New Tools and a Population‐Based Approach to Improve Kidney Transplant Opportunity and Preemptive Waitlisting

ABSTRACT

Background

While kidney transplantation provides longer survival and improved quality of life, there are numerous barriers to access. Provider‐led discussions and early referrals are critical for improving access. However, reducing variability in referrals and addressing barriers through standardizing documentation of transplant option discussion and systematically tracking CKD care can be impactful.

Methods

We identified advanced CKD patients from an integrated health care system using a population‐based approach. We used new electronic health record (EHR) tools, searchable system‐wide, to require documentation of transplant discussion and evaluated over five years the number of CKD patients within the kidney transplant continuum and individual patient characteristics.

Results

At baseline, we identified 1747 patients with an eGFR ≤20 mL/min/1.73 m². Of those, 554 (31.7%) were in the transplant continuum (referred, evaluated, and waitlisted). After the EHR tool was implemented, documentation of transplant discussion improved to 100% within a few months. At five‐year follow‐up, the transplant continuum population almost doubled from 554 to 974. Those added after the intervention were older (p = 0.003) and more likely to identify as Black (p = 0.006). Of the waitlisted patients, 65% were referred to transplant centers before starting dialysis. For incident CKD members, outside the cohort, the transplant documentation rate and referrals also remain high with the referral volume per year steadily increased from 233 to 369, and kidney transplant volume per year also steadily increased from 89 to 141.

Conclusion

This study demonstrates that a new, yet simple, EHR tool can be impactful for sustainably increasing documentation of transplant option discussion, waitlisted members, and transplant opportunity.

Abbreviations

CKD

chronic kidney disease

eGFR

estimated glomerular filtration rate

EHR

electronic health record

ESKD

end‐stage kidney disease

KPMAS

Kaiser Permanente Mid‐Atlantic States

NDI

neighborhood deprivation index

QI

quality improvement.

1. Introduction

Kidney transplantation offers superior survival and quality of life compared to dialysis for many patients with advanced chronic kidney disease (CKD) [1, 2]. However, there are numerous patient, provider, and health system barriers to accessing kidney transplantation [2, 3, 4, 5]. Older age, non‐White race, and lower socioeconomic status have all been documented as patient characteristics associated with late referral of CKD patients to nephrologists [4]. Poor communication is the most prevalent provider‐level barrier to the early steps of kidney transplantation [6].

Provider‐led discussions on the benefits of kidney transplantation are critical for improving access to transplantation [7, 8, 9, 10]. However, there is variability in the probability of kidney transplant referral for advanced CKD and end‐stage kidney disease (ESKD) patients [11, 12, 13, 14, 15]. There is little data on what steps can be taken to minimize referral variability or systemic ways to identify and document discussion of transplant options for advanced CKD patients in a large population.

Integrated healthcare systems, with standardized processes and comprehensive electronic health records (EHRs), are uniquely positioned to develop quality improvement (QI) initiatives to systematically track CKD patients through the CKD stages and transplant continuum and address provider‐level barriers to transplantation. At Kaiser Permanente Mid‐Atlantic States (KPMAS), we leveraged our large, integrated EHR system to implement a novel EHR tool centered on improving patient‐provider communication to increase rates of kidney transplantation discussion, referral, and preemptive wait‐listing for advanced CKD patients.

2. Methods

Our study included members from KPMAS, an integrated healthcare system serving about 800 000 members in Northern Virginia, Washington, D.C., Suburban Maryland, and the greater Baltimore area at the time of the study period. Our study population included KPMAS adults (ages ≥18 years) with advanced CKD with an estimated glomerular filtration rate (eGFR) ≤20 mL/min/1.73 m². Our cohort was extracted on June 13, 2018 and followed through to June 14, 2023 (Figure 1A,B). Patients with acute kidney injury, a history of kidney transplant, and those who terminated membership prior to June 13, 2018, were excluded from the analysis. The KPMAS EHR, a comprehensive KP HealthConnect Epic‐based EHR, was used to abstract data on patient characteristics and outcomes through follow‐up. The study was approved by the KPMAS Institutional Review Board.

FIGURE 1.

