | 16.86 | 10.22 |
| Sibling | 11.85 | 22.90 | 20.97 | 18.05 |
| Parent | 17.66 | 24.01 | 23.14 | 21.30 |
| Friend | 17.70 | 24.36 | 27.83 | 22.87 |
| Aunt/Uncle | 17.61 | 0.00 | 0.00 | 6.78 |
| Coworker | 17.66 | 0.00 | 0.00 | 6.80 |
| Recipient Gender | ||||
| Male | 49.93 | 48.89 | 46.40 | 48.52 |
| Female | 50.07 | 51.11 | 53.60 | 51.48 |
| Respondent Gender | ||||
| Male | 51.54 | 46.68 | 44.06 | 47.74 |
| Female | 48.46 | 53.32 | 55.94 | 52.26 |
| Respondent Age | ||||
| <29 | 33.06 | 28.73 | 28.51 | 30.33 |
| 30–39 | 34.62 | 38.34 | 37.43 | 36.63 |
| 40–49 | 16.32 | 17.37 | 17.49 | 17.00 |
| 50–59 | 10.83 | 9.85 | 10.80 | 10.52 |
| >=60 | 5.16 | 5.71 | 5.77 | 5.52 |
| Respondent Education | ||||
| <=High School | 10.88 | 10.90 | 10.23 | 10.68 |
| Some College | 32.23 | 36.66 | 34.91 | 34.41 |
| 4- year degree | 33.10 | 40.21 | 40.80 | 37.66 |
| Graduate degree | 10.47 | 12.24 | 14.06 | 12.12 |
| Currently in school | 13.32 | 0.00 | 0.00 | 5.13 |
| Respondent Race/Ethnicity | ||||
| White | 77.46 | 76.46 | 76.11 | 76.73 |
| Black | 6.59 | 6.29 | 6.97 | 6.62 |
| Hispanic | 3.78 | 4.55 | 4.29 | 4.17 |
| Asian | 6.13 | 6.70 | 6.17 | 6.32 |
| Oth/Multi | 6.04 | 6.00 | 6.46 | 6.16 |
| Has Child | ||||
| No | 56.15 | 47.55 | 48.97 | 51.30 |
| Yes | 43.85 | 52.45 | 51.03 | 48.70 |
| Has Spouse | ||||
| No | 100.00 | 6.93 | 9.89 | 43.67 |
| Yes | 0.00 | 93.07 | 90.11 | 56.33 |
| Has Sibling | ||||
| No | 13.65 | 9.79 | 10.63 | 11.54 |
| Yes | 86.35 | 90.21 | 89.37 | 88.46 |
| Has Parent | ||||
| No | 10.88 | 9.50 | 10.57 | 10.36 |
| Yes | 89.12 | 90.50 | 89.43 | 89.64 |
| Has Friend | ||||
| No | 0.00 | 6.93 | 6.86 | 4.24 |
| Yes | 100.00 | 93.07 | 93.14 | 95.76 |
| Has Aunt/Uncle | ||||
| No | 12.13 | 100.00 | 100.00 | 66.18 |
| Yes | 87.87 | 0.00 | 0.00 | 33.82 |
| Is Employed | ||||
| No | 20.75 | 100.00 | 100.00 | 69.49 |
| Yes | 79.25 | 0.00 | 0.00 | 30.51 |
The respondent age distribution skewed young, with a median and modal age of 30–39 (36.6%), with ≤29 (30.3%) the 2nd-most common response. Respondents ≥60 were rare (5.5%). Respondents were highly educated, with a median educational attainment of a 4-year college degree. The sample over-represents non-Hispanic White (76.5%) and Asian (6.6%) compared to their shares of the population, while non-Hispanic Black (6.9%) and Hispanic (4.1%) respondents are underrepresented.
4.2. Descriptive Analysis
Table 2 examines mean willingness to donate as a joint function of respondent relationship to the recipient and respondent race/ethnicity. Overall, respondents are most willing to donate when the recipient is their spouse (mean = 8.4), followed by their children (7.8), sibling (7.5), parent (7.4), friend (6.3), aunt/uncle (5.7), and coworker (4.2). Within respondent race/ethnicity groups, this ranking is highly similar, with some minor differences in ordering and relative differences. For instance, Black respondents assign their spouses a lower mean rating (7.5) than White respondents (8.4), but similar ratings for their children (7.8 for Whites, 7.7 for Blacks). Black respondents are also more willing to consider donating to a coworker (5.3) than any other racial/ethnic group. Hispanic respondents on average assign a higher willingness to siblings (8.0) and parents (8.0) than children (7.5), unlike Black and White respondents. Asian respondents’ averages show a very similar pattern as Hispanic respondents, with siblings (7.8) and parents (7.3) receiving higher average willingness than children (6.7).
Table 2:
Mean Willingness to Donate by Respondent Race/Ethnicity & Recipient Relationship to Respondent
| Respondent Race/Ethnicity | ||||||
|---|---|---|---|---|---|---|
| Recipient Relationship to Respondent | White (N) | Black (N) | Hispanic (N) | Asian (N) | Other/Multi (N) | Total (N) |
| Child | 7.84 | 7.68 | 7.53 | 6.66* | 7.75 | 7.77 |
| (635) | (63) | (15) | (35) | (40) | (788) | |
| Spouse | 8.42 | 7.52* | 8.24 | 8.10+ | 8.67+ | 8.37 |
| (470) | (27) | (17) | (29) | (33) | (576) | |
| Sibling | 7.47+ | 7.15 | 8.02 | 7.82+ | 7.30 | 7.49 |
| (780) | (68) | (46) | (73) | (50) | (1,017) | |
| Parent | 7.41+ | 6.97 | 7.97 | 7.26 | 7.48 | 7.41 |
| (890) | (77) | (66) | (86) | (81) | (1,200) | |
| Friend | 6.25+ | 5.94+ | 6.00 | 6.14+ | 7.12 | 6.26 |
| (965) | (94) | (54) | (93) | (83) | (1,289) | |
| Aunt/Uncle | 5.74+ | 5.94+ | 5.25+ | 5.59 | 5.19+ | 5.67 |
| (292) | (17) | (20) | (22) | (31) | (382) | |
| Coworker | 4.14+ | 5.30+* | 3.82+ | 4.39 | 4.38+ | 4.23 |
| (292) | (27) | (17) | (18) | (29) | (383) | |
| Total (N) | 7.00 | 6.73 | 6.99 | 6.84 | 7.05 | 6.98 |
| (4,324) | (373) | (235) | (356) | (347) | (5,635) | |
: Column t-test significant at P<0.05;
: Row t-test significant at P<0.05;
Child is the reference for within-column t-tests;
White is the reference for withinrow t-tests
Table 3 examines these same patterns separately by respondent and recipient gender. Women report higher average willingness to donate (7.1) than men (6.8). By relationship, men’s and women’s average willingness are similar, but women assign a higher average willingness to their children (8.0 for women, 7.4 for men), aunts/uncles (5.9 for women, 5.4 for men), and coworkers (4.5 for women, 4.0 for men). Their implied relationship rankings are similar except that, unlike men, women report higher average willingness for siblings (7.6) than parents (7.4).
