Thematic Analysis of the Health Records of a National Sample of US Veterans With Advanced Kidney Disease Evaluated for Transplant

Key Points Question What types of clinical care are involved in the kidney transplant evaluation process in real-world clinical settings? Findings In this qualitative study of the electronic health records of 211 US veterans with advanced kidney disease who were referred for kidney transplant evaluation, 4 dominant themes were identified describing clinical care during the evaluation process: far-reaching and inflexible medical evaluation, psychosocial valuation, surveillance over compliance, and disempowerment and lack of transparency. Meaning In this study, clinician documentation in the medical record indicated that, to be considered for a kidney transplant, patients were required to participate in a rigid, demanding, and opaque evaluation process over which they and their local clinicians had little control.

K idney transplant is a preferred treatment for endstage kidney disease. 1,2 However, because there are risks associated with transplant and the supply of donor organs is limited, patients are required to participate in a formal medical and psychosocial evaluation before they can be added to the deceased donor transplant waiting list or can be accepted for a transplant from a living donor. Once selected, patients may spend years on the national waiting list before receiving a transplant from a deceased donor. During this time, they are closely monitored by the transplant center for signs of deteriorating health and/or the development of new contraindications to transplant.
The National Kidney Allocation System relies on an explicit and publicly available algorithm that was developed with input from diverse stakeholders to determine how patients on the waiting list are selected for transplant from a deceased donor. 3 However, the processes by which transplant center teams evaluate and select patients for inclusion on the deceased donor waiting list and for transplant from a living donor are less transparent. [4][5][6] Previous studies describing the transplant process have focused on the factors and outcomes associated with referral, waiting list inclusion, and receipt of a kidney transplant [7][8][9][10][11] and on patients' experiences with particular aspects of the transplant process. 12- 16 To our knowledge, few previous studies have described how the entire process of referral, evaluation, and selection for kidney transplant proceeds in real-world clinical settings. 17 A more complete understanding of this process could help to improve the quality of care for patients and families, guide process improvement, and support public discussion about strategies to promote fair allocation of donor kidneys. 5,18 The US Department of Veterans Affairs (VA) health care system is the only national health care system that supports all steps in the kidney transplant process internally, including referral and evaluation, selection, and surgery. On an annual basis, approximately 1800 patients are referred to 1 of 7 VA kidney transplant centers for evaluation, and approximately 160 kidney transplants are performed within the VA system. 19 An estimated two-thirds of referred patients are invited to participate in an in-person evaluation, and the remainder are declined as transplant candidates without further evaluation. 20 Approximately one-half of patients invited for an in-person evaluation at a VA transplant center are selected for inclusion on the national kidney transplant waiting list. 20 To understand how evaluation and selection for kidney transplant unfolds in real-world clinical settings, we conducted a qualitative study of clinician documentation in the electronic health records (EHRs) of patients with advanced kidney disease who were referred to VA transplant centers.

Cohort Selection and Data Collection
Using a previously published approach, [21][22][23][24] we conducted a qualitative study of clinician documentation in the EHRs of members of a representative national sample of US veterans with advanced kidney disease who were referred for evalua-tion at 1 or more VA kidney transplant centers. The study was approved by the VA Central Institutional Review Board with a waiver of informed consent for all study procedures because of minimal risk to patients.
We identified a national cohort of 130 374 patients with advanced kidney disease (defined as an estimated glomerular filtration rate <20 mL/min/1.73 m 2 on at least 2 occasions at least 3 months apart) between October 1, 1999, and December 31, 2014 (eFigure 1 in the Supplement). We used linked clinical and administrative data from the VA, the Centers for Medicare & Medicaid Services, and the US Renal Data System (a national registry for patients with end-stage kidney disease) to assemble and characterize the study cohort. Patients entered the cohort on the date of their second finding of an estimated glomerular filtration rate of less than 20 mL/min/1.73 m 2 . Patients were followed up until their date of death or the end of the follow-up period on October 8, 2019. We then selected a random sample of 4000 patients who entered the cohort between January 1, 2004, and December 31, 2014, for in-depth medical record review. We obtained all clinician progress notes that were stored in the VA Corporate Data Warehouse between the time of cohort entry and the end of the follow-up period. We used the Veterans Indexed Search for Analysis tool, a Lucene-based text search instrument, 22 to identify all mentions of the term transplant (which included mentions that contained the root word, such as transplantation and transplants) in note titles or free text in patients' medical records between the time of cohort entry and the end of the follow-up period.
A text search of the EHRs of 4000 patients yielded 97 515 notes with at least 1 mention of the term transplant belonging to 1896 unique patients. One of us (C.R.B., a nephrologist and research fellow) reviewed notes with at least 1 transplant mention to exclude the following groups: (1) 481 patients for whom mentions containing the term transplant did not pertain to the kidney transplant evaluation process (eg, patients who had received a kidney transplant before cohort entry or patients for whom the transplant mention was in reference to a different type of organ transplant or appeared only as part of a standard template element in the medical record) and (2) 1204 patients with a valid transplant mention who were not referred to a VA transplant center during the follow-up period. Using this approach, we identified an analytic cohort of 211 patients (5.3%) who were referred to a VA transplant center during the follow-up period.

