Healthcare‐Related Sociocultural Factors, Racial Disparities, and Kidney Transplant Outcomes in the Kidney Transplant Fast Track Program

ABSTRACT

Background

In the United States, streamlining the kidney transplantation (KT) evaluation process may reduce disparities and barriers to KT access. Prior work showed that the Kidney Transplant Fast Track (KTFT) program shortened this process and reduced racial disparities in waitlisting and overall KT. In the present study, we examined whether race/ethnicity and sociocultural factors (e.g., medical mistrust, healthcare‐related discrimination/racism) remained independently associated with time to active waitlisting and KT receipt (overall, living [LDKT] and deceased donor KT [DDKT]) and whether sociocultural factors attenuated racial/ethnic differences in these outcomes.

Methods

In this secondary analysis, data came from the KTFT study, a prospective KT candidate cohort. Participants were recruited before KT evaluation start (05/2015–06/2018), coinciding with baseline measure collection, then followed via medical record through 08/2022. We used hierarchically‐adjusted Fine–Gray proportional hazards models in this exploratory analysis.

Results

Among 1108 KT candidates (243 Black, 783 White, 82 Other), medical mistrust was associated with lower cumulative incidence of waitlisting, but no other sociocultural factors were associated with outcomes. Racial and ethnic differences emerged for KT type: Black participants had a greater cumulative incidence of DDKT, and participants categorized as other race/ethnicity had a lower cumulative incidence of LDKT, relative to White participants.

Conclusions

Although KTFT reduced racial/ethnic disparities in waitlisting and overall KT receipt, we identified racial/ethnic differences in LDKT and DDKT. Medical mistrust was a significant barrier to waitlisting. Findings suggest that even when the KT evaluation process is streamlined, sociocultural factors and race/ethnicity may remain associated with KT outcomes.

Trial Registration

ClinicalTrials.gov identifier: NCT02342119

Abbreviations

CKD

chronic kidney disease

DDKT

deceased donor kidney transplantation

KAS

kidney allocation system

KTFT

kidney transplant fast track

KT

kidney transplantation

LDKT

living donor kidney transplantation

1. Introduction

Kidney transplantation (KT) is the optimal therapy for kidney failure [1, 2, 3]. Yet, many patients, particularly those from racially and ethnically minoritized populations in the United States, face persistent barriers to KT access [3, 4, 5, 6, 7, 8, 9, 10]. Before being considered eligible for KT, patients must complete a potentially burdensome, multi‐step KT evaluation process spanning several clinical specialties and settings [11, 12, 13, 14, 15, 16, 17]. Typically, the patient is responsible for managing these various appointments and tests across different clinics [15, 16], and evidence suggests that Black patients take longer than White patients to complete the KT evaluation process [7], contributing to broader racial disparities in KT access [3]. This disparity may reflect structural barriers across the KT continuum, including upstream delays in referral [7, 11, 13, 14, 15], and may be exacerbated by interpersonal and healthcare‐system related challenges that can influence how patients engage with the KT evaluation process—including experiences of discrimination and perceived racism in healthcare, as well as medical mistrust, which Black patients report more frequently than White patients in KT contexts [7, 8, 18].

To reduce racial and ethnic disparities during evaluation, Myaskovsky et al. implemented the Kidney Transplant Fast Track (KTFT) program within a large, single‐center KT clinic [19, 20]. KTFT is a concierge‐based approach designed to reduce clinic‐level logistical barriers in KT evaluation by streamlining and standardizing the KT evaluation process via scheduling all KT patients’ appointments and coordinating receipt of all pre‐KT testing [19, 20]. Compared to historical controls, KTFT receipt was associated with greater cumulative incidence of waitlisting and KT after adjusting for sociodemographic and medical characteristics. Moreover, the racial disparity in waitlisting and overall KT receipt (i.e., KT receipt from living and deceased donors) between Black and White patients was no longer significant within the KTFT cohort [19].

These improvements are especially notable in the context of broader national policy reforms. For example, the 2014 implementation of the Kidney Allocation System (KAS) [21] allowed patients to begin accruing time on the KT waitlist starting from dialysis initiation, regardless of when they were officially added to the waitlist [21]. Since KAS implementation [21], national trends suggest that racial and ethnic disparities in deceased donor KT (DDKT) have narrowed, but living donor KT (LDKT) disparities persist [3, 4, 5]. Importantly, KAS did not address barriers in KT evaluation and waitlisting—the stages targeted by KTFT [19].

Despite the demonstrated success of KTFT's streamlining aspect in improving racial equity in KT waitlisting and receipt in its patient cohort [19], important questions remain. First, we do not yet know whether disparities persisted in KT type; the prior study did not examine racial or ethnic differences between LDKT and DDKT [19]. Given the advantages of LDKT over DDKT, including decreased KT wait times and increased long‐term survival [22, 23, 24, 25], identifying ongoing disparities in KT type is critical.

Second, although healthcare‐related sociocultural factors (henceforth: sociocultural factors), including experiences of discrimination or perceived racism in healthcare, medical mistrust, and trust in physicians, have been identified as key correlates of KT‐related outcomes [7, 8, 18], they were not assessed in the prior KTFT study [19]. Despite their intuitive relevance, the association of sociocultural factors with the longitudinal progression through key KT milestones has not been empirically examined within the KTFT cohort [19], nor more broadly in settings where structural barriers, like a lengthy KT evaluation process, have been addressed. We conceptualize these four sociocultural factors as interpersonal or systemic experiences within healthcare that can shape patient‐provider relationships, trust in healthcare providers and systems, and engagement with complex healthcare processes such as KT evaluation. Prior cross‐sectional research across chronic disease contexts demonstrates that these sociocultural experiences are associated with poorer communication, and unmet or delayed care [7, 8, 13, 26, 27, 28, 29, 30, 31]. In previous work, Hamoda et al. found that sociocultural factors, not race, were associated with lower odds of attending the initial KT evaluation appointment [8]. This finding was significant, given race has historically served as a proxy for these factors [32]. Unlike race, sociocultural factors may be modifiable targets of racial and ethnic disparities. Thus, examining these factors within the KTFT cohort may reveal whether their associations with key KT‐related milestones persists despite reductions in racial disparities, and may provide insights for strategies that enhance KT access.