Table 3:
Mean Willingness to Donate by Respondent/Recipient Gender and Recipient Relationship to Respondent
| Respondent Gender | Recipient Gender | |||||
|---|---|---|---|---|---|---|
| Recipient Relationship to Respondent | Male (N) | Female (N) | Total (N) | Male (N) | Female (N) | Total (N) |
| Child | 7.44 | 8.02* | 7.77 | 7.73 | 7.80 | 7.77 |
| (341) | (477) | (788) | (391) | (397) | (788) | |
| Spouse | 8.34+ | 8.40+ | 8.37 | 8.40+ | 8.34+ | 8.37 |
| (242) | (334) | (576) | (334) | (242) | (576) | |
| Sibling | 7.35 | 7.61+ | 7.49 | 7.40+ | 7.58 | 7.49 |
| (472) | (545) | (1,017) | (501) | (516) | (1,017) | |
| Parent | 7.41 | 7.40+ | 7.41 | 7.14+ | 7.64* | 7.41 |
| (631) | (569) | (1,200) | (566) | (634) | (1,200) | |
| Friend | 6.22+ | 6.30+ | 6.26 | 6.15+ | 6.35+ | 6.26 |
| (618) | (671) | (1,289) | (559) | (730) | (1,289) | |
| Aunt/Uncle | 5.44+ | 5.90+ | 5.67 | 5.64+ | 5.69+ | 5.67 |
| (169) | (186) | (382) | (191) | (191) | (382) | |
| Coworker | 3.99+ | 4.48+ | 4.23 | 3.97+ | 4.50+ | 4.23 |
| (190) | (193) | (383) | (192) | (191) | (383) | |
| Total (N) | 6.83 | 7.11 | 6.98 | 6.90 | 7.05 | 6.98 |
| (2,690) | (2,945) | (5,635) | (2,734) | (2,901) | (5,635) | |
: Column t-test significant at P<0.05;
: Row t-test significant at P<0.05;
Child is the reference for within-column t-tests;
Male is the reference for withinrow t-tests
Comparing ratings by recipient gender, average willingness is highly similar across recipient gender and within relationship types, with two exceptions: on average, respondents report higher willingness for women than men among parent recipients (7.6 for mothers, 7.1 for fathers) and coworker recipients (4.5 for women coworkers, 4.0 for men coworkers).
4.3. Multivariable Regression Analyses
Table 4a, 4b, 4c, and 4d reports the results of our multivariable regression analyses, conducted separately by relationship. These results provide strong evidence that respondents are especially willing to donate to their children, spouses/partners, siblings, and parents, and are less willing to donate to their friends, aunts/uncles, and coworkers, net of demographic covariates.
Table 4a:
Regression Analysis
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Relationship | Coefficient | B | p | B | p | B | p | B | p | B | p |
| Child (N=2744) | 1.41 | 0.000 | 1.40 | 0.000 | 1.43 | 0.000 | 1.41 | 0.000 | 1.06 | 0.000 | |
| Black Resp. | −0.36 | 0.062 | −0.36 | 0.116 | −0.36 | 0.062 | −0.36 | 0.062 | −0.35 | 0.066 | |
| Hispanic Resp. | −0.19 | 0.455 | −0.14 | 0.604 | −0.19 | 0.455 | −0.19 | 0.455 | −0.20 | 0.424 | |
| Asian Resp. | −0.47 | 0.050 | −0.45 | 0.089 | −0.47 | 0.050 | −0.47 | 0.050 | −0.49 | 0.042 | |
| Other/Multiracial Resp. | −0.03 | 0.872 | −0.12 | 0.612 | −0.03 | 0.875 | −0.03 | 0.873 | −0.04 | 0.855 | |
| Female Resp. | 0.08 | 0.357 | 0.08 | 0.348 | 0.08 | 0.355 | 0.08 | 0.414 | 0.08 | 0.367 | |
| Female Recip. | 0.06 | 0.549 | 0.05 | 0.568 | 0.06 | 0.553 | 0.06 | 0.549 | 0.05 | 0.584 | |
| Age ≤29 | 0.04 | 0.767 | 0.04 | 0.778 | 0.04 | 0.766 | 0.04 | 0.767 | −0.04 | 0.804 | |
| Age 40–49 | −0.05 | 0.656 | −0.05 | 0.652 | −0.05 | 0.658 | −0.05 | 0.656 | −0.17 | 0.189 | |
| Age 50–59 | −0.08 | 0.570 | −0.08 | 0.576 | −0.08 | 0.570 | −0.08 | 0.571 | −0.23 | 0.141 | |
| Age ≥60 | −0.45 | 0.021 | −0.45 | 0.021 | −0.45 | 0.021 | −0.45 | 0.021 | −0.61 | 0.005 | |
| Celationship x Black Resp. | 0.00 | 0.992 | |||||||||
| Celationship x Hispanic Resp. | −0.40 | 0.603 | |||||||||
| Celationship x Asian Resp. | −0.09 | 0.888 | |||||||||
| Relationship x Other/Multi. Resp. | 0.43 | 0.410 | |||||||||
| Relationship x Female Resp. | −0.03 | 0.897 | |||||||||
| Relationship x Female Recip. | 0.00 | 0.994 | |||||||||
| Relationship x Age ≤29 | 0.40 | 0.300 | |||||||||
| Relationship x Age 40–49 | 0.57 | 0.046 | |||||||||
| Relationship x Age 50–59 | 0.71 | 0.029 | |||||||||
| Relationship x Age ≥60 | 0.72 | 0.066 | |||||||||
| Spouse/Partner (N=3174) | 0.98 | 0.000 | 1.02 | 0.000 | 1.11 | 0.000 | 0.88 | 0.000 | 0.90 | 0.000 | |
| Black Resp. | −0.33 | 0.043 | −0.26 | 0.126 | −0.33 | 0.043 | −0.33 | 0.043 | −0.32 | 0.046 | |
| Hispanic Resp. | 0.02 | 0.911 | 0.07 | 0.753 | 0.02 | 0.907 | 0.02 | 0.907 | 0.03 | 0.891 | |
