Racial disparities in kidney transplants is a public health issue that has persisted for more than 50 years. As compared with White kidney recipients, African American kidney recipients have nearly twice the risk of graft loss at 5 years after transplant. Despite recent data demonstrating modest improvements in this disparity, a kidney transplanted today is still expected to function about half as long in African American people [1-3]. Marginal improvements in African American recipients are due to improved access to transplants, stemming from organ allocation policy changes and reductions in early acute rejection through a better understanding of immunological risk and the use of potent immunosuppression [4-6]. However, recent studies from our team and others demonstrate that posttransplant outcome disparities in African American people can predominately be explained by issues that arise late (≥2 years) after transplant [7-9]. These issues include late medication nonadherence leading to high tacrolimus variability and rejection and poor control of diabetes and hypertension [7-13]. In models adjusting for these late issues, the risk of graft loss in African American recipients was reduced by up to 75%, as compared with non–African American recipients. [7-11].

For the past decade, our research team has focused on improving posttransplant outcomes and reducing disparities in African American kidney recipients by developing and using mobile health (mHealth) and telehealth interventions [14-20]. Our mHealth app and intervention are founded on the self-determination theory and are designed to improve patient autonomy, competence, and relatedness [21-23]. This is coupled with the use of pharmacist-led motivational interviewing (MI) delivered during televisits, which is designed to improve care and identify and address barriers preventing optimal adherence and comorbidity management, including medication costs [24-27]. Our mHealth-based telehealth intervention improves access to care through remote patient monitoring and medication therapy management, addressing structural barriers, racial bias, and therapeutic inertia that induce disparities [28-30]. Through robust preliminary research, we developed, refined, and tested the functionality, acceptance, and durability of our intervention [14-17]. We completed a 60-patient (39/60, 65% African American) pilot study (National Institute of Diabetes and Digestive and Kidney Diseases [NIDDK] K23DK099440), demonstrating significant improvements in the control of diabetes and hypertension through our mHealth intervention; improvements in hypertension control were more substantial in African American recipients [16]. In 2020, we completed a 12-month randomized controlled trial (RCT; Agency for Healthcare Research and Quality [AHRQ] R18HS023754) in 136 patients (87/136, 64% African American) that used our mHealth, remote patient monitoring, pharmacist-led telehealth intervention to improve medication adherence and reduce hospital readmissions and costs by over 40% [14,20]. These 2 clinical trials demonstrated that our intervention is highly accepted in African American people; 55% of our kidney transplant population is African American, while 65% (88/136) of study participants were African American. Dropout rates were 3% [16,20]. An economic analysis demonstrated that the intervention provided a net cost savings of US $368,839 through reduced hospitalizations and had a return of investment of US $4.30 for every dollar spent [31].

The next clear step in our research trajectory is to conduct a long-term, large-scale study and assess whether this 4-component health services intervention can improve health outcome disparities in African American recipients of kidney transplant. Thus, we propose an RCT to use this system to deliberately address posttransplant disparities in African American people entitled the “Multifaceted Intervention to Improve Graft Outcome Disparities in African American Kidney Transplants (MITIGAAT)” study. The overarching hypothesis for the MITIGAAT study is that late nonadherence and suboptimal control of diabetes and hypertension are more common in African American recipients of kidney transplant and are major contributors to health disparities. A multimodal intervention that addresses these issues will significantly reduce health care disparities among African American recipients of kidney transplant.

This is a single-center, 24-month, 2-arm, 1:1 RCT involving 190 participants (95 in each arm), with roughly 55% (100 participants) being African American. All patients will be recruited from the Medical University of South Carolina (MUSC), in Charleston, South Carolina (ClinicalTrial.gov NCT06023615). This study was funded by the National Institute of Health’s (NIH) NIDDK. We will also conduct a retrospective longitudinal cohort study, comparing a large contemporary national cohort of adult veteran kidney recipients transplanted between 2015 and 2021 with our RCT intervention arm. This study meets all guidelines set by ClinicalTrials.gov and has been approved by the local Institutional Review Board. The statistician conducting the analyses will be blinded to the treatment arm (semiblind). The randomization will occur through computer-generated random numbers, with no stratification, in blocks of 10. REDCap (Research Electronic Data Capture; Vanderbilt University) will be used for computerized randomization and arm assignment, which will be done by a study team member.