Qualitative Analyses
One of us (C.R.B.) reviewed all notes with transplant mentions that belonged to the 211 cohort members referred to a VA transplant center and abstracted passages pertaining to the transplant evaluation process from their EHRs. Two of us who are experienced in qualitative methodology (C.R.B. and A.M.O., a physician scientist) used inductive thematic analysis, an approach for analyzing unstructured text that facilitates discovery of previously unidentified factors associated with a phenomenon of interest, 25 to independently review and openly code abstracted passages for concepts associated with the transplant evaluation. One of us (A.M.O.) reviewed all records for a random selection of 55 cohort members who were referred to a VA transplant center, and another one of us (C.R.B.) reviewed records for all 211 patients. We used an iterative and consensus-based approach to define codes and resolve differences in interpretation until we reached thematic saturation, the point at which no new concepts were identified with additional review. 26,27 We then collapsed associated codes into larger thematic categories, returning frequently to the primary passages to ensure that emergent themes were grounded in the data. 25,27 We used ATLAS.ti, version 8 (Scientific Software Development), to organize and store text and to support coding.

Statistical Analysis
SAS, version 9.4 (SAS Institute) was used to assemble the cohort. Stata, version 15 (StataCorp LLC) was used to describe the characteristics of patients referred for transplant evaluation.

Results
Among a national sample of 4000 patients with advanced kidney disease who received care in the VA health care system from January 1, 2004, to December 31, 2014, 211 patients were referred to a VA transplant center during follow-up through October 8, 2019. Of those, 202 patients (95.7%) were male, 118 patients (55.9%) were White, and 75 patients (35.5%) were Black or African American; the mean (SD) age of cohort members was 57.9 (9.5) years ( Table 1). A total of 51 patients (24.2%) had initiated maintenance dialysis by the time of cohort entry. Nine patients (4.3%) were enrolled in Medicaid in the year before cohort entry. Based on qualitative analysis of notes containing transplant mentions, we identified 4 themes related to the transplant process: (1) far-reaching and inflexible medical evaluation, (2) psychosocial valuation, (3) surveillance over compliance, and (4) disempowerment and lack of transparency.

Far-reaching and Inflexible Medical Evaluation
Documentation in patients' EHRs suggested that clinicians approached the transplant evaluation with the goal of fully char-acterizing patients' known health conditions and identifying undiagnosed conditions (example quotations are included in Table 2). As part of the evaluation, patients were expected to attend multiple clinic visits with different specialists and might be referred for invasive procedures, such as cardiac catheterization, liver biopsy, dental extractions, and surgeries, many of which would not have been indicated as part of routine care. Testing conducted as part of the transplant evaluation could involve risks and sometimes led to a cascade of further testing and treatments.
Participation in the transplant evaluation could also have implications for decisions about other aspects of patients' care, Other e 2 (0.9) a Race as documented in Medicare and Department of Veterans Affairs files may not have been based on self-reporting. b Medicaid enrollment at any time in the year before cohort entry was ascertained from Medicare enrollment files. c Referral transplant centers were nonexclusive; 23 patients were referred to multiple centers. d The notes in the electronic health record did not specify the transplant center to which the patient was referred. e The patients visited a different Department of Veterans Affairs transplant center for consideration of dual-organ transplant.
even when these aspects were not associated with the transplant process. Patients and their clinicians might have delayed or avoided receiving tests or treatments because of concern that these might adversely impact the patients' transplant candidacy. In general, the transplant evaluation process made little accommodation for the distinct circumstances, priorities, and values of individual patients. The need for frequent testing and interaction with the health care system during the transplant evaluation process could be overwhelming and emotionally burdensome for some patients and families. Setbacks or deviations from expectations in the evaluation process were sometimes associated with depression and even suicidal ideation.