| Asian Resp. | −0.21 | 0.212 | −0.21 | 0.249 | −0.21 | 0.217 | −0.21 | 0.217 | −0.21 | 0.214 | |
| Other/Multiracial Resp. | 0.22 | 0.188 | 0.20 | 0.253 | 0.22 | 0.188 | 0.22 | 0.188 | 0.22 | 0.187 | |
| Female Resp. | 0.16 | 0.042 | 0.16 | 0.042 | 0.13 | 0.136 | 0.13 | 0.136 | 0.16 | 0.043 | |
| Female Recip. | 0.19 | 0.017 | 0.19 | 0.017 | 0.23 | 0.010 | 0.23 | 0.010 | 0.20 | 0.016 | |
| Age ≤29 | −0.02 | 0.844 | −0.02 | 0.830 | −0.02 | 0.861 | −0.02 | 0.861 | −0.04 | 0.735 | |
| Age 40–49 | −0.15 | 0.177 | −0.15 | 0.178 | −0.15 | 0.181 | −0.15 | 0.181 | −0.14 | 0.249 | |
| Age 50–59 | −0.29 | 0.044 | −0.30 | 0.042 | −0.29 | 0.046 | −0.29 | 0.046 | −0.33 | 0.036 | |
| Age ≥60 | −0.69 | 0.001 | −0.69 | 0.000 | −0.69 | 0.000 | −0.69 | 0.000 | −0.77 | 0.000 | |
| Relationship x Black Resp. | −0.59 | 0.256 | |||||||||
| Relationship x Hispanic Resp. | −0.36 | 0.554 | |||||||||
| Relationship x Asian Resp. | 0.00 | 0.993 | |||||||||
| Relationship x Other/Multi. Resp. | 0.11 | 0.811 | |||||||||
| Relationship x Female Resp. | −0.23 | 0.337 | |||||||||
| Relationship x Female Recip. | 0.23 | 0.337 | |||||||||
| Relationship x Age ≤29 | 0.12 | 0.676 | |||||||||
| Relationship x Age 40–49 | −0.06 | 0.844 | |||||||||
| Relationship x Age 50–59 | 0.25 | 0.502 | |||||||||
| Relationship x Age ≥60 | 0.69 | 0.176 | |||||||||
Table 4b:
Regression Analysis
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Relationship | Coefficient | B | p | B | p | B | p | B | p | B | p |
| Sibling (N=4985) | 0.42 | 0.000 | 0.40 | 0.000 | 0.38 | 0.003 | 0.40 | 0.002 | 0.25 | 0.091 | |
| Black Resp. | −0.24 | 0.091 | −0.19 | 0.221 | −0.24 | 0.092 | −0.24 | 0.090 | −0.24 | 0.095 | |
| Hispanic Resp. | −0.09 | 0.619 | −0.19 | 0.324 | −0.09 | 0.621 | −0.09 | 0.621 | −0.07 | 0.687 | |
| Asian Resp. | −0.13 | 0.381 | −0.25 | 0.128 | −0.13 | 0.383 | −0.13 | 0.381 | −0.13 | 0.383 | |
| Other/Multiracial Resp. | 0.09 | 0.559 | 0.16 | 0.317 | 0.09 | 0.557 | 0.09 | 0.556 | 0.09 | 0.527 | |
| Female Resp. | 0.14 | 0.048 | 0.14 | 0.048 | 0.14 | 0.049 | 0.13 | 0.098 | 0.14 | 0.049 | |
| Female Recip. | 0.09 | 0.238 | 0.08 | 0.245 | 0.07 | 0.372 | 0.09 | 0.241 | 0.08 | 0.264 | |
| Age ≤29 | 0.02 | 0.819 | 0.02 | 0.785 | 0.02 | 0.825 | 0.02 | 0.819 | −0.06 | 0.548 | |
| Age 40–49 | −0.06 | 0.571 | −0.06 | 0.563 | −0.06 | 0.574 | −0.06 | 0.572 | −0.03 | 0.808 | |
| Age 50–59 | −0.28 | 0.032 | −0.28 | 0.033 | −0.28 | 0.032 | −0.28 | 0.032 | −0.34 | 0.018 | |
| Age ≥60 | −0.68 | 0.000 | −0.67 | 0.000 | −0.68 | 0.000 | −0.68 | 0.000 | −0.76 | 0.000 | |
| Relationship x Black Resp. | −0.24 | 0.497 | |||||||||
| Relationship x Hispanic Resp. | 0.50 | 0.238 | |||||||||
| Relationship x Asian Resp. | 0.54 | 0.114 | |||||||||
| Relationship x Other/Multi. Resp. | −0.48 | 0.234 | |||||||||
| Relationship x Female Resp. | 0.07 | 0.699 | |||||||||
| Relationship x Female Recip. | 0.05 | 0.793 | |||||||||
| Relationship x Age ≤29 | 0.41 | 0.060 | |||||||||
| Relationship x Age 40–49 | −0.12 | 0.639 | |||||||||
| Relationship x Age 50–59 | 0.33 | 0.280 | |||||||||
| Relationship x Age ≥60 | 0.42 | 0.289 | |||||||||
| Parent (N=5051) | 0.55 | 0.000 | 0.54 | 0.000 | 0.69 | 0.000 | 0.32 | 0.007 | 0.59 | 0.000 | |
| Black Resp. | −0.22 | 0.122 | −0.22 | 0.170 | −0.21 | 0.130 | −0.22 | 0.123 | −0.22 | 0.120 | |
| Hispanic Resp. | 0.01 | 0.939 | −0.17 | 0.405 | 0.01 | 0.930 | 0.02 | 0.926 | 0.01 | 0.960 | |
| Asian Resp. | −0.13 | 0.348 | −0.12 | 0.451 | −0.13 | 0.360 | −0.13 | 0.341 | −0.13 | 0.350 | |
| Other/Multiracial Resp. | 0.06 | 0.695 | 0.12 | 0.466 | 0.06 | 0.684 | 0.04 | 0.761 | 0.05 | 0.701 | |
| Female Resp. | 0.18 | 0.010 | 0.18 | 0.009 | 0.18 | 0.009 | 0.08 | 0.304 | 0.18 | 0.008 | |
| Female Recip. | 0.10 | 0.151 | 0.10 | 0.153 | 0.17 | 0.040 | 0.10 | 0.162 | 0.11 | 0.144 | |
| Age ≤29 | 0.02 | 0.817 | 0.02 | 0.839 | 0.02 | 0.842 | 0.03 | 0.763 | 0.00 | 0.964 | |
| Age 40–49 | −0.03 | 0.737 | −0.03 | 0.742 | −0.03 | 0.753 | −0.03 | 0.744 | 0.02 | 0.846 | |
| Age 50–59 | −0.31 | 0.022 | −0.31 | 0.021 | −0.31 | 0.021 | −0.32 | 0.018 | −0.24 | 0.106 | |
| Age ≥60 | −0.68 | 0.006 | −0.69 | 0.006 | −0.69 | 0.006 | −0.70 | 0.005 | −0.64 | 0.025 | |
| Relationship x Black Resp. | 0.00 | 0.999 | |||||||||