Baseline and progression of cohort of patients with eGFR < 20 mL/min/1.73m². (A) Baseline cohort as of June 2018. Shown is the pathway for identifying our study cohort and the baseline transplant discussion status and transplant continuum stages of our cohort. (B) Progression of study cohort from June 13, 2018 to June 14, 2023. Shown is the progression and final waitlist and transplantation status of our cohort at our final study time point.

2.1. Intervention Description

We began by capturing all laboratory results documented in the Epic EHR for patients within the KPMAS system, regardless of venue (office, emergency room/primary care physician, etc.), updated every 2 hours, to determine if a patient had CKD. This allowed a population‐based approach to CKD care. For all patients with an eGFR <30 mL/min/1.73 m² and no contact with a nephrologist in the past, a message was sent to the primary care physician (PCP) to place a Nephrology referral. This allowed for outreach that was not dependent on individual providers to acknowledge and assess CKD care, but instead a population‐based approach to CKD care. With the CKD registry in place, for those seen and followed by a nephrologist, each nephrologist had a dashboard, obtained through “care team,” that he or she could run any time, showing their list of CKD members with their latest eGFR. While previously documentation of transplant discussion was done for some patients in the main body of a progress note, there was no way to systematically assess transplant discussion for all members with advanced CKD.

The next step was to create a menu (smartform) within the Epic EHR to allow for discrete data evaluation for new inputs involving standardizing CKD care. For each patient with CKD stage 4/5/ESKD on their problem list, we included the kidney replacement options that the patient had selected (Figure S1), which enabled us to assess option documentation in a searchable way system‐wide. The EPIC smartform was rolled out and has stayed in place since early 2018, and training to nephrologists and kidney case managers on how to document was conducted in early 2018. A list of the patients with eGFR ≤20 mL/min/1.73 m² missing documentation of transplant discussion, both for the cohort and incident CKD members added after the cohort, according to the new EHR tool, was sent to each provider monthly with a reminder to document transplant option discussion and refer eligible candidates. Initially this was done manually, but later our team implemented a best practice alert (BPA) with the primary nephrologist‐facing prompt on the EHR screen to discuss kidney transplantation options and document the outcome of the discussion.

All patients with eGFR ≤20 mL/min/1.73 m² attended a Kidney Replacement Options Class. In this class, patients discussed the benefits of kidney transplants, home dialysis options, and having a life care plan, among other topics. After the class, the kidney case managers and nephrologists documented the dialysis modality and chosen transplant option. We use the patient selection criteria from our local transplant centers we refer to and our national guideline. Even if patients do not meet selection criteria, if they are interested in kidney transplant, they are given options to be referred and to have a determination by the transplant centers. Patients are also offered a second opinion at another transplant center when declined by one center.

With all this data stored, we were able to create a dashboard that would show which patients had no transplant option discussions, which patients were lacking referral, or which patients were ready for reevaluation. The CKD dashboard for the population‐based approach was operational for several months, and the EHR tool for documenting transplant discussion for a few months during the training and education of nephrologists and kidney case managers on how to document, prior to the cohort extraction in June 2018.

2.2. Patient Characteristics

Patient characteristics were obtained from the Epic EHR. Patient demographics included age, sex assigned at birth (male versus female), and self‐reported race (Black, White, Asian, or other). Other race includes Native Hawaiian/Pacific Islander, American Indian/Alaskan Native, multiracial, or members of unknown races. Patient comorbidities and clinical history included blood type (A, B, O), primary cause of ESKD (diabetes, hypertension, glomerulonephritis, cystic kidney diseases and others including congenital uropathy, kidney cell carcinoma, sarcoidosis, etc.), body mass index, eGFR at the time of referral, eGFR at the time of wait‐listing, and dialysis duration prior to transplant. Census tract‐level neighborhood deprivation index (NDI) scores were included as a proxy for a patient's social status.

Data on the patient's transplant process includes dates and status of first referral, first evaluation completion, first placement on the waitlist, and first kidney transplant. Days to events were calculated using the referral date to the evaluation start, waitlist placement, or transplantation.

2.3. Statistical Analysis

Descriptive statistics were used to report the patient characteristics for the overall sample and were stratified by preemptive waitlist status. Comparison tests were used to compare patient characteristics across preemptive waitlist status; Wilcoxon rank sum and Fisher's exact test were used for continuous and categorical variables, respectively. Complete case analysis was used to account for missing data and numbers in the continuous variables provided in the text or in the tables for the individual variables. We employed a complete case analysis methodology, assuming that the absence of data was completely at random, based on our clinical workflow and the standardized documentation of EHRs. We opted not to implement any imputation methods to handle missing data. Patients with missing values for certain characteristics were excluded from that particular descriptive analysis; however, they were retained in the overall cohort unless otherwise specified. All statistical analyses were conducted using SAS 9.4 (Cary, NC) with a p‐value less than 0.05 considered statistically significant.