Psychosocial Evaluation
Interviews conducted by social workers and mental health specialists as part of the psychosocial assessment included detailed questions about patients' living arrangements, employment history, substance use, financial status, and legal history as well as in-depth questions about family members (example quotations are included in Table 3). Ongoing sub-stance use (eg, alcohol, tobacco, and marijuana) constituted an unconditional contraindication to transplant regardless of the circumstances or frequency of the substance use. Clinicians' questions and recommendations sometimes strayed into patients' personal affairs and could be upsetting or embarrassing. Documentation in medical records suggested that patients' awareness of clinicians' conflicting obligations to advocate for them while also being required to report potentially compromising information to the transplant center could stifle honest communication and undermine trust. Referral letters to the transplant center sometimes conveyed a moral tone alluding to whether patients were deserving of transplant.
As part of the transplant process, patients were expected to identify a support person who could care for them during the immediate posttransplant period. Documentation by some transplant specialists suggested a preference for support persons with Gastroenterology consultation note He states that a gastric lesion was seen on MRI and he was to have EGD but declined…he is afraid that his workup for renal transplant will be delayed if anything is found.

Social work transplant evaluation note
We discussed his pulmonary nodule…They recommended CT-guided biopsy, but he initially refused this, as he personally feels it's unlikely that he has cancer, and is willing to assume the risk…[He] wants a kidney transplant, and this issue needs to be resolved before that.

Psychiatric clinic note
The patient presented with a lot of anxiety related to upcoming medical procedures that are prerequisite for a possible kidney transplant…fears regarding "not waking up" from some of these procedures.

Suicide risk prevention assessment
He reported disappointment and exacerbation in depression as a result of being dropped from a list of possible recipients for a kidney transplant. He admitted to suicidal ideation.

Nephrology clinic note
He is an extremely compliant patient, functional and active, very dedicated to his children…He has been pleading with us to refer him for consideration for transplantation as he has 2 young children with no one else to provide for them.

Psychiatry transplant evaluation note
Long-standing supportive marriage and stable financial circumstances are also good predictors of positive outcome.

Nephrology clinic note
This gentleman appears to be in many ways a "loner"…he states he did not marry, and he is averse to marriage now. This does not of course disqualify him from transplant consideration, but I want to be more sure he does have ex-fiancée's support and some insight into this dynamic. formal relationships (ie, a spouse or family member) and adequate financial resources. We found examples of support persons who were prepared to relocate and/or quit their jobs to fulfill this responsibility. We also found examples of patients who could not be accepted as transplant candidates because they did not have family or friends who were willing and able to fulfill this role.

Surveillance Over Compliance
In addition to completing the medical and psychosocial evaluation, patients were expected to demonstrate personal responsibility for improving their health and social circumstances (example quotations are included in Table 4). The extent to which a patient adhered to treatment recommendations could be viewed by clinicians as an indication of the patient's level of motivation to receive a kidney transplant and as a "litmus test" for how the patient would fare after transplant. We also found examples of clinicians encouraging patients to adhere to treatment recommendations by reminding them that adherence could impact their transplant candidacy. Clinicians at both the local medical center and the transplant center routinely reviewed patients' medical records to assess the extent to which they were adhering to clinical recommendations and to identify risk factors for nonadherence. Missed clinic appointments and missed or shortened dialysis treatments were interpreted as indicators of "noncompliance" with little consideration for extenuating circumstances or the values and preferences of individual patients. Nonadherence to prescribed medications raised questions about whether patients would be able and willing to adhere to posttransplant medication regimens. However, clinicians tracked and encouraged adherence to a range of other medical recommendations that appeared to be less relevant to posttransplant care.

Disempowerment and Lack of Transparency
Patients often began the transplant evaluation around the same time that they started or were preparing to start dialysis treatment. When facing the prospect or experiencing the reality of long-term dialysis treatment, some patients expressed a willingness to do anything to avoid having to rely on dialysis for the remainder of their lives (example quotations are included in Table 5).
To help patients move forward with the transplant process, clinicians at the patient's local medical center sometimes had to assume a more technical role, performing tests and administering treatments required by the transplant center even when these did not make sense to them, deviated from perceived best practices, or could be harmful.
Patients, family members, and local clinicians were often unsure or mistaken about the patient's status in the transplant process. Documentation from specialists responsible for discrete aspects of the evaluation suggested that they did not always know how the results of their testing might impact a patient's transplant candidacy. Clinician documentation suggested that patients were sometimes surprised to find that their candidacy had been declined by the transplant center.
Documentation from the transplant center rarely made explicit mention of the team's role as stewards of a scarce resource. Letters to patients and referring clinicians focused instead on the transplant team's assessment of the risk of negative outcomes and described what the team had determined to be in the best interests of individual patients. Transplant center assessments took little account of the priorities of individual patients and their willingness to assume risk. Clinicians' notes in the medical record sometimes described patients' distress at what they perceived to be arbitrary or unfair decisions about their transplant candidacy.