In this exploratory secondary analysis of a prospective observational cohort, we addressed these two gaps. All participants underwent the KTFT program, which addressed a key KT barrier by streamlining the evaluation process. Within this context, this study was motivated by the question of whether: (1) race and ethnicity, and (2) sociocultural factors, remained independently associated with multiple KT outcomes (i.e., time to: active waitlisting, and KT, LDKT, and DDKT receipt), and whether sociocultural factors attenuated racial and ethnic differences in these outcomes. Additionally, we examined these relationships using hierarchical models that adjusted for demographic, medical, and psychosocial factors, donor recruitment and preference, and transplant knowledge and concerns, thereby evaluating whether our key variables remained associated with KT outcomes after accounting for other commonly cited explanations of KT disparities (e.g., KT knowledge and donor preferences; Table S1). Although waitlisting and overall KT receipt were included in the primary KTFT paper as time‐to‐event outcomes, these models were limited to demographic and medical covariates [19]. The present analysis introduced a hierarchical modeling framework incorporating an expanded set of variables, representing the first comprehensive evaluation of these sociocultural factors as predictors in longitudinal models across multiple KT outcomes, including KT type, within the KTFT cohort.

2. Methods

2.1. Study Sample and Procedures

This prospective observational cohort study of KT candidates is from a larger clinical trial, “Increasing Equity in Transplant Evaluation and Living Donor Kidney Transplantation” [19, 20]. In the original study, patients were recruited from a single center, 05/2015–06/2018. Eligibility for recruitment included adult, English‐speaking patients referred for KT, that had not previously undergone KT, did not have cognitive or sensory impairments that would prevent them from participating, and had scheduled their initial KT evaluation appointment.

Among 1472 eligible patients, 1288 provided verbal consent after scheduling, but before attending, their initial KT evaluation appointment. During this period, we collected telephonic or paper‐based baseline surveys. Of these, only 1118 attended the initial evaluation appointment, and 1108 provided written consent for medical record review. Waitlisting and KT statuses were followed via medical records through 08/2022. See Myaskovsky et al. for details on patient flow [19], and Bornemann et al. for detailed information on study protocol [20].

The current study was approved by institutional review boards at the University of Pittsburgh and the University of New Mexico; both institutions signed a data use agreement. Study conduct complied with the Declaration of Helsinki and the Declaration of Istanbul on Organ Trafficking and Transplant Tourism. We followed “Strengthening the Reporting of Observational Studies in Epidemiology” reporting guidelines (Table S2) [33].

2.2. Study Measures

2.2.1. Outcome Variables

Outcome variables for this study included the cumulative incidence of (1) active KT waitlisting; (2) KT receipt (DDKT, LDKT, or unknown donor); (3) LDKT receipt; and (4) DDKT receipt.

2.2.2. Independent Variables

All independent variables were collected at baseline (i.e., after patients scheduled, but before attending, their initial KT evaluation appointment). The primary independent variables were race and ethnicity, and sociocultural factors. For race and ethnicity, patients first reported if they were Hispanic or Latino, and then asked their race (i.e., American Indian/Native American, Asian, Black/African American, Native Hawaiian/Pacific Islander, White, and/or “Other” race). We combined race and ethnicity into a race/ethnicity variable that yielded three groups: non‐Hispanic Black, non‐Hispanic White, and “Other race/ethnicity” (the latter group was created due to the small number of participants who did not indicate either “Black” or “White” race; all other races and ethnicities were collapsed into this category). Sociocultural factors were operationalized with the following four variables: (1) experience of discrimination in healthcare (“discrimination”; 7‐item scale reporting personal experiences of discrimination with healthcare providers; response range: “Never” to “Always”) [34, 35], perceived racism in healthcare (“racism”; 4‐item scale reporting how common racism is in healthcare; response range: “Strongly disagree” to “Strongly agree”) [36], medical mistrust of the healthcare system (“mistrust”; 7‐item scale reporting how hospital systems are untrustworthy, incompetent, and not acting in patients’ best interest; response range: “Strongly disagree” to “Strongly agree”) [36, 37, 38], and trust in physician (11‐item scale reporting how much respondents trust their physicians; response range: “Totally disagree” to “Totally agree”) [39]. Each measure used a 5‐point Likert scale. Items were averaged to create scale scores from 1 to 5, with higher scores reflecting greater discrimination, racism, mistrust, or trust in physician.

Patients reported demographic characteristics, transplant knowledge and concerns, psychosocial factors, donation preference and recruitment, and sociocultural factors in baseline surveys. We abstracted baseline medical factors via electronic medical review (see Table S1 for a full description of variables).

2.3. Statistical Analysis

We examined missingness in our data and ensured analytic assumptions were met. We log‐ or squared‐transformed independent variables with issues related to normality, skewness, or outliers, or converted continuous variables to categorical (e.g., we dichotomized the discrimination measure into “Ever” and “Never” because of its skewed distribution) [35]. Missing cases for each variable are reported under Table 1, with no systematic missingness pattern by racial/ethnic group. For categorical variables, we calculated frequencies and percentages. For continuous variables, we reported medians and first and third quartiles, which are resistant to potential outliers and closely approximate means for normal distributions. We compared participants by the three race/ethnicity categories on all survey items using Pearson's Chi‐squared Tests or One‐way Analysis of Variance (ANOVA) with Sidak correction (Table 1).

TABLE 1.

Characteristics of participants at baseline: Total sample, and by race and ethnicity.