3. Results

From our 800 000+ members within KPMAS at the time of study period, we identified a cohort of 1857 patients with eGFR ≤20 mL/min/1.73 m² as of June 13, 2018. After excluding patients with prior kidney transplant and those whose membership had terminated, the study cohort included 1747 members (Figure 1A). Of those, 554 (31.7%) were in the transplant continuum (referral, evaluation, and waitlisted). Of those in the transplant continuum, 225 (12.9%) members were waitlisted, and 83 (4.8%) were preemptively waitlisted. Among those not referred, 1016 (58.2%) did not have documentation of the transplant option discussion (Figure 1A). In three months after the EHR tool was implemented, the population without a documented transplant discussion improved from 1016 to only 59. At five‐year follow‐up, the cohort population in the transplant continuum almost doubled from 554 to 974, with 420 patients added after the intervention (Table 1, Figure 1B).

TABLE 1.

Descriptive statistics of KPMAS adult advanced CKD cohort in the transplant continuum at baseline and five years after the new EHR tool implementation as of June 2023.

	At baseline N = 554		After 5 years N = 420		Fisher's exact Test p‐value
Discrete variables	N	%	N	%
Sex assigned at birth
Female	230	41.5	177	42.1	0.8443
Male	324	58.5	243	57.9
Race and ethnicity
Asian/Native Hawaiian/Pacific Islander	68	12.3	34	8.1	0.0055
Black/African American	359	64.8	293	69.8
Hispanic/Latino	68	12.3	36	8.6
White	48	8.7	54	12.9
Other/unknown	11	2.0	3	0.7
Primary spoken language
English	478	86.3	386	91.9	0.0058
Other	76	13.7	34	8.1
Marital status
Married/domestic partner/common law	96	17.3	81	19.3	0.5515
Divorced/widowed/separated	23	4.2	13	3.1
Other/unknown	435	78.5	326	77.6
BMI closest to referral, kg/m2
17–24.9	99	17.9	88	21.0	0.0738
25–30	188	33.9	117	27.9
>30	259	46.8	213	50.7
Unknown	8	1.4	2	0.5
Continuous variables	Median (N)	IQR	Median (N)	IQR	Wilcoxon rank‐sum test p‐value
Age at time of waitlisting, years	58 (373)	50–66	62 (77)	55–70	0.0032
eGFR (ml/min/1.73 m2) closest to referral by dialysis status at referral
Dependent on dialysis	NA (225)	NA	NA (254)	NA	NA
No dialysis	16 (329)	13–18	16 (166)	12–18	0.6319
eGFR (ml/min/1.73 m2) closest to waitlist	13 (138)	10–17	14 (27)	11–16.5	0.7157
NDI, median	−0.22 (544)	−0.69–0.37	−0.10 (420)	−0.68–0.49	0.0655
Stages in the transplant continuum, median
Members with referral but no evaluation date, frequency	65	11.7%	176	41.9%
Members with referral and evaluation but not waitlisted, frequency	121	21.8%	169	40.2%
Time from referral to evaluation start, days	212 (489)	128–425	176.5 (244)	104.5–358.5	0.0251
Time from evaluation start to waitlist, days	124.5 (368)	43–303.5	154 (75)	72–267	0.1784
Time from waitlist to surgery, days	1066 (210)	455–1679	228 (27)	42–530	<0.0001

We compared the characteristics of patients before (N = 554) to after intervention (N = 420) (Table 1). Those added after the intervention had a higher age at the time of wait‐listing (p = 0.003) and were more likely to identify as Black/African American (p = 0.006). This population also trended towards a higher BMI (p = 0.07) and a higher NDI (p = 0.07). Importantly, the time from referral to evaluation was shorter after intervention (p = 0.025), as was the time from waitlist to transplantation (p<0.0001). For incident CKD members, outside the cohort, the documentation of transplant option discussion remains very high with referral volume per year steadily increased from 233 in 2018 to 369 in 2024, and renal transplant volume per year also steadily increased from 89 in 2018 to 141 in 2024. A comparison of the characteristics of preemptively waitlisted and non‐preemptively waitlisted patients is provided in Tables S1–S2. Of the waitlisted patients, 65% [(168 + 139)/476)] were referred to transplant centers before starting dialysis. However, only 168 members were waitlisted preemptively, and 139 members started dialysis before they were waitlisted.