| Relationship x Hispanic Resp. | 0.62 | 0.096 | |||||||||
| Relationship x Asian Resp. | −0.04 | 0.901 | |||||||||
| Relationship x Other/Multi. Resp. | −0.26 | 0.434 | |||||||||
| Relationship x Female Resp. | −0.28 | 0.090 | |||||||||
| Relationship x Female Recip. | 0.43 | 0.008 | |||||||||
| Relationship x Age ≤29 | 0.10 | 0.585 | |||||||||
| Relationship x Age 40–49 | −0.28 | 0.263 | |||||||||
| Relationship x Age 50–59 | −0.33 | 0.325 | |||||||||
| Relationship x Age ≥60 | −0.18 | 0.752 | |||||||||
Table 4C:
Regression Analysis
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Relationship | Coefficient | B | p | B | p | B | p | B | p | B | p |
| Friend (N=5396) | −1.12 | 0.000 | −1.15 | 0.000 | −0.98 | 0.000 | −1.01 | 0.000 | −0.94 | 0.000 | |
| Black Resp. | −0.26 | 0.056 | −0.22 | 0.163 | −0.26 | 0.053 | −0.26 | 0.054 | −0.26 | 0.052 | |
| Hispanic Resp. | −0.05 | 0.771 | 0.02 | 0.913 | −0.05 | 0.772 | −0.05 | 0.750 | −0.05 | 0.770 | |
| Asian Resp. | −0.06 | 0.645 | −0.02 | 0.882 | −0.07 | 0.637 | −0.06 | 0.660 | −0.07 | 0.609 | |
| Other/Multiracial Resp. | 0.04 | 0.788 | −0.18 | 0.260 | 0.04 | 0.794 | 0.03 | 0.809 | 0.04 | 0.796 | |
| Female Resp. | 0.17 | 0.013 | 0.17 | 0.012 | 0.18 | 0.006 | 0.21 | 0.005 | 0.17 | 0.012 | |
| Female Recip. | 0.08 | 0.236 | 0.08 | 0.243 | 0.15 | 0.061 | 0.10 | 0.171 | 0.08 | 0.240 | |
| Age ≤29 | 0.03 | 0.772 | 0.03 | 0.765 | 0.02 | 0.775 | 0.03 | 0.744 | 0.07 | 0.458 | |
| Age 40–49 | −0.09 | 0.358 | −0.09 | 0.375 | −0.09 | 0.363 | −0.09 | 0.358 | −0.05 | 0.656 | |
| Age 50–59 | −0.36 | 0.003 | −0.36 | 0.003 | −0.36 | 0.003 | −0.36 | 0.003 | −0.28 | 0.043 | |
| Age ≥60 | −0.60 | 0.001 | −0.60 | 0.001 | −0.60 | 0.001 | −0.60 | 0.001 | −0.37 | 0.066 | |
| Relationship x Black Resp. | −0.14 | 0.648 | |||||||||
| Relationship x Hispanic Resp. | −0.29 | 0.461 | |||||||||
| Relationship x Asian Resp. | −0.15 | 0.636 | |||||||||
| Relationship x Other/Multi. Resp. | 0.88 | 0.006 | |||||||||
| Relationship x Female Resp. | −0.27 | 0.086 | |||||||||
| Relationship x Female Recip. | −0.20 | 0.204 | |||||||||
| Relationship x Age ≤29 | −0.20 | 0.286 | |||||||||
| Relationship x Age 40–49 | −0.18 | 0.445 | |||||||||
| Relationship x Age 50–59 | −0.38 | 0.174 | |||||||||
| Relationship x Age ≥60 | −0.81 | 0.017 | |||||||||
| Aunt/Uncle (N=1906) | −0.32 | 0.064 | −0.23 | 0.242 | −0.45 | 0.063 | −0.21 | 0.383 | −0.64 | 0.030 | |
| Black Resp. | 0.00 | 0.986 | 0.01 | 0.984 | 0.00 | 0.995 | 0.01 | 0.976 | 0.00 | 0.987 | |
| Hispanic Resp. | −0.14 | 0.702 | −0.13 | 0.740 | −0.14 | 0.694 | −0.12 | 0.733 | −0.13 | 0.711 | |
| Asian Resp. | 0.01 | 0.981 | 0.15 | 0.617 | 0.01 | 0.969 | 0.00 | 0.990 | 0.01 | 0.968 | |
| Other/Multiracial Resp. | −0.22 | 0.419 | −0.10 | 0.736 | −0.22 | 0.423 | −0.22 | 0.412 | −0.21 | 0.430 | |
| Female Resp. | 0.24 | 0.069 | 0.23 | 0.075 | 0.23 | 0.072 | 0.28 | 0.054 | 0.23 | 0.078 | |
| Female Recip. | −0.14 | 0.309 | −0.15 | 0.282 | −0.19 | 0.214 | −0.14 | 0.317 | −0.14 | 0.312 | |
| Age ≤29 | 0.02 | 0.900 | 0.02 | 0.907 | 0.02 | 0.900 | 0.03 | 0.880 | −0.08 | 0.670 | |
| Age 40–49 | 0.06 | 0.774 | 0.05 | 0.791 | 0.06 | 0.777 | 0.06 | 0.763 | −0.01 | 0.972 | |
| Age 50–59 | −0.30 | 0.216 | −0.29 | 0.234 | −0.30 | 0.213 | −0.31 | 0.209 | −0.35 | 0.190 | |
| Age ≥60 | −0.19 | 0.648 | −0.19 | 0.649 | −0.20 | 0.641 | −0.20 | 0.640 | −0.30 | 0.501 | |
| Relationship x Black Resp. | 0.05 | 0.950 | |||||||||
| Relationship x Hispanic Resp. | −0.03 | 0.977 | |||||||||
| Relationship x Asian Resp. | −0.89 | 0.226 | |||||||||
| Relationship x Other/Multi. Resp. | −0.48 | 0.442 | |||||||||
| Relationship x Female Resp. | 0.26 | 0.440 | |||||||||
| Relationship x Female Recip. | −0.23 | 0.508 | |||||||||
| Relationship x Age ≤29 | 0.58 | 0.163 | |||||||||
| Relationship x Age 40–49 | 0.38 | 0.430 | |||||||||
| Relationship x Age 50–59 | 0.29 | 0.638 | |||||||||
| Relationship x Age ≥60 | 0.77 | 0.491 | |||||||||
Table 4D:
Regression Analysis
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Relationship | Coefficient | B | p | B | p | B | p | B | p | B | p |
| Coworker (N=1719) | −2.55 | 0.000 | −2.67 | 0.000 | −2.75 | 0.000 | −2.64 | 0.000 | −2.27 | 0.000 | |
| Black Resp. | −0.05 | 0.845 | −0.26 | 0.369 | −0.04 | 0.868 | −0.05 | 0.847 | −0.05 | 0.863 | |