		Race and ethnicity			3‐group comparison
Characteristic	Total (N = 1108)	Non‐Hispanic Black (n = 243)	Non‐Hispanic White (n = 783)	Other Race/Ethnicity a (n = 82)	Test statistic Χ 2 (df) or F statistic b	p value
Demographics factors
Age at T1 completion: Median (Q1, Q3)	59 (49, 67)	57 c (46, 64)	60 d (50, 68)	57 (41, 63)	10.28	<0.001
Gender: n (column %)					0.91 (2)	0.636
Male	695 (62.73)	150 (61.73)	497 (63.47)	48 (58.54)
Female	349 (38)	93 (38.27)	286 (36.53)	34 (41.46)
Educational level (more than High School degree): n (column %)	595 (53.70)	121 (49.79)	425 (54.28)	49 (59.76)	2.81 (2)	0.246
Family Income (less than $50,000): n (column %)	772 (72.28)	204 (88.31) c	510 (67.37)	58 (72.50) e	38.74 (2)	<0.001
Marital Status (married/domestic partnership): n (column %)	531 (47.92)	70 (28.81) c	426 (54.41) d	35 (42.68) e	49.67 (2)	<0.001
Employment status (part‐ or full‐time employed): n (column %)	268 (24.19)	47 (19.34)	201 (25.67)	20 (24.39)	4.05 (2)	0.132
Number in social network (i.e., number of potential living kidney donors): Median (Q1, Q3)	19.00 (11.00, 32.50)	20.00 (12.00, 35.00)	19.00 (11.00, 32.00)	19.00 (11.00, 33.00)	1.34	0.261
Log (Number in social network/potential living kidney donors): Median (Q1, Q3)	2.94 (2.40, 3.50)	3.00 (2.48, 3.56)	2.94 (2.40, 3.47)	2.94 (2.40, 3.50)	1.61	0.201
Insurance: n (column %)					72.08 (4)	<0.001
Private only	202 (18.23)	28 (11.52) c	161 (20.56) d	13 (15.85)
Public only	478 (43.14)	159 (65.43) c	276 (35.25) d	43 (52.44)
Both private and public	428 (38.63)	56 (23.05) c	346 (44.19) d	26 (31.71)
Medical factors
Type of dialysis: n (column %)					42.79 (4)	<0.001
Hemodialysis	568 (51.26)	159 (65.43) c	357 (45.59) d	52 (63.41)
Peritoneal dialysis	128 (11.55)	30 (12.35)	88 (11.24)	10 (12.20)
No dialysis	412 (37.18)	54 (22.22) c	338 (43.17) d	20 (24.30)
Dialysis duration: n (column %)					89.17 (6)	<0.001
No time on dialysis	416 (37.55)	56 (23.05) c	340 (43.42) d	20 (24.39)
<1 year on dialysis	401 (36.19)	87 (35.80)	283 (36.14)	31 (37.80)
1–5 years on dialysis	233 (21.03)	67 (27.57) c	145 (18.52) d	21 (25.61)
>5 years on dialysis	58 (5.23)	21 (25.61) c	15 (1.92) d	10 (12.20)
Body Mass Index (BMI): Median (Q1, Q3)	29.57 (25.57, 34.97)	28.51 (25.05, 34.44)	29.66 (25.54, 35.12)	32.16 (26.76, 35.46)	1.94	0.144
Burden of kidney disease: Median (Q1, Q3)	4.00 (3.00, 4.67)	4.00 (3.00, 4.67)	4.00 (3.00, 4.67)	4.00 (3.00, 4.67)	0.47	0.628
Charlson Comorbidity Index (CCI): Median (Q1, Q3)	4.00 (3.00, 5.00)	4.00 c (3.00, 6.00)	4.00 (3.00, 5.00)	4.00 e (3.00, 5.00)	6.45	0.002
Psychosocial characteristics
Social support: Median (Q1, Q3)	43.00 (37.00, 47.00)	42.00 c (36.00, 46.00)	43.00 d (38.00, 47.00)	41.00 (35.00, 45.00)	7.63	0.001
Anxiety (moderate‐to‐severe): n (column %)	54 (4.87)	10 (4.12)	38 (4.85)	6 (7.32)	1.36 (2)	0.507
Depression (moderate‐to‐severe): n (column %)	71 (6.41)	19 (7.82)	43 (5.49)	9 (10.98)	4.76 (2)	0.093
Family loyalty: Median (Q1, Q3)	3.19 (2.81, 3.56)	3.38 c (2.94, 3.75)	3.19 (2.75, 3.50)	3.25 e (2.88, 3.56)	12.86	<0.001
Any religious objection to LDKT: n (column %)	579 (52.97)	142 (59.41) c	388 (50.13)	49 (61.25)	8.69	0.013
Religiosity: Median (Q1, Q3)	7.00 (4.50, 9.00)	8.50 c (6.00, 9.00)	7.00 (4.00, 8.50)	7.00 e (4.00, 9.00)	14.28	<0.001
Health literacy: Median (Q1, Q3)	4.00 (3.00, 4.67)	4.00 (3.33, 4.67)	4.00 (3.00, 4.67)	4.00 (3.33, 4.67)	0.12	0.890
Transplant knowledge and concerns
Transplant knowledge: Median (Q1, Q3)	11.00 (9.00, 13.00)	10.00 c (8.00, 12.00)	11.00 (9.00, 13.00)	11.00 e (9.00, 13.00)	22.76	<0.001
Number of learning activities: Median (Q1, Q3)	2.00 (1.00, 3.00)	2.00 c (1.00, 3.00)	2.00 (1.00, 3.00)	2.00 (1.00, 3.00)	16.78	<0.001
Total hours of learning activities: Median (Q1, Q3)	4.00 (2.00, 9.00)	3.00 (1.00, 6.00)	4.50 (2.00, 10.00)	4.00 (1.00, 10.00)	0.15	0.865
Transplant concerns: Median (Q1, Q3)	87.00 (78.00, 96.00)	91.00 c (79.00, 101.00)	86.00 (77.00, 95.00)	89.00 (81.00, 95.00)	5.77	0.003
Donor recruitment and preference variables
Have a living donor at baseline (Yes): n (column %)	484 (43.76)	115 (47.52)	333 (42.58)	36 (43.90)	1.83 (2)	0.400
Donor Recruitment and Preference: n (column %)					1.16 (2)	0.561
Deceased donor	326 (29.42)	57 (23.46)	240 (30.65)	29 (35.37)	7.15 (4)	0.128
Living donor	746 (67.33)	178 (73.25)	516 (65.90)	52 (63.41)
Unable to answer	36 (3.25)	8 (3.29)	27 (3.45)	1 (1.22)
Willing to accept LDKT volunteer (Yes): n (column %)	969 (88.74)	210 (86.42)	683 (89.05)	76 (92.68)	2.66 (2)	0.265
Willing to ask for LDKT volunteer (Yes): n (column %)	554 (52.02)	142 (60.68) c	372 (49.53)	40 (50.00)	9.03 (2)	0.011
Healthcare‐related sociocultural factors
Any experience of discrimination (Yes): n (%)	335 (30.23)	149 (61.32) c	141 (18.01) d	45 (54.88)	190.40 (2)	<0.001
Perceived racism: Median (Q1, Q3)	2.25 (2.00, 2.75)	2.50 c (2.00, 3.25)	2.00 d (1.75, 2.50)	2.25 e (2.00, 3.00)	54.44	<0.001
Medical mistrust: Median (Q1, Q3)	3.00 (2.57, 3.57)	3.43 c (3.00, 3.83)	2.86 d (2.43, 3.43)	3.14 (2.71, 3.67)	35.50	<0.001
Trust in physician: Median (Q1, Q3)	4.00 (3.64, 4.45)	4.00 c (3.55, 4.27)	4.00 (3.64, 4.45)	3.91 (3.55, 4.18)	4.87	0.008