4. Discussion

This QI initiative offers opportunities for every advanced CKD patient, strategically identified using a population‐based approach, rather than being dependent on individual providers to have a discussion of transplant options. Following implementation of the new EHR tool, the rate of documentation of transplant option discussion increased significantly, and the number of members in the transplant continuum almost doubled. Moreover, for incident advanced CKD members added after the cohort, the referral volume as well as documentation of transplant option discussion rate and transplant opportunities continues to be high after the intervention. This study demonstrates that a new, yet simple, EHR tool can drastically improve transplantation metrics for CKD patients. Similar to our findings, Farouk et al. showed a simple, scalable intervention can significantly improve kidney transplant referrals in outpatient settings [16]. Many other studies have focused on timely referral for ESKD patients on maintenance dialysis [2, 17, 18]. Like these other studies, our EHR tool benefits from its simplicity of implementation and ease of use for physicians, suggesting that our approach could be replicated by other health systems by leveraging existing tools to systematically address transplantation barriers.

Moreover, our intervention added more older and Black/African American members. Studies have shown barriers to referral for African Americans [4, 19]. Bolognese et al. [20] found Black patients were referred for transplant later, while Kutner et al. [21]. concluded that early kidney transplant discussion appeared to reduce barriers to Black patients' waiting list placement before the start of dialysis. Similarly, Bartolomero et al. also found age to be one of the factors considered by nephrologists in excluding patients from referral [22]. In our study, advanced CKD patients were strategically identified using a population‐based approach, rather than being dependent on individual providers. This population‐based approach and the new EHR tools have helped to address, partly, age and racial disparity. Our intervention also added higher trending BMI and NDI but did not reach statistical significance, and future work will address closing remaining gaps in care and disparities.

For incident CKD members, outside the cohort, the transplant documentation rate and referrals also remain high since the intervention with steadily increasing volume of transplanted members each year. This indicates that the tool had a lasting positive impact on provider behavior and patient outcomes.

Although our study was limited to KPMAS, our large cohort size, five‐year study period, and sustained performance for incident advanced CKD members, beyond the cohort, demonstrate the rigor and effectiveness of this tool. Our analysis benefits from the granularity obtained from referral to evaluation to wait‐listing and through the transplant continuum. This study demonstrates that a new, yet simple, EHR tool can be developed using existing frameworks and have an impactful, sustainable improvement in documentation of transplant option discussion, waitlisted members, and transplant opportunity.

Author Contributions

Research idea and study design: Aklilu A. Yishak, Atif Jensen; data acquisition: Aklilu A. Yishak, Allison V. Tomimatsu, Megha S. Bhatnagar, Karishma N. Mohadikar; data analysis/interpretation: Aklilu A. Yishak, Megha S. Bhatnagar, Jennifer C. Gander, Karishma N. Mohadikar, R. Clayton Bishop; statistical analysis: Megha S. Bhatnagar, Karishma N. Mohadikar, Jennifer C. Gander; supervision: Aklilu A. Yishak, Rebecca LaLonde, Adrienne N. Deneal, Ronald Potts.

Funding

This study was supported by Kaiser Permanente National Transplant Services & Mid‐Atlantic Permanente Medical Group. The funders had no role in the study design, data collection, analysis, reporting, or the decision to submit for publication.

Conflicts of Interest

The authors declare no relevant conflicts of interest.

Supporting information

Supplementary File 1: ctr70465‐sup‐0001‐SuppMat.docx

Yishak A. A., Bhatnagar M. S., Tomimatsu A. V., et al. “New Tools and a Population‐Based Approach to Improve Kidney Transplant Opportunity and Preemptive Waitlisting.” Clinical Transplantation 40, no. 2 (2026): e70465. 10.1111/ctr.70465

Data Availability Statement

Deidentified data used in this study are available on reasonable request from the corresponding author with an appropriate data use agreement.