Discussion
Our thematic analysis of the medical records of a national sample of patients with advanced kidney disease who were referred for kidney transplant evaluation provides insights into how this process proceeds in real-world clinical settings. Because kidney transplant was a desirable goal for patients, they had little choice but to participate in an evaluation and selection process that could be inflexible and demanding and over which they and their local clinicians had little control. The need to formally incorporate societal considerations that go beyond the needs of individual patients (eg, fair allocation of a scarce resource) distinguishes the evaluation and selection process for kidney and other types of solid organ transplants from most other clinical care processes. Nonetheless, there is general agreement that this process should strive to uphold patients' goals, values, and preferences to the greatest extent possible. 28 Our analysis suggests that there are opportunities to make the evaluation process more personcentered. Documentation in the medical record suggested that patients and their local clinicians were often unsure about what to expect from the evaluation process or about the rationale for evaluation requirements or selection decisions, which may impede shared decision-making. 12,29 There was also little accommodation for the needs of patients and families and few opportunities for patients' local clinicians to use their clinical judgment and provide individualized care.
Transplant centers are given substantial flexibility to define their own systems for evaluating and selecting transplant candidates with the aim of being responsive to the particular needs of the populations they serve. 28 However, this approach can produce variability in the selection and evaluation processes across centers [30][31][32][33] and limit transparency, accountability, and opportunities for stakeholder input, features that are critical for ensuring a fair allocation system. [4][5][6] The use of subjective and nonevidence-based selection criteria [34][35][36] and the rigidity with which these criteria are applied may have the unintended consequence of leaving patients and families vulnerable to the harms of overtesting, 37 implicit bias, 38,39 and discriminatory practices. 40,41 In particular, we suspect that requirements for complete abstinence from substances (including legal substances), 42 narrow definitions of acceptable social support, 43,44 and insistence on "compliance" with a wide range of medical recommendations 45 may have disproportionately impacted the most disadvantaged members of the cohort.
Our findings highlight the need for a more person-centered and equitable approach to the kidney transplant evaluation and selection process (eFigure 2 in the Supplement). An evaluation and selection process that is more substantially guided by highquality evidence and communally accepted principles of fairness and social justice could limit the burden of irrelevant testing procedures and protect patients from harmful, arbitrary, and/or discriminatory evaluation and selection practices. 37,40,46 More open communication and collaboration between the transplant center and patients' local medical centers could also allow for greater transparency and flexibility in the evaluation process. A collaborative approach could empower local clinicians to individualize the evaluation process within established parameters to accommodate the needs and circumstances of individual patients and families. 47 Greater transparency in the transplant evaluation and selection process could also serve to improve public accountability and trust 5,18 and support shared decision-making by providing patients and their families with more realistic expectations about the evaluation. 47

Strengths and Limitations
This study has several strengths and limitations. The national scope and integrated nature of the VA health care system allowed us to identify a representative sample of patients who were referred for transplant evaluation and to track their involvement from an early stage in the transplant process. However, because of the distinctive organizational structure of the VA transplant program, some of the themes that emerged from this analysis may not be transferable to other health systems or to groups of patients who are not well represented in the VA population, such as women and young adults. Some of our findings (eg, uncertainty about transplant status) are particularly notable because the VA system has an organizational policy regarding kidney transplant evaluation that is more explicit, centralized, and accountable than that of many other health care organizations. 19,48 This study is also limited because not all aspects of clinical encounters were documented in the medical record and medical record documentation does not provide direct information about the experiences or perspectives of patients, families, or clinicians.

Conclusions
This qualitative analysis of documentation in the EHRs of a representative national sample of US veterans with advanced kidney disease who were referred to a VA transplant center suggests that there are opportunities t o m a ke t h e t r a n s p l a nt e v a l u a t i o n p r o c e s s m o r e person-centered and equitable. Our findings highlight the need for a more evidence-based, individualized, and collaborative approach to evaluation and selection for kidney transplant.