The primary outcome of this study is to determine the impact of multilevel health services intervention on achieving improved adherence to tacrolimus. In addition, we will determine the impact of multilevel health services intervention on blood pressure (BP) and glucose control (in those with diabetes). We will conduct a cost-benefit analysis, assessing the estimated costs of hospitalizations and emergency department (ED) visits compared with the cost to deliver interventions. Finally, we will track the incidence of acute rejection, graft loss, and death in the intervention patients; compare our data with a large national cohort of veteran patients of kidney transplant; and assess for racial disparities for these health outcomes. We will enroll both African American and non–African American people to allow us to directly compare these groups across treatment arms to assess disparities.

Patients will be recruited from the MUSC kidney transplant roster. Adult (≥18 years old at the time of transplant) recipients of kidney transplant having a posttransplant period of more than 2 years who meet the eligibility criteria will be approached by researchers for consideration of participation. After discussing the details of the study, providing time, and gaining informed consent, patients will be randomized in a 1:1 fashion (blocks of 10) to either the enhanced control group (usual care and attention control) or the intervention group (usual care and 4-level intervention [Table 1]). For the aim-4 cohort study, we will include all adult solitary kidney recipients transplanted between 2015 and 2021, with more than 2 years of follow-up, to match the RCT. Complete discussion of this follow-up cohort study is beyond the scope of this RCT methodology paper.

Data capture will be accomplished using electronic extraction, manual chart abstraction, and through direct patient interviews (when necessary) by a study coordinator. Data will be collected by conducting a thorough review of the patient’s medical records with patient phone calls at 3-month intervals. Data collection will include all baseline donor or recipient demographics and transplant characteristics as well as medications, laboratory data, pathology, and clinical events that occur during the study. Data will also be captured through the web-based portal and linked to abstracted data within the research database. The REDCap system will be used for data management. REDCap is a secure, web-based application designed exclusively to support data capture for research studies, which we have used in our previous clinical trials. REDCap provides (1) an intuitive interface for data entry (with data validation); (2) audit trails for tracking data manipulation and export procedures; (3) automated export procedures for seamless data downloads to common statistical packages (SPSS [IBM], SAS [SAS Institute], Stata [StataCorp], and R [R Development Core Team]); (4) procedures for importing data from external sources; and (5) advanced features, such as branching logic and calculated fields.

This study is powered to detect clinically meaningful and statistically significant improvements in medication adherence (aim 1), systolic BP (SBP) control (aim 2), diabetes mellitus (DM) control (aim 3), and graft survival (aim 4) while also demonstrating significant reductions in African American disparities for these end points (overarching aim). The analyses take full advantage of all measurements captured during the entire 2-year study using repeated measures methodology. For tacrolimus variability (aim 1), there will be 9 assessments (baseline and every 3 months), aligning with when laboratories are measured. Over the 2-year study, a 5% trajectory difference between arms for tacrolimus variability is considered clinically meaningful and is feasibly achievable given our pilot data (time×treatment). Tacrolimus concentration coefficient of variation (tac CV) has an expected SD of 8.5 with an approximate Gaussian distribution; thus, meeting normality assumptions. A sample size of 87 per group (N=174) achieves >0.999 power, assuming both arms start with a tac CV of 35% and change over the 2-year study to 32% in the control arm and 27% in the treatment arm. The study is powered at 0.800 to specifically test for reductions in racial disparities between the intervention and control arms, using a 3-way interaction term in the model (time×treatment×race). For BP and glucose control (aim 2), we will capture all home- and ambulatory-based measures aggregated into monthly assessments, totaling 25 (baseline and once monthly for 24 periods). We have 0.998 power to detect a difference in SBP changes between arms, assuming a mean SBP of 135 mm Hg in both arms at baseline and a 5 mm Hg reduction in the treatment arm, using 25 repeated measures (monthly). SBP has an expected SD of 9 with a Gaussian distribution; thus, meeting normality assumptions. We have 0.800 power to test for reductions in racial disparities between the arms for SBP, using the time×treatment×race 3-way interaction term. We have 0.998 power to detect a difference between arms for glucose changes, assuming a mean baseline glucose of 160 mg/dL in both arms, 50% (87/174) of patients having DM at the time of randomization, and a trajectory difference in glucose of 24 mg/dL between arms (25 repeated measures). We have 0.803 power to test for significant reductions in racial disparities between the 2 arms for glucose control, modeled using the time×treatment×race interaction term. Mean glucose has an expected SD of 10.8 and distribution approximates Gaussian; thus, meeting normality assumptions. To account for dropouts and censoring events, we will increase each arm sample size to 95, totaling 190 participants. This is an inflation of approximately 10%, which is significantly above the expected totals for dropouts (estimated to be between 2% and 3%), and power reductions due to censoring events.