Note: Missing: n = 40 Family Income (n = 12 Non‐Hispanic Black, n = 26 Non‐Hispanic White, n = 2 Other race/ethnicity); n = 1 Burden of Kidney Disease (n = 1 Non‐Hispanic White); n = 1 Social Support (n = 1 Non‐Hispanic White); n = 1 Family Loyalty (n = 1 Non‐Hispanic White); n = 15 Any Religious Objection to LDKT (n = 4 Non‐Hispanic Black, n = 9 Non‐Hispanic White, n = 2 Other race/ethnicity); n = 4 Religiosity (n = 2 Non‐Hispanic Black, n = 2 Non‐Hispanic White); n = 1 Total Hours of Learning Activities (n = 1 Non‐Hispanic White); n = 2 Have a Living Donor at Baseline (n = 1 Non‐Hispanic Black, n = 1 Non‐Hispanic White); n = 16 Willing to Accept LDKT Volunteer (n = 16 Non‐Hispanic White); n = 43 Willing to Ask for LDKT Volunteer (n = 9 Non‐Hispanic Black, n = 32 Non‐Hispanic White, n = 2 Other race/ethnicity); n = 8 Perceived Racism in Healthcare (n = 1 Non‐Hispanic Black, n = 6 Non‐Hispanic White, n = 1 Other race/ethnicity); n = 4 Medical Mistrust (n = 1 Non‐Hispanic Black, n = 3 Non‐Hispanic White).

^a Other Race/Ethnicity includes n = 5 American Indian/Native American, n = 7 Asian, n = 24 Hispanic, n = 1 Native Hawaiian/Pacific Islander, n = 45 Mixed race or other.

^b Categorical variables used Chi‐square tests; continuous variables used one‐way ANOVA with Sidak correction. Bolded values indicate statistically significant omnibus test of p ≤ 0.05.

^c Significant difference of p ≤ 0.05 between Non‐Hispanic Black participants and Non‐Hispanic White participants.

^d Significant difference of p ≤ 0.05 between Non‐Hispanic White participants and other race/ethnicity participants.

^e Significant difference of p ≤ 0.05 between Non‐Hispanic Black participants and other race/ethnicity participants.

^c,d,eStatistical significance of p ≤ 0.05 confirmed by Pairwise Chi‐Square Tests with Bonferroni Correction (categorical variables) and one‐way ANOVA with Sidak correction (continuous variables).

2.3.1. Fine–Gray Proportional Hazards Model

To evaluate race/ethnicity and sociocultural factors as predictors of KT outcomes, we performed Fine–Gray competing risk survival analyses with cumulative incidence function on KT waitlisting, and among those waitlisted, overall KT, LDKT, and DDKT. For analyses of active KT waitlisting, time was measured from evaluation initiation to active waitlisting, regardless of interruptions or restarts, as long as waitlisting occurred before study completion. Censored events included waitlist rejection, incomplete evaluation, evaluation closure by patient or clinic, or ongoing evaluations at study completion. Death was categorized as a competing event. For analyses of overall KT, LDKT, or DDKT receipt, time was measured from active waitlisting to KT. Censored events included waitlist removal or waitlist maintenance by study completion. Death was a competing event for all models. Because receipt of one KT type precludes receipt of another, DDKT receipt was considered a competing event for LDKT models, and vice‐versa. Given the presence of competing risks, Fine–Gray proportional hazards analyses was preferred over standard Cox models, as the latter would censor events like death [40, 41], a distinct terminal consequence that prevents active waitlisting or transplantation.

We employed a hierarchical analytic approach [42] to evaluate associations between race/ethnicity, sociocultural factors, and KT‐related outcomes in this exploratory analysis. We first modeled the unadjusted association of race/ethnicity with each outcome to establish baseline differences (Block 1). We then added sociocultural factors to examine associations with outcomes and evaluate how they affected race/ethnicity estimates (Block 2). Finally, we added an expanded set of covariates, including demographic, medical, and psychosocial factors; donor recruitment and preference variables; and transplant knowledge or concerns, to evaluate whether race/ethnicity and/or sociocultural factors were associated with outcomes following further adjustment (Block 3). Our prior publication reported on medical‐ and demographic‐adjusted time‐to‐event models with race/ethnicity predicting active waitlisting and overall KT receipt [19], but our current hierarchical modeling approach allowed us to assess how associations between key predictors (i.e., race/ethnicity, sociocultural factors) and KT outcomes changed across sequentially‐adjusted models.

2.3.2. Model Selection

All variables were considered for inclusion in the fully‐adjusted Fine–Gray models across the four outcomes. Candidate covariates were identified a priori from prior literature across established domains, including demographic, medical, psychosocial, transplant knowledge or concerns, and donor preference and recruitment variables (Table S1). Given the large number of potentially correlated variables within these domains, we applied the least absolute shrinkage and selection operator (LASSO) to identify a more parsimonious set of covariates for inclusion in each fully‐adjusted model. LASSO applies regularization and shrinkage techniques to identify predictors for model inclusion, reduces overfitting, and addresses multicollinearity by retaining relevant variables and discarding redundant ones [43, 44].

This LASSO method accounted for competing risks data using the “fastcmprsk” R package [45]. To ensure that race/ethnicity and key sociocultural constructs of interests were consistently evaluated across outcomes, the following variables were forced into the LASSO procedure: race/ethnicity, discrimination, racism, mistrust, and trust in physicians. For each model determined by LASSO, we confirmed the proportional hazards assumption [46].

