results from the closed-ended questions are presented in Maass et al. [13].The paper survey was administered between March 23 and October 1, 2015 to kidney transplant recipients and providers. The survey included both closed- and open-ended questions reflecting four main areas: provider-patient relationships; general and transplant-related health; elements of clinical care; and affect and well-being. The three open-ended response questions were: “How has your life changed since having a kidney transplant?”; “What concerns you most about your health care and future quality of life?” and “Is there something we didn’t ask you that you wish we had? If so, please tell us what it is.”To participate in the study, patients were required to have had a primary kidney-only transplant at 18 years of age or older at UM; to be at least 1 month post-transplant; to have a currently functioning graft; and to not have had transplants of any other organ nor an additional kidney transplant prior to survey enrollment. Kidney transplants may have come from either living or deceased donors. Patients were recruited in-person during regular follow-up visits to the UM transplant center, a convenience sampling approach [14]. Each participant could read and understand English and provided written informed consent. Researchers were not present while participants completed the survey. Family members could have been in the room when the patient was consented and began the survey, and providers may have entered while participants were working on the survey.Demographics and transplant-related data were collected through retrospective chart review. Some demographic data from this study have been published elsewhere [13]. Reporting of race and ethnicity conformed to US federal guidelines [15, 16]. These data were used to compare participants who responded to any of the open-ended questions to those who left these questions blank. We adapted the COREQ guidelines to a qualitative survey analysis [17]. We report this alignment in Additional file 1.The analysis of the open-ended questions was conducted using methods from the grounded theory framework [18]. Themes of patient responses were developed through cascading analysis of responses, rather than through an a priori construct. In several steps, we iteratively reviewed patient responses and generated thematic categories based on the responses. After the final coding was completed, we combined categories into major themes.Specifically, we first selected a random sample of 30 responses for each of the three questions. Three research staff members (ESG, HS, and SMC) each reviewed the responses for two of the three open-ended questions and developed initial response categories.Next, the three research staff members reviewed all of the responses to the open-ended questions and identified additional response categories. The project team reviewed the revised list of categories, and the team combined overlapping categories to produce a preliminary list of ten, nine, and six categories for each question, respectively.Using the preliminary lists of categories, all of the responses were coded (0 or 1) in Microsoft Excel to record whether they corresponded to each category, where 0 indicated no, and 1 indicated yes [19, 20]. Categories were not mutually exclusive; a response could correspond to multiple categories. Each question was coded by two independent reviewers (ESG, HS, SMC, or ELT), and inter-coder reliability was assessed at 94% or higher. During the coding process, new categories were proposed for each of the questions. After further team discussion, the new list of categories included 13, 11, and 13 categories for each question, respectively.The responses to each of the questions were next re-coded according to the new list of categories. Each question was coded by two independent reviewers (HS, SMC, or ELT). No further categories were proposed, and the categorization was considered to have reached saturation [21, 22]. Initial discrepancies between coded responses were resolved through review by a researcher not involved with the coding (KLM). Where there was further uncertainty, discrepancies were resolved through consensus between two team members (KLM and ELT). Researchers (ESG, HS, SMC, and ELT) then identified representative quotations for each category for each question. Through a team discussion and review of the quotations, the categories were synthesized into a smaller number of overarching themes for a total of 6, 6, and 4 themes for each of the three questions, respectively. Participants did not provide feedback on the findings.The response rate to the full survey was 66%, and demographic characteristics were similar between those who consented and those who declined to participate. Participants in the survey were older, transplanted more recently, more likely to be male and have received a living unrelated donor transplant compared to the UM kidney transplant population that met eligibility for the study. A comparison of the demographic characteristics of participants who responded to the open-ended questions to those who did not is shown in Table 1. Participants who answered the open-ended questions versus those who did not were younger at the time of transplant (mean age 48.3 vs. 52.0, p = 0.05), were more likely to