For aim 4, this study has adequate power to detect a statistically significant difference between the intervention arm and the national contemporary Veterans Affairs (VA) cohort for graft survival. Based on our previous research using the national VA database, we expect to identify roughly 8500 kidney transplants from the VA transplanted between 2015 and 2021, representing a contemporary cohort (33% of which are expected to be African American). Annual graft loss attrition rates are expected to be 5% in the VA cohort (with a mean of 7-year follow-up time) and 3% in the RCT treatment arm (with 2-year follow-up time). Given these, we will have 0.780 power to detect a significant difference in graft survival between the VA national cohort and our intervention patients. To assess if the intervention reduced racial disparities in graft loss, we would compare the African American persons in the RCT treatment arm to the roughly 2800 African American persons in the national VA cohort. We expect annual graft attrition rates to be 7% in the VA cohort (7-year mean follow-up) and 4% in the treatment arm (2-year follow-up). Given these estimates, we have 0.812 power to assess if the treatment intervention significantly reduced racial disparities for graft survival, as compared with the national contemporary VA cohort. These power analyses are valid across different data distribution assumptions (Gaussian and log-transformed). Power analyses were conducted using SAS 9.4 (SAS Proc GLMPower for aims 1-2; SAS Proc Power; two-sample survival test=log rank for aim 4; [SAS Institute]).

Patients randomized to the intervention arm will be provided the same usual care as the control group. In addition, these participants will receive comprehensive supplemental remote monitoring and follow-up by using our smartphone-enabled mHealth app and dashboard, integrated with home-based monitoring of BPs and glucoses, and pharmacist-led scheduled televisits Figure 1. Participants in this group will be provided with a smartphone and data plan if they are not current owners of a device that is compatible with the mHealth app (iOS [Apple Inc] or Android [Google]). All will also be provided with a Bluetooth-enabled, automated, cuff-style bicep home BP monitor. Those with diabetes will be provided a Bluetooth-enabled glucometer with testing supplies. On the mobile device, our mHealth app will be installed which displays the patient’s medication list and alerts them when it is time to take each medication, requiring them to respond by a push button when they have taken the specified medications, providing a time stamp of intake as part of a multi-method adherence tracking system [14]. The intervention will include pharmacist-led telemonitoring of patient adherence to medications, appointments, BP measures, glucose readings in DMs, and electronic health record (EHR) information (tacrolimus variability, appointments, and insurance status) and 21 scheduled telehealth visits with patients that cover specific topics.

The data safety monitoring plan (DSMP) will include the use of a safety officer and the MUSC Institutional Review Board to monitor study-related safety, clinical outcomes, and potential adverse events. In addition, the DSMP will use the study statistician to review the data generated by the MITIGAAT study, ensure data integrity, and assess the potential for futility. Summaries of adverse event reports and safety concerns raised by the safety officer will be made to the NIH in yearly progress reports unless the nature of a particular event is such that it bears reporting to the NIH immediately. The designated safety officer for the MITIGAAT study is a well-experienced transplant physician who is not directly involved in the intervention component of the study. The designated statistician responsible for data oversight and creating the reports needed for the DSMP meetings is an experienced biostatistician with knowledge in monitoring clinical research data integrity. Both the safety officer and the biostatistician will coordinate data review and analysis and communicate with the study principal investigator and the coinvestigators. The functions of the designated safety officer are to (1) provide scientific oversight, (2) review all serious adverse effects or complications related to the study, (3) monitor accrual, (4) review summary reports relating to compliance with protocol requirements, and (5) provide advice on resource allocation.

This study was submitted to the MUSC Institutional Review Board (IRB) for a full board review. The IRB reviewed and approved the protocol and reviewed consent forms to ensured protection of patient privacy and safety. The study was approved on April 18, 2023, under ID Pro00127666. Potential participants provide informed consent before participation and can opt out at any time. In order to protect participants against any risk regarding loss of personal information, all obligations under the HIPPA will be met. Any data or information shared for dissemination will be deidentified and the confidentiality of all participants will be strictly maintained. The only persons with access to protected health information will include study investigators, research coordinators, and those approved by the IRB of record. All data will be secured on university servers, behind firewalls, with passwords protecting entry in these systems. All protected health information will be obtained and managed in accordance with the HIPAA Privacy Rule (45 CFR Parts 160 and 164).