| Hispanic Resp. | −0.04 | 0.894 | 0.02 | 0.955 | −0.04 | 0.906 | −0.04 | 0.903 | −0.05 | 0.888 | |
| Asian Resp. | 0.12 | 0.652 | 0.05 | 0.876 | 0.12 | 0.656 | 0.13 | 0.635 | 0.12 | 0.651 | |
| Other/Multiracial Resp. | −0.05 | 0.856 | −0.11 | 0.718 | −0.06 | 0.817 | −0.05 | 0.855 | −0.06 | 0.814 | |
| Female Resp. | 0.24 | 0.060 | 0.25 | 0.056 | 0.24 | 0.060 | 0.21 | 0.141 | 0.25 | 0.058 | |
| Female Recip. | −0.07 | 0.589 | −0.07 | 0.590 | −0.15 | 0.313 | −0.07 | 0.587 | −0.07 | 0.600 | |
| Age ≤29 | −0.07 | 0.666 | −0.07 | 0.681 | −0.07 | 0.688 | −0.07 | 0.662 | 0.01 | 0.944 | |
| Age 40–49 | −0.10 | 0.593 | −0.11 | 0.585 | −0.11 | 0.578 | −0.11 | 0.590 | 0.00 | 0.986 | |
| Age 50–59 | −0.27 | 0.258 | −0.26 | 0.278 | −0.28 | 0.250 | −0.27 | 0.262 | −0.27 | 0.320 | |
| Age ≥60 | −0.81 | 0.048 | −0.81 | 0.047 | −0.81 | 0.045 | −0.80 | 0.048 | −0.62 | 0.157 | |
| Relationship x Black Resp. | 1.24 | 0.076 | |||||||||
| Relationship x Hispanic Resp. | −0.30 | 0.713 | |||||||||
| Relationship x Asian Resp. | 0.51 | 0.499 | |||||||||
| Relationship x Other/Multi. Resp. | 0.31 | 0.644 | |||||||||
| Relationship x Female Resp. | 0.41 | 0.221 | |||||||||
| Relationship x Female Recip. | 0.18 | 0.589 | |||||||||
| Relationship x Age ≤29 | −0.45 | 0.258 | |||||||||
| Relationship x Age 40–49 | −0.59 | 0.248 | |||||||||
| Relationship x Age 50–59 | −0.07 | 0.900 | |||||||||
| Relationship x Age ≥60 | −1.24 | 0.220 | |||||||||
Specifically, among respondents with children, those whose children were selected in the vignette assigned much higher willingness ratings than those where some other recipient type was selected (B=1.41, p<0.001), net of controls. The interactions of the relationship variable with race/ethnicity, respondent gender, and recipient gender were all substantively small and statistically insignificant. However, respondents aged 40–49 (B=0.57, p=0.046), 50–59 (B=0.71, p=0.029), and ≥60 (B=0.72, p=0.066) all assigned statistically significantly higher willingness ratings than respondents age 30–39 when children were designated the recipient. The total effect of this interaction across all age categories was statistically significant at the p<0.10 level (F=1.98, p=0.095).
Similarly, among respondents with a spouse/partner (all in waves 2–3), those whose spouse/partner were selected in the living kidney donation vignette assigned much higher willingness ratings than those where some other recipient type was selected (B=0.98, p<0.001). The interactions of the relationship variable with each demographic variable besides age yielded no substantively meaningful or statistically significant coefficients.
Among respondents with siblings, those whose siblings were selected in the living kidney donation vignette assigned moderately higher willingness ratings than those where some other recipient type was selected (B=0.42, p<0.001). The interactions of the relationship variable with race/ethnicity and each gender variable yielded no substantively meaningful or statistically significant coefficients. For age, respondents age 20–29 assigned statistically significantly higher willingness ratings than 30–39 year olds when siblings were designated the recipient. However, the age interaction with relationship was not statistically significant in an F-test across all age categories.
Among respondents with parents, those whose parents were selected in the living kidney donation vignette assigned moderately higher willingness ratings than those where some other recipient type was selected (B=0.55, p<0.001). The interactions of the relationship variable with respondent race/ethnicity, gender, and age yielded no substantively meaningful or statistically significant coefficients. However, those responding to scenarios where mothers rather than fathers were selected assigned moderately higher willingness ratings (B=0.43, p=0.008). Two additional interactive coefficients were statistically significant only at the p<0.10 level: Hispanic respondents assigned moderately higher willingness ratings to their parents than White respondents (B=0.62, p=0.096), and women respondents assign lower willingness to their parents than men respondents (B=−0.28, p=0.090).