3. Results

Our study included 1108 patients (male n[%] = 695[63]; non‐Hispanic Black n[%] = 243[22]; non‐Hispanic White n[%] = 783[71]; Other race/ethnicity n[%] = 82[7]). Median age was 59 (interquartile range: 49–67). Participants significantly differed by race/ethnicity on several variables (Table 1). Frequencies and percentages of key KT‐related outcomes, including censored and competing events, can be found in footnotes under Tables 2, 3, 4.

TABLE 2.

Fine–Gray proportional sub‐distribution hazards models for time (in days) from initial KT evaluation to active waitlisting (competing event: death).

	SHR	p value	95% CI
Block 1 (unadjusted race/ethnicity model a )
Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	0.63	<0.001	0.50–0.79
Other race/ethnicity	0.77	0.121	0.56–1.07
Block 2 (race/ethnicity + sociocultural factors b )
Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	0.70	0.005	0.55–0.90
Other race/ethnicity	0.80	0.197	0.58–1.12
Experiences of discrimination in healthcare (Reference: Never experienced)	0.97	0.778	0.78–1.21
Perceived racism in healthcare	1.04	0.538	0.92–1.18
Medical mistrust	0.79	<0.001	0.69–0.90
Trust in physician	1.15	0.077	0.98–1.34
Block 3 (adjusted model with LASSO‐selected covariates c )
* Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	0.88	0.361	0.68–1.15
Other race/ethnicity	0.80	0.229	0.55–1.15
* Experiences of discrimination in healthcare (Reference: Never experienced)	1.16	0.191	0.93–1.46
* Perceived racism in healthcare	1.05	0.426	0.93–1.19
* Medical mistrust	0.85	0.014	0.74–0.97
* Trust in physician	1.04	0.670	0.88–1.22
Age	0.98	<0.001	0.98–0.99
Family income of less than $50k per year (Reference: $50k or more per year)	0.85	0.170	0.63–1.07
Marital status (Single, widowed, or divorced) (Reference: Married or partnered)	0.68	<0.001	0.55–0.84
Employment status (Unemployed) (Reference: Employed full‐/part‐time)	0.71	0.002	0.57–0.88
Dialysis modality (Reference: Not on dialysis)
Hemodialysis	1.09	0.866	0.42–2.84
Peritoneal dialysis	1.06	0.909	0.39–2.91
Dialysis duration (Reference: No time on dialysis)
Less than 1 year on dialysis	0.69	0.457	0.26–1.84
1–5 years on dialysis	0.55	0.251	0.21–1.51
More than 5 years on dialysis	0.32	0.044	0.11–0.97
CCI	0.84	<0.001	0.79–0.90
Burden of kidney disease	0.91	0.030	0.83–0.99
Transplant knowledge	1.05	0.013	1.01–1.08
Number of learning activities	1.05	0.318	0.96–1.14
Moderate‐to‐severe depression (Reference: No depression)	0.56	0.016	0.34–0.89
Social support	1.01	0.090	0.99–1.03

Note: Bolded values indicate statistical significance of p < 0.05.

^*Denotes variables that were forced into the LASSO procedure in Block 3 due to a priori aims. The remaining variables in Block 3 were selected for inclusion by LASSO. The total study sample was included in this analysis (N = 1108). The number of events in this analysis was as follows: n = 525 (47%) participants were actively waitlisted, n = 411 (37%) were censored, and n = 172 (16%) experienced a competing event (i.e., death).

^aSee Figure S1A.

^bSee Figure S1B.

^cSee Figure S1C.

TABLE 3.

Fine–Gray proportional sub‐distribution hazards models for time (in days) from active waitlisting to KT of any type (competing event: Death).

	SHR	p value	95% CI
Block 1 (unadjusted race/ethnicity model a )
Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	1.13	0.379	0.86–1.47
Other race/ethnicity	1.14	0.592	0.71–1.82
Block 2 (race/ethnicity + sociocultural factors b )
Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	1.26	0.128	0.94–1.70
Other race/ethnicity	1.15	0.563	0.71–1.86
Experiences of Discrimination in Healthcare (Reference: Never experienced)	0.91	0.478	0.69–1.19
Perceived Racism in Healthcare	0.91	0.268	0.77–1.08
Medical Mistrust	1.00	0.977	0.85–1.18
Trust in Physician	1.07	0.498	0.88–1.29
Block 3 (adjusted model with LASSO‐selected covariates c )
* Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	1.12	0.541	0.78–1.59
Other race/ethnicity	1.16	0.548	0.72–1.85
* Experiences of Discrimination in Healthcare (Reference: Never experienced)	0.89	0.439	0.67–1.19
* Perceived Racism in Healthcare	0.93	0.422	0.78–1.11
* Medical Mistrust	1.06	0.491	0.90–1.25
* Trust in Physician	1.03	0.774	0.84–1.26
Age	0.99	0.008	0.98–0.99
Education (High school degree or less) (Reference: More than high school)	0.82	0.135	0.63–1.06
Health Insurance (Reference: Public and private insurance)
Private insurance only	1.04	0.767	0.78–1.39
Public insurance only	0.88	0.473	0.63–1.24
Dialysis duration (Reference: No time on dialysis)
Less than 1 year on dialysis	0.79	0.070	0.61–1.02
1–5 years on dialysis	1.16	0.421	0.81–1.67
More than 5 years on dialysis	2.28	0.088	0.88–5.86
Health literacy	1.09	0.266	0.94–1.27
Any religious objection to LDKT (Reference: No religious objection)	0.82	0.098	0.65–1.04
Social Support	1.02	0.222	0.99–1.04
Transplant Knowledge	1.00	0.948	0.96–1.04
Hours of learning activities	1.01	0.055	0.96–1.04
Not having a living donor (Reference: Having a living donor at baseline)	0.77	0.043	0.60–0.99
Willing to accept a living donor (Reference: Not willing to accept a living donor)	0.78	0.219	0.52–1.16

Note: Bolded values indicate statistical significance of p < 0.05.