All participants will be compensated US $120 for their data use.

For the primary outcome (aim 1, tacrolimus variability), repeated measures analyses will be conducted using generalized linear mixed modeling (GLMM). We will assess for tac CV distribution and use the appropriate link function and data transformation, if necessary, to ensure model assumptions are not violated. GLMM is the ideal model to use for this clinical trial as it provides robust estimates, takes full advantage of longitudinal data captured during the entire 2-year study, and accounts for a wide array of continuous data distribution types (eg, Gaussian, log-normal, exponential, and gamma) and discrete data distribution types for secondary outcomes (eg, Poisson, binomial, and negative binomial). Assessments for tacrolimus variability will be made every 3 months, aligning with when laboratories are typically drawn. For treatment effect estimates, both unadjusted (raw) and multivariable adjusted modeling will be conducted in an iterative manner. The time×treatment interaction term will serve as the primary measure of efficacy, which estimates the longitudinal trajectory difference between arms over the entire 2-year study. A time×treatment×race 3-way interaction term will be used to assess for changes in racial disparities, as this term estimates if the longitudinal trajectory changes between arms differ by race. Statistical significance for this term indicates that the treatment effect significantly differs by race (African American vs non–African American).

For secondary outcomes, which include BP and glucose (in those with diabetes), data will be aggregated into 25 repeated measure assessments (baseline and monthly for 2-year study). GLMM methodology will also be used for these outcome analyses. The primary efficacy measure for all secondary outcomes will be the time×treatment interaction term estimates (measure of difference in the slope between arms) from GLMM models. To specifically test for the impact of the intervention on racial disparities, the 3-way (time×treatment×race) interaction term will again be used. For both the primary and secondary outcomes, graft loss and death will be considered censoring events, and analyses will include patient data accrued up until these events occurred. To assess for differences in infection rates by treatment arm and race, Poisson or negative binomial modeling will be conducted. Model estimates for rate ratios and 95% CI will be reported. Acute rejection, graft loss, and death will be reported using Kaplan-Meier survival curves and compared using the Log Rank test. Multivariable modeling will be used to adjust for confounders, using frailty Cox regression analysis. Hazard ratios and 95% CIs will be reported. The proportional hazards assumption is tested using Schoenfeld residual plots, log (−log) plots, and ASSESS PH in SAS.

Given we are capturing a large amount of patient-generated data, missingness may be an issue and we have devised a comprehensive plan to assess and address this. First, we will have all patients in both arms use a smartphone enabled with our mHealth app to capture and automatically send home-based, patient-generated data (BPs, glucoses, and adherence). If the patient fails to send this data, the study coordinator will reach out to the patient and facilitate capture. Second, we will use the study coordinator to conduct manual chart abstraction to reduce missing data for key baseline variables and outcomes (graft loss, death, and acute rejection). Following this, if missing data for any variable is >5%, we will conduct multiple imputation and maximum likelihood methods to address missingness. We will use fully conditional multiple imputation regression, with 20 imputations for each missing value. For outcome assessments in all aims, both unadjusted and adjusted modeling will be performed on each imputation and PROC MIANALYZE will be used for pooling model outputs to obtain estimates, 95% CIs, and P values. We will compare estimates generated from complete-case and multiple imputation analysis. To assess for missing not at random, we will implement sensitivity analysis using the missing not at random tool of PROC multiple imputation. We will report both sets of results (complete case and imputation results) in papers with implications addressed in the discussion [32].

The aims of this study are to improve adherence and control of late clinical issues that are predominant factors for racial disparities in kidney recipients, through a technology-enabled, telehealth-delivered, 4-level intervention. The key clinical issues for this study include tacrolimus variability, BP, and glucose control (in those with DM). We will also assess the impact on the intervention on health care use (hospitalizations and ED visits) and conduct a cost-benefit analysis. Finally, we will assess the impact of the intervention on acute rejection and graft survival rates versus a large contemporary national cohort. The study design and intervention were developed using our formative research, providing insights into interventions that change behaviors, improve access, and address provider bias and structural racism, while being highly acceptable to patients [7-9,11-21].