Among respondents with friends, those whose friends were selected in the living kidney donation vignette assigned significantly lower willingness ratings than those where some other recipient type was selected (B=−1.12, p<0.001). This deficit was statistically significantly smaller among other/multiracial respondents compared to White respondents (B=0.88, p=0.006), but not for other racial/ethnic groups. The total effect of the relationship-race/ethnicity interaction was statistically significant only at the p<0.10 level (F=2.26, p=0.061). Neither gender interaction was statistically significant at the p<0.05 level, but women respondents assigned lower willingness ratings to their friends than men respondents (B=−0.27, p=0.086). Finally, respondents age≥60 reported statistically significantly lower willingness ratings than respondents age 30–39 when friends were designated the recipient. However, the age interaction with relationship was not statistically significant in an F-test across all age categories.
Among respondents with aunts/uncles (all in wave 1), respondents whose aunts/uncles were selected in the living kidney donation vignette assigned statistically significantly lower willingness ratings than those where some other relationship type was selected only at the p<0.10 level (B=− 0.32, p=0.064). None of the demographic interactions were statistically significant.
Among respondents who were employed, those whose coworkers were selected in the living kidney donation vignette assigned significantly lower willingness ratings than those where some other recipient type was selected (B=−2.55, p<0.001). None of the demographic interactions were statistically significant at the p<0.05 level, but Black respondents did assign significantly higher (at the p<0.10 level) willingness ratings to coworkers than did White respondents (B=1.24, p=0.076).
Across all non-interactive models, the main effects of respondent race/ethnicity and respondent or recipient gender are generally modest and not statistically significant. However, these effects did sometimes differ across relationship-specific models. For instance, in the child models, Black (B=−0.36, p=0.062) and Asian (B=−0.47, p=0.050) respondents assigned lower willingness ratings than White respondents net of other factors, but this was not the case for either racial/ethnic group for the parent, aunt/uncle, or coworker models, and was the case for Black respondents but not Asian respondents for the sibling and spouse/partner models. A more consistent finding is women’s higher average willingness to donate compared to men’s; the model 1 coefficient is positive in all models tested, is statistically significant at the p<0.05 level in 4 out of 7 models (spouse/partner, sibling, parent, friend), and in a further 2 of the 7 models is statistically significant at the p<0.10 level but not the p<0.05 level (aunt/uncle, coworker). Recipient gender was statistically significant only in the spouse/partner models (B=0.19, p=0.017).
5. Discussion
Growing racial/ethnic disparities in LDKT and persistent but smaller gender disparities therein imply that more research on basic factors driving these disparities is needed to explain and ameliorate them. In this paper, we consider a novel factor with the potential to partially explain these disparities: differential willingness to donate by donor race/ethnicity or donor/recipient gender. Using data derived from an experimental vignette study in which respondents were offered living kidney donation scenarios in which a randomly-chosen member of their network was selected as the hypothetical transplant candidate, we show that although relationship effects on willingness to donate are highly important, they only sometimes significantly differ by respondent or recipient gender or respondent race/ethnicity. The only statistically significant interactions at the p<0.05 level are between parent relationship and recipient gender (favoring mothers over fathers) and friend relationship and respondent other/multiracial race/ethnicity (with Other/Multiracial respondents assigning higher ratings than Whites). Additionally, at the p<0.10 level, Hispanics show higher willingness than Whites to donate to their parents, women show lower willingness than men to donate to their parents, women show lower willingness than men to donate to their friends, and Blacks show higher willingness than Whites to donate to their coworkers.
These findings offer several implications for understanding LDKT disparities. For racial/ethnic LDKT disparities, these results suggest that marginalized racial/ethnic groups are not less willing to consider living kidney donation to the same recipient type compared to non-Hispanic Whites. Indeed, every relationship-race/ethnicity interaction with a statistically significant coefficient at the p<0.05 or p<0.10 level showed evidence of marginalized racial/ethnic groups’ higher willingness to donate to that recipient type rather than another relationship type in their network compared to non-Hispanic Whites. Thus, it is unlikely that racial/ethnic differences in LDKT rates from specific relationships are due to lower willingness to donate in that relationship category compared to Whites. Instead, we argue in line with many others40–43 that institutional and structural barriers to LDKT are the most likely explanation for these disparities.
Turning to gender disparities, it is surprising that the few respondent gender-relationship interactions with statistically significant coefficients (at the p<.10 level) show evidence of women’s lesser willingness to donate compared to men for that particular relationship. This was the case for donation to parents and friends. Considering that prior research24,28,29,32 shows that women are much more likely to become living kidney donors, this pattern of relationship-specific lower willingness to donate compared to men suggests that these patterns are not entirely due to differential willingness to donate. Instead, it is likely that these differences are due to other factors such as potential donor availability and health.
For disparities by recipient gender, only the mother advantage over fathers was statistically significant. This result is consistent with prior research on gender disparities in LDKT receipt, which shows that women’s rates of LDKT receipt from their children exceeded men’s prior to 2005, and shortly thereafter women’s advantage in paired LKD rates emerged.3 If women’s advantage in paired LKD rates is due to high rates of children enabling a paired LKD (perhaps due to pregnancy-induced presensitization precluding direct donation), these findings would suggest that children’s higher rates of donation to their mothers may be explained better by differential willingness than by medical or kin availability factors.
Some relationship effects also differed statistically significantly by age. Only the age interaction with child recipient was statistically significant at the p<0.10 level. This showed a Ushaped pattern in which respondents age 20–29 and ≥40 report greater willingness to donate than respondents age 30–39. This result may be an artifact of children’s age and the respondent’s age at children’s birth. The children of 20–29 year olds will generally be very young, as will a high percentage of the children of 30–39 year olds. When 30–39 year olds’ children are adults, this means they had them when they were very young if they are biological children. If respondents understand that it is often difficult to transplant adult kidneys into young children, or if respondents who had children at a young age respond differently than those who had them at older ages, this could explain this finding. Future research should investigate these matters further.
These results also highlight that, compared to average respondent willingness to donate to these recipients, parent-to-child and niece/nephew-to-aunt/uncle may be underutilized. In the case of parent-to-child, this is likely due to parents’ advanced age when their children reach typical kidney failure age. However, aunt/uncle recipients have only moderately lower average willingness than friends, yet friends are a common source of living kidney donations while nieces/nephews are not.3 Researchers and clinicians should explore common barriers to donation for niece/nephew potential donors and how they might be ameliorated.