^*Denotes variables that were forced into the LASSO procedure in Block 3 due to a priori aims. The remaining variables in Block 3 were selected for inclusion by LASSO. Participants included in this analysis were those actively waitlisted (n = 525). Among those actively waitlisted, the number of events in this analysis was as follows: n = 330 (63%) participants were transplanted, n = 142 (27%) were censored, and n = 53 (10%) experienced a competing event (i.e., death).

^aSee Figure S2A.

^bSee Figure S2B.

^cSee Figure S2C.

TABLE 4.

Fine–Gray proportional sub‐distribution hazards models from active waitlisting to receipt of LDKT (competing events: Death, DDKT) and DDKT (competing events: Death, LDKT).

	LDKT receipt a			DDKT receipt b
	SHR	p value	95% CI	SHR	p value	95% CI
Block 1 (unadjusted race/ethnicity model c )
Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	0.41	0.005	0.22–0.77	1.94	<0.001	1.41–2.66
Other race/ethnicity	0.42	0.081	0.16–1.11	1.77	0.038	1.03–3.05
Block 2 (race/ethnicity + sociocultural factors d )
Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	0.48	0.023	0.26–0.91	2.10	<0.001	1.44–3.04
Other race/ethnicity	0.48	0.144	0.18–1.28	1.71	0.066	0.96–3.03
Experiences of discrimination in healthcare (Reference: Never experienced)	0.73	0.218	0.44–1.20	1.06	0.746	0.75–1.50
Perceived racism in healthcare	1.03	0.870	0.76–1.38	0.83	0.097	0.67–1.03
Medical mistrust	0.89	0.404	0.68–1.17	1.09	0.401	0.89–1.34
Trust in physician	0.96	0.818	0.69–1.34	1.15	0.245	0.91–1.47
Block 3 (adjusted model with LASSO‐selected covariates e )
* Race (Reference: Non‐Hispanic White)
Black, non‐Hispanic	0.47	0.060	0.21–1.03	1.59	0.028	1.05–2.41
Other race/ethnicity	0.36	0.027	0.14–0.89	1.78	0.077	0.94–3.38
* Experiences of Discrimination in Healthcare (Reference: Never experienced)	0.82	0.463	0.48–1.39	1.02	0.935	0.69–1.49
* Perceived Racism in Healthcare	0.79	0.173	0.56–1.11	0.88	0.269	0.70–1.10
* Medical Mistrust	1.06	0.676	0.81–1.40	1.02	0.860	0.81–1.28
* Trust in Physician	0.73	0.112	0.49–1.08	1.20	0.177	0.92–1.56
Age	0.98	0.025	0.96–0.99	1.00	0.498	0.98–1.01
Education (High school degree or less) (Reference: More than high school)	0.45	0.002	0.27–0.75	1.18	0.303	0.86–1.63
Family income of less than $50k per year (Reference: $50k or more per year)	0.41	<0.001	0.26–0.65	1.51	0.031	1.04–2.19
Employment status (Reference: Employed full‐/part‐time)	0.75	0.225	0.47–1.19	1.38	0.083	0.96–1.98
Social network (log transformed)	1.11	0.456	0.85–1.45	−	−	−
Dialysis modality (Reference: Not on dialysis)
Hemodialysis	11.85	<0.001	5.01–28.04	−	−	−
Peritoneal dialysis	14.16	<0.001	4.29–46.75	−	−	−
Dialysis duration (Reference: No time on dialysis)
Less than 1 year on dialysis	0.04	<0.001	0.01–0.10	1.17	0.355	0.84–1.65
1–5 years on dialysis	0.03	<0.001	0.01–0.10	1.84	0.004	1.21–2.81
More than 5 years on dialysis	<0.001	<0.001	<0.001–<0.001	4.50	0.003	1.70–11.96
Religiosity	1.07	0.159	0.98–1.16	−	−	−
Any religious objection to LDKT (Reference: No religious objection)	0.82	0.355	0.53–1.25	0.97	0.854	0.71–1.32
Having a living donor at baseline (Reference: Not having a living donor)	2.29	<0.001	1.44–3.64	1.00	0.997	0.70–1.42

Note: Bolded values indicate statistical significance of p < 0.05.

^*Denotes variables that were forced into the LASSO procedure in Block 3 due to a priori aims. The remaining variables in Block 3 were selected for inclusion by LASSO. Cells marked with “–” indicates that the variable was not included in the model for that outcome (i.e., not selected by LASSO) and are intentionally left blank. Participants included in these analyses were those actively waitlisted, and whose KT types were known (n = 500; 25 cases were removed from analysis). The number of events included in each analysis are as follows:

^a n = 113 (23%) participants received LDKT, n = 142 (28%) were censored, n = 245 (49%) experienced a competing event (i.e., DDKT or death).

^b n = 192 (38%) participants received DDKT, n = 142 (28%) were censored, n = 166 (33%) experienced a competing event (i.e., LDKT or death).

^c See Figure S3A/S4A.

^d See Figure S3B/S4B.

^e See Figure S3C/S4C.

3.1. Time to Active Waitlisting

In Block 1, Black participants (sub‐distribution hazard ratio [SHR] = 0.63; 95% CI: 0.50–0.79; p < 0.001), but not Other race/ethnicity participants (SHR = 0.77; 95% CI: 0.56–1.07; p = 0.121), had a significantly lower cumulative incidence of waitlisting compared with White participants. These findings establish baseline racial and ethnic differences within the KTFT cohort prior to further model adjustment. In Block 2, after introducing sociocultural factors, the same statistical pattern was observed (Black participants: SHR = 0.70, 95% CI: 0.55–0.90, p = 0.005; Other race/ethnicity participants: SHR = 0.80, 95% CI: 0.58–1.12, p = 0.197). In addition, medical mistrust was significantly associated with reduced cumulative incidence of waitlisting (SHR = 0.79; 95% CI: 0.69–0.90; p < 0.001). In Block 3, there were no statistically significant racial and ethnic differences regarding time to active waitlisting (Black participants: SHR = 0.88, 95% CI: 0.68–1.15, p = 0.361; Other race/ethnicity participants: SHR = 0.80, 95% CI: 0.55–1.15, p = 0.229). However, greater medical mistrust maintained its association with waitlisting (SHR = 0.85; 95% CI: 0.74–0.97; p = 0.014; Table 2 and Figure S1A–C).

3.2. Time From Active Waitlisting to KT

In Block 1, there were no statistically significant differences between Black participants (SHR = 1.13; 95% CI: 0.86–1.47; p = 0.379) or Other race/ethnicity participants (SHR = 1.14; 95% CI: 0.71–1.82; p = 0.592) compared with White participants in the cumulative incidence of KT. This pattern was maintained in Block 2 (Black: SHR = 1.26, 95% CI: 0.94–1.70, p = 0.128; Other race/ethnicity: SHR = 1.15, 95% CI: 0.71–1.86, p = 0.563), and Block 3 (Black: SHR = 1.12, 95% CI: 0.78–1.59, p = 0.541; Other race/ethnicity: SHR = 1.16, 95% CI: 0.72–1.85, p = 0.548). No sociocultural factors were associated with KT receipt (Table 3 and Figure S2A–C).

3.3. Time From Active Waitlisting to LDKT

There was a statistically significant lower cumulative incidence of LDKT among Black participants (SHR = 0.41; 95% CI: 0.22–0.77; p = 0.005), but not Other race/ethnicity participants (SHR = 0.42; 95% CI: 0.16–1.11; p = 0.081), compared with White participants in Block 1. This pattern persisted in Block 2, with slightly smaller differences (Black: SHR = 0.48, 95% CI: 0.26–0.91, p = 0.023; Other race/ethnicity: SHR = 0.48, 95% CI: 0.18–1.28, p = 0.144). However, in Block 3, the difference between Black and White participants was attenuated for LDKT (SHR = 0.47; 95% CI: 0.21–1.03; p = 0.060), but became statistically significant between Other race/ethnicity and White participants (SHR = 0.36; 95% CI: 0.14–0.89; p = 0.027). No statistically significant associations between sociocultural factors and LDKT receipt were identified (Table 4 and Figure 3A–C).

3.4. Time From Active Waitlisting to DDKT

In Block 1, a greater cumulative incidence of DDKT was observed among Black participants (SHR = 1.94; 95% CI: 1.41–2.66; p < 0.001) and Other race/ethnicity participants (SHR = 1.77; 95% CI: 1.03–3.05; p = 0.038) compared with White participants. Black participants maintained a significantly higher cumulative incidence of DDKT receipt compared with White participants in Block 2 (SHR = 2.10; 95% CI: 1.44–3.04; p < 0.001), while Other race/ethnicity participants did not (SHR = 1.71; 95% CI: 0.96–3.03; p = 0.066). In Block 3, Black participants still had a significantly greater cumulative incidence of DDKT over time (SHR = 1.59; 95% CI: 1.05–2.41; p = 0.028), but Other race/ethnicity participants did not (SHR = 1.78; 95% CI: 0.94–3.38; p = 0.077). No statistically significant associations between sociocultural factors and DDKT receipt were identified (Table 4 and Figure S4A–C).

4. Discussion

In this exploratory secondary analysis of a prospective observational cohort of patients who underwent the KTFT approach to KT evaluation, we examined whether sociocultural factors predicted multiple KT outcomes in time‐to‐event models, including waitlisting, KT, LDKT, and DDKT, and whether racial and ethnic differences existed for KT type. We found that medical mistrust was the only sociocultural factor associated with KT outcomes, such that it remained inversely associated with active waitlisting even when included alongside race/ethnicity and other covariates. Importantly, this association was maintained even after a major structural barrier to KT access (i.e., lengthy evaluation process) was explicitly addressed by KTFT [19], indicating that some barriers to waitlisting persisted despite improved clinic‐level processes. Although overall racial and ethnic disparities in waitlisting and KT receipt were attenuated within the KTFT context [19], we identified racial and ethnic differences in LDKT and DDKT that persisted after accounting for sociocultural factors and other covariates. This highlights distinctions in KT access that are not captured when KT receipt is considered as a single outcome. Together, these findings demonstrate that while KTFT addressed racial/ethnic disparities in waitlisting and overall KT, residual disparities remained in relation to medical mistrust and racial/ethnic differences in KT type. This work extends prior KTFT analyses by adjusting for an expanded set of covariates (e.g., sociocultural factors), and provides the first longitudinal characterization of racial and ethnic variations in LDKT and DDKT receipt within this cohort.

This study builds on the existing body of mostly cross‐sectional literature [8, 26, 28, 47] by using a prospective cohort design to examine how sociocultural factors are associated with multiple KT outcomes in longitudinal, time‐to‐event models—from evaluation initiation, through KT receipt—using a patient cohort that underwent a streamlined approach to KT evaluation. Whereas prior analyses of the KTFT cohort focused solely on waitlisting and overall KT receipt, and only adjusted for demographic and medical factors [19], the present study expanded the model and considered not only sociocultural factors, but also psychosocial factors, donor preference and recruitment, and transplant knowledge and concerns.

These latter variables were measured in the KTFT cohort because of their associations with KT access in previous work (Table S1), but to avoid redundancies, we used LASSO to construct parsimonious models. Although several variables were not retained because they did not provide additional predictive value in the Fine–Gray models after accounting for other correlated variables, they still may be important to KT outcomes. Ultimately, applying a hierarchical modeling approach across multiple KT milestones allowed us to evaluate how racial/ethnic and sociocultural associations with KT outcomes changed with sequential adjustment, representing a more comprehensive and systematic evaluation of factors associated with KT access than previously reported in this cohort.

Although previous studies have shown that discrimination, racism, medical mistrust, and trust in physician, are important to consider within a KT evaluation context [7, 8, 18], none have evaluated how these factors operate once major systemic barriers have been minimized; examining the KTFT cohort allowed us to do just that. Contrary to our expectations, we found that sociocultural factors were not significantly associated with any KT receipt outcome, suggesting that their influence may occur earlier in the chronic kidney disease (CKD) trajectory, or that KTFT may have mitigated their influence beyond waitlisting. These findings also suggest that some psychosocial, transplant knowledge/concerns, and donor preference‐related factors commonly discussed in KT literature may not provide additional explanatory value for KT outcomes once structural barriers to evaluation are minimized, and when other correlated clinical/demographic factors are considered. However, they may still contribute to disparities earlier in the KT continuum, or through pathways not captured in the present study.