This research is subject to a number of limitations. First, the sample is not drawn from the population of potential living kidney donors, but instead a convenience web-based sample obtained through Amazon’s Mechanical Turk platform. Although this pool of respondents enabled us to affordably collect a large sample of responses (N=5,635), the ability of this sample to generalize to the realistic potential donor population should be investigated in future research.44 Second, the sample’s racial/ethnic diversity was limited compared to the general population’s, which also limited our statistical power to test for racial/ethnic differences in willingness to be evaluated for donation by potential recipient relationship. Third, the three waves of data collection in this sample included different relationship sets and different vignette presentation styles in their design, which may limit the validity of the conclusions.
Supplementary Material
Highlights.
Racial/ethnic and gender disparities in living donor kidney transplants are large.
We surveyed participants’ responses to randomly assigned living donation vignettes.
Respondents are more willing to donate to close family than other relationships.
Respondents are more willing to donate to their mothers than their fathers.
Older respondents are more willing to donate to their parents than others.
Acknowledgements
This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases grants R01DK114888 and R01DK132953. We also acknowledge assistance provided by the Population Research Institute at Penn State University, which is supported by an infrastructure grant by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (P2CHD041025).
Abbreviations:
- ESKD
End-stage kidney disease
- DDKT
deceased donor kidney transplantation
- LDKT
living donor kidney transplantation
- LKD
living kidney donation
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Lentine KL, Smith JM, Hart A, et al. OPTN/SRTR 2020 Annual Data Report: Kidney. American Journal of Transplantation. 2022;22(S2):21–136. doi: 10.1111/ajt.16982 [DOI] [PubMed] [Google Scholar]
- 2.Purnell TS, Luo X, Cooper LA, et al. Association of Race and Ethnicity With Live Donor Kidney Transplantation in the United States From 1995 to 2014. JAMA. 2018;319(1):49–61. doi: 10.1001/jama.2017.19152 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Daw J A Third Era? Trends in the Living Kidney Donation Rate Per Active Waiting List-year, 1996–2021. Presented at: American Transplant Congress; 2023. [Google Scholar]
- 4.Daw J The ties that transplant: The social capital determinants of the living kidney donor relationship distribution. Social Science Research. 2023;113:102888. doi: 10.1016/j.ssresearch.2023.102888 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Purnell TS, Boulware LE. Racial Disparities in Kidney Transplant and Living Donation. In: STEEL J, ed. Living Donor Advocacy: An Evolving Role Within Transplantation. Springer; 2014:327–345. doi: 10.1007/978-1-4614-9143-9_23 [DOI] [Google Scholar]
- 6.Kumar K, Tonascia JM, Muzaale AD, et al. Racial differences in completion of the living kidney donor evaluation process. Clinical Transplantation. 2018;32(7):e13291. doi: 10.1111/ctr.13291 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lopez-Soler RI, Garcia-Roca R, Lee DD. Disparities in living donation. Current Opinion in Organ Transplantation. 2021;26(5):542–546. doi: 10.1097/MOT.0000000000000912 [DOI] [PubMed] [Google Scholar]
- 8.Ross LF, Thistlethwaite JR. Gender and race/ethnicity differences in living kidney donor demographics: Preference or disparity? Transplantation Reviews. 2021;35(3):100614. doi: 10.1016/j.trre.2021.100614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Wesselman H, Ford CG, Leyva Y, et al. Social Determinants of Health and Race Disparities in Kidney Transplant. CJASN. 2021;16(2):262–274. doi: 10.2215/CJN.04860420 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Daw J, Roberts M, Gillespie A, Verdery AM, Purnell TS. Testing the Differential Access Hypothesis That Black Kidney Transplant Candidates Perceive Social Network Access to Fewer Potential Living Donors Than White Candidates. Prog Transpl. 2023;33(2):130–140. doi: 10.1177/15269248231164174 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Waterman AD, Rodrigue JR, Purnell TS, Ladin K, Boulware LE. Addressing Racial and Ethnic Disparities in Live Donor Kidney Transplantation: Priorities for Research and Intervention. Seminars in Nephrology. 2010;30(1):90–98. doi: 10.1016/j.semnephrol.2009.10.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rodrigue JR, Kazley AS, Mandelbrot DA, Hays R, LaPointe Rudow D, Baliga P. Living Donor Kidney Transplantation: Overcoming Disparities in Live Kidney Donation in the US— Recommendations from a Consensus Conference. CJASN. 2015;10(9):1687–1695. doi: 10.2215/CJN.00700115 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rodrigue JR, Schold JD, Mandelbrot DA. The decline in living kidney donation in the United States: Random variation or cause for concern? Transplantation. 2013;96(9). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.United States Renal Data System. 2022 USRDS Annual Data Report: Epidemiology of Kidney Disease in the United States. National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2022. https://adr.usrds.org/2022 [Google Scholar]
- 15.Killian AC, Shelton B, MacLennan P, et al. Evaluation of Community-Level Vulnerability and Racial Disparities in Living Donor Kidney Transplant. JAMA Surgery. 2021;156(12):1120–1129. doi: 10.1001/jamasurg.2021.4410 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bratton C, Chavin K, Baliga P. Racial disparities in organ donation and why. Current Opinion in Organ Transplantation. 2011;16(2):243–249. doi: 10.1097/MOT.0b013e3283447b1c [DOI] [PubMed] [Google Scholar]
- 17.Lentine KL, Mandelbrot D. Addressing Disparities in Living Donor Kidney Transplantation: A Call to Action. CJASN. 2018;13(12):1909–1911. doi: 10.2215/CJN.06250518 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Brar A, Markell M. Impact of gender and gender disparities in patients with kidney disease. Current Opinion in Nephrology and Hypertension. 2019;28(2):178–182. doi: 10.1097/MNH.0000000000000482 [DOI] [PubMed] [Google Scholar]