Furthermore, consistent with KTFT's goal and our prior study [19], we once again observed no racial differences in overall KT receipt. However, we were able to build on this initial observation by exploring racial and ethnic differences in KT type. White participants were more likely to receive LDKT, but Black and Other race/ethnicity participants more commonly received DDKT. Interestingly, racial/ethnic differences in KT type shifted with model adjustment. For example, Black‐White differences in LDKT receipt attenuated after accounting for additional factors, whereas significant differences emerged between White and Other race/ethnicity participants only after adjustment. Notably, Black participants remained significantly more likely to receive DDKT compared with White participants even in the fully‐adjusted model, whereas the difference between Other race/ethnicity and White participants was reduced beyond the univariate model. These findings clarify that interventions addressing disparities in KT receipt must consider LDKT and DDKT separately, because there may be masked differential effects in KT type. Further, results suggest that, rather than sociocultural factors, perhaps broader contextual factors (e.g., access to living donors) influence KT outcomes, even within an intervention designed to promote KT access.

Patients in this KTFT cohort benefitted from a streamlined approach to KT evaluation. Still, the persistent association between medical mistrust and waitlisting, as well as the sustained racial/ethnic differences in KT type, underscore the limitations of relying on clinic‐level, equity‐focused interventions alone. In one prior analysis of the KTFT cohort, we found that patients reported improvements in discrimination, racism, and medical mistrust after completing KTFT, but Black participants’ trust in physicians declined [18]. Taken together with our current findings, this points to the need for systemic interventions to be paired with more personalized strategies that intentionally build trust and address patients’ prior experiences at institutional, as well as interpersonal and patient levels. This is particularly relevant for KT evaluation processes, which require sustained patient engagement with—and complex navigation of—healthcare systems. Although KTFT addressed a critical structural barrier, additional strategies at the individual‐ and interpersonal‐levels, like culturally‐concordant peer support to improve patient understanding of optimal KT treatments, or provider trainings to enhance shared decision‐making processes, may build trust in healthcare providers and systems, and support equitable outcomes [48]. Notably, in this same cohort, a culturally‐tailored educational intervention did not improve KT‐related outcomes overall [49], underscoring that education alone may be insufficient when patients face broader systemic barriers, or have limited agency to act on new knowledge. Embedding trust‐building and patient‐empowerment strategies within systemic interventions, like KTFT, may close remaining gaps in KT access disparities. Moreover, interventions introduced even earlier in the CKD continuum may also meaningfully contribute to reduced disparities. For example, timely referral to nephrology, early nephrology engagement, and education about kidney failure treatments and KT options may address inequities that manifest long before KT evaluation begins [50].

4.1. Limitations

In the present study, we did not conduct a randomized‐controlled trial, or include a comparison group not exposed to the KTFT program; therefore, we cannot make causal inferences regarding any of the observed associations. Baseline measures were collected once, prior to the initial KT evaluation appointment. Past research suggests sociocultural factors may evolve over time [18], so it is possible we were limited in our ability to detect associations with longer‐term outcomes like KT receipt. Additionally, baseline measures were either self‐administered or obtained with the help of an interviewer. As such, responses were subject to social acceptability bias, either due to the presence of the interviewer, or for fear their responses would influence their candidacy for active waitlisting—despite the fact participants were assured their responses would not affect their healthcare. This study was conducted at a single KT clinic in the United States, thus limiting generalizability of findings. The Other race/ethnicity group was small (7% of sample) and heterogeneous, so findings regarding this group should be interpreted with caution. Further, we did not examine race/ethnicity‐by‐sociocultural factor interactions, warranting future examination. Nevertheless, the present study had a diverse sample of patients from urban and rural backgrounds, and we were able to examine multiple KT outcomes prospectively, and over long‐term follow‐up.

4.2. Conclusions

In this prospective observational cohort study of KT candidates who experienced a fast‐tracked approach to KT evaluation, we observed that medical mistrust was inversely associated with active waitlisting, and identified racial/ethnic differences in LDKT and DDKT receipt. Our analytic approach accounted for a broad range of relevant factors, including psychosocial, donor recruitment and preferences, and transplant knowledge and concerns, offering a comprehensive evaluation of factors associated with KT access. These findings suggest that although clinic‐level interventions like KTFT are critical to address structural barriers, additional measures may be necessary to reduce racial and ethnic disparities in KT outcomes. Personalized strategies that build trust and address patients’ prior experiences in healthcare at the individual‐ and interpersonal‐levels may supplement systemic efforts. Future research should explore the impact of multi‐level and upstream interventions on KT outcomes.

Author Contributions

Concept and design: Miriam Vélez‐Bermúdez, Mary Amanda Dew, and Larissa Myaskovsky. Acquisition, analysis, or interpretation of data: Miriam Vélez‐Bermúdez, Yuridia Leyva, Yiliang Zhu, Mary Amanda Dew, and Larissa Myaskovsky. Drafting of the manuscript: Miriam Vélez‐Bermúdez and Larissa Myaskovsky. Critical review of the manuscript for important intellectual content: All authors. Statistical analysis: Miriam Vélez‐Bermúdez, Yuridia Leyva, and Yiliang Zhu. Obtained funding: Larissa Myaskovsky. Supervision: Larissa Myaskovsky, Yiliang Zhu, Unruh Unruh, and Mary Amanda Dew.

Funding

This work was supported in part by grants R01DK101715 from the National Institute of Diabetes and Digestive and Kidney Diseases (PI: Larissa Myaskovsky); UL1TR001857 from the National Center for Advancing Translational Sciences; T32HL007736 from the National Heart, Lung, and Blood Institute in part to Dr. Miriam Vélez‐Bermúdez (PI: T. Resta); and C‐3924 from Dialysis Clinic Inc.

Role of the Funder/Sponsor

The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information: ctr70640‐sup‐0001‐SuppMat.docx

Vélez‐Bermúdez M., Leyva Y., Puttarajappa C., et al. “Healthcare‐Related Sociocultural Factors, Racial Disparities, and Kidney Transplant Outcomes in the Kidney Transplant Fast Track Program.” Clinical Transplantation 40, no. 8 (2026): e70640. 10.1111/ctr.70640

Data Availability Statement

Dr. Larissa Myaskovsky had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The data are not publicly available due to privacy or ethical restrictions.