- 19.Sheikh SS, Locke JE. Gender disparities in transplantation. Curr Opin Organ Transplant. 2021;26(5):513–520. doi: 10.1097/MOT.0000000000000909 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Mustian MN, Kumar V, Stegner K, et al. Mitigating Racial and Gender Disparities in Access to Living Donor Kidney Transplantation: Impact of the Nation’s Longest Single-Center Kidney Chain. Ann Surg. 2019;270(4):639–646. doi: 10.1097/SLA.0000000000003484 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Carrero JJ, Hecking M, Chesnaye NC, Jager KJ. Sex and gender disparities in the epidemiology and outcomes of chronic kidney disease. Nat Rev Nephrol. 2018;14(3):151–164. doi: 10.1038/nrneph.2017.181 [DOI] [PubMed] [Google Scholar]
- 22.Kayler LK, Rasmussen CS, Dykstra DM, et al. Gender Imbalance and Outcomes in Living Donor Renal Transplantation in The United States. American Journal of Transplantation. 2003;3(4):452–458. doi: 10.1034/j.1600-6143.2003.00086.x [DOI] [PubMed] [Google Scholar]
- 23.Piccoli GB, Al Rukhaimi M, Liu ZH, Zakharova E, Levin A. What we know and do not know about women and kidney diseases; Questions unanswered and answers unquestioned: Reflection on World Kidney Day and International Woman’s Day. Braz J Med Biol Res. 2018;51:e7315. doi: 10.1590/1414-431X20177315 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Vilayur E, van Zwieten A, Chen M, et al. Sex and Gender Disparities in Living Kidney Donation: A Scoping Review. Transplantation Direct. 2023;9(9):e1530. doi: 10.1097/TXD.0000000000001530 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Gill J, Joffres Y, Rose C, et al. The Change in Living Kidney Donation in Women and Men in the United States (2005–2015): A Population-Based Analysis. JASN. 2018;29(4):1301–1308. doi: 10.1681/ASN.2017111160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Gibney EM, Doshi MD, Hartmann EL, Parikh CR, Garg AX. Health Insurance Status of US Living Kidney Donors. Clin J Am Soc Nephrol. 2010;5(5):912–916. doi: 10.2215/CJN.07121009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Zimmerman D, Donnelly S, Miller J, Stewart D, Albert SE. Gender disparity in living renal transplant donation. American Journal of Kidney Diseases. 2000;36(3):534–540. doi: 10.1053/ajkd.2000.9794 [DOI] [PubMed] [Google Scholar]
- 28.Kayler LK, Meier-Kriesche HU, Punch JD, et al. Gender Imbalance in Living Donor Renal Transplantation. Transplantation. 2002;73(2):248. [DOI] [PubMed] [Google Scholar]
- 29.Bromberger B, Spragan D, Hashmi S, et al. Pregnancy-Induced Sensitization Promotes Sex Disparity in Living Donor Kidney Transplantation. JASN. 2017;28(10):3025–3033. doi: 10.1681/ASN.2016101059 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.August P, Suthanthiran M. Sex and Kidney Transplantation: Why Can’t a Woman Be More Like a Man? J Am Soc Nephrol. 2017;28(10):2829–2831. doi: 10.1681/ASN.2017060657 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Melk A, Babitsch B, Borchert-Mörlins B, et al. Equally Interchangeable? How Sex and Gender Affect Transplantation. Transplantation. 2019;103(6):1094–1110. doi: 10.1097/TP.0000000000002655 [DOI] [PubMed] [Google Scholar]
- 32.Kayler LK, Armenti VT, Dafoe DC, Burke JF, Francos GC, Ratner LE. Patterns of Volunteerism, Testing, and Exclusion Among Potential Living Kidney Donors. Health Care for Women International. 2005;26(4):285–294. doi: 10.1080/07399330590925772 [DOI] [PubMed] [Google Scholar]
- 33.Thompson TL, Robinson JD, Kenny RW. Gender Differences in Family Communication About Organ Donation1. Sex Roles. 2003;49(11/12):587–596. [Google Scholar]
- 34.Tuohy KA, Johnson S, Khwaja K, Pavlakis M. Gender Disparities in the Live Kidney Donor Evaluation Process. Transplantation. 2006;82(11):1402–1407. doi: 10.1097/01.tp.0000248953.64931.15 [DOI] [PubMed] [Google Scholar]
- 35.Gillespie A, Hammer H, Kolenikov S, et al. Sex Differences and Attitudes toward Living Donor Kidney Transplantation among Urban Black Patients on Hemodialysis. CJASN. 2014;9(10):1764–1772. doi: 10.2215/CJN.12531213 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Weinberg J, Freese J, McElhattan D. Comparing Data Characteristics and Results of an Online Factorial Survey between a Population-Based and a Crowdsource-Recruited Sample. SocScience. 2014;1:292–310. doi: 10.15195/v1.a19 [DOI] [Google Scholar]
- 37.Mullinix KJ, Leeper TJ, Druckman JN, Freese J. The Generalizability of Survey Experiments. Journal of Experimental Political Science. 2015;2(2):109–138. doi: 10.1017/XPS.2015.19 [DOI] [Google Scholar]
- 38.Levay KE, Freese J, Druckman JN. The Demographic and Political Composition of Mechanical Turk Samples. SAGE Open. 2016;6(1):2158244016636433. doi: 10.1177/2158244016636433 [DOI] [Google Scholar]
- 39.Porter ND, Verdery AM, Gaddis SM. Enhancing big data in the social sciences with crowdsourcing: Data augmentation practices, techniques, and opportunities. PLOS ONE. 2020;15(6):e0233154. doi: 10.1371/journal.pone.0233154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Mohottige D, McElroy LM, Boulware LE. A Cascade of Structural Barriers Contributing to Racial Kidney Transplant Inequities. Advances in Chronic Kidney Disease. 2021;28(6):517–527. doi: 10.1053/j.ackd.2021.10.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Monson RS, Kemerley P, Walczak D, Benedetti E, Oberholzer J, Danielson KK. Disparities in completion rates of the medical pre-renal transplant evaluation by race/ethnicity and gender. Transplantation. 2015;99(1):236–242. doi: 10.1097/TP.0000000000000271 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Patzer RE, McClellan WM. Influence of race, ethnicity and socioeconomic status on kidney disease. Nature Reviews Nephrology. 2012;8(9):533–541. doi: 10.1038/nrneph.2012.117 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Lu Y, Norman SP, Doshi MD. Understanding Structural Racism as a Barrier to Living Donor Kidney Transplantation and Transplant Care. Curr Transpl Rep. 2022;9(2):119–126. doi: 10.1007/s40472-021-00338-x [DOI] [Google Scholar]
- 44.Pew Research Center. Comparing Two Types of Online Survey Samples; 2023. https://www.pewresearch.org/methods/2023/09/07/comparing-two-types-of-online-surveysamples/
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.