To estimate utility values for IgAN health states, we employed a vignette-based approach. This approach was used instead of EQ-5D, which is often preferred by the National Institute for Health and Care Excellence [20], because it was challenging to recruit a sufficient number of patients within a reasonable timeline due to the rarity of IgAN. In addition, EQ-5D might not be sufficiently sensitive to various IgAN health states we aimed to evaluate in this study. Specifically, time trade-off (TTO) interviews based on the developed vignettes were conducted with consenting members of the UK general population. The TTO method asks respondents about the number of years of life in a given imperfect health state they would be willing to ‘trade off’ in order to live in full health [21] and is a well-established method to measure utilities for health states across different disease areas [22].

Rare diseases such as IgAN present challenges in estimating health state utilities using preference-based approaches due to the difficulties in recruiting a sufficiently large sample of patients or caregivers within the study timeframe [23]. In addition, generic preference-based approaches may not be sensitive to certain disease attributes. Thus, vignette-based approaches may be more desirable when a study aims to isolate the utility impact of specific attributes of a disease [23]. In the case of IgAN, we aimed to estimate the utilities of various disease health states related to CKD stages, proteinuria, dialysis, and nephrotic syndrome. Given these, a vignette-based study that uses a general population is the most suitable approach for the current study. It is also a commonly used method in previous literature to derive health state utilities for rare diseases [24].

The study was reviewed and received an exemption determination from ongoing oversight by the Pearl Institutional Review Board according to FDA 21 CFR 56.104 and 45CFR46.104(b)(2): (2) Tests, Surveys, Interviews. Informed consent was obtained from all participating individuals.

We included health states describing various CKD stages (1–5), proteinuria status (defined by proteinuria < 1 g/day or ≥ 1 g/day based on the treatment target in the KDIGO guidelines [4]), dialysis status, and nephrotic syndrome. Because symptoms in CKD stage 1 and 2 are similar [25], the two CKD stages were combined into one health state.

To support study participants’ understanding of each health state, vignettes were developed to describe associated symptoms and QoL (Supplementary Table 1, see electronic supplementary material [ESM]). First, a preliminary vignette was developed for each health state based on information collected in a targeted literature review. Vignettes included descriptions of urine test outcomes, symptoms and their severities, QoL, resource use, and risks of other diseases. To ensure that the vignettes provided accurate and comprehensive descriptions of the health states, feedback on the preliminary vignettes was obtained through one-on-one web-based interviews in the UK. Five experienced nephrologists and one representative from a patient advocacy group participated in the interviews. In addition to having a child with IgAN, the patient advocacy group representative had extensive experience interacting with patients with various types of IgAN and significant research experience in this field.

Participants were recruited through an online panel from members of the UK general public. To be included in the study, participants had to meet the following criteria: (1) at least 18 years or older, (2) able to speak and read English, and (3) able and willing to provide an informed consent.

In an initial survey, individuals who were willing to participate were screened for eligibility and asked to provide informed consent as well as information on demographics, comorbidities, and self-reported general health status on a visual analogue scale (VAS) from 0 (worst imaginable health) to 100 (best imaginable health). Eligible respondents who completed the survey were scheduled for a TTO interview. Only participants who completed both the initial screening survey and the TTO interview were compensated (consistent with fair market value) for their participation in the study. A target sample size of 200 was determined based on a margin of error < 14% of the standard deviation (SD) at a significance level of 0.05.

Because unsupervised online surveys may compromise data quality due to decreased participant engagement, we used computer-assisted telephone interviews (CATIs) to guide participants through the TTO tasks [26]. CATIs were held over telephone/teleconference while a computerized questionnaire was administered via screen sharing between 28 January 2022 and 4 March 2022. Interviewers recorded the respondent’s answers directly into the computer. Respondents could also ask for clarification from the interviewer if they were not clear on a question.

To ensure that participants understood the concept of TTO, they first completed a training task and a practice task. Participants were then randomly assigned to view the vignettes of the IgAN health states (Supplementary Table 1, see ESM) in either increasing or decreasing severity. After reading the vignette for a health state, participants rated the value of the health state on a VAS from 0 (worst imaginable health) to 100 (best imaginable health) and responded to a series of TTO questions illustrated with visual aids (Fig. 1). Inclusion of the VAS rating prior to the TTO task is recommended by the EuroQoL protocols to validate TTO-based results [27]. In the TTO tasks, participants were repeatedly asked to compare living 10 years in an IgAN health state versus living fewer years in full health until they were indifferent between the two options. Health states worse than dead were determined by participants being indifferent between living a certain number of years in an IgAN health state versus immediate death (zero years in full health). Another example of valuing health states worse than dead is presented in Fig. 1B. The point of indifference was used to estimate the number of years of life in the IgAN health state they would be willing to give up in order to live in full health. Given that the design of TTO tasks (e.g., time horizon, structure of the hypothetical lives, iteration procedures, visual aids, and respondent training) influences how participants value health states [28], the TTO questions in this study followed the data collection procedures in the EuroQol Valuation Technology (EQ-VT) protocol [29]. The EQ-VT protocol includes TTO as one of the main valuation tasks and makes recommendations on the data collection procedures based on best practices and empirical research [27]. Following the EQ-VT protocol improves comparability between the utility values from this study and those from EQ-5D, which are commonly used in cost-effectiveness analysis models.

Example visual aids for TTO questions in the survey

Prior to data collection, the initial survey and the TTO tasks were pre-tested among two eligible participants to ensure clarity of the questions and vignettes. In addition, a soft launch was conducted by collecting data from 10 participants to confirm that they interpreted the questions correctly. Feedback from the soft launch was provided to moderators to ensure optimal understanding and interpretation of the TTO questions by participants in the full launch.

Sociodemographic characteristics, comorbidities, and current health status of the participants were summarized descriptively.

TTO utilities ranged from − 1 (worse than dead) to + 1 (perfect health), with the smallest difference being 0.05, and were calculated based on responses to the TTO questions. Specifically, if a health state (h) was considered better than dead, the utility value was calculated as U(h) = x/10, where x denotes the number of years in full health at the point of indifference. If a health state (h) was considered worse than dead, the utility value was calculated as U(h) = (x − 10)/10. The value of health states was also rated on a VAS from 0 (worst imaginable health) to 100 (best imaginable health) and divided by 100 to calculate VAS scores.

To evaluate the impact of proteinuria and dialysis on health utility in patients with IgAN, the utility values of proteinuria < 1 g/day versus ≥ 1 g/day for CKD stage 1 & 2, 3, and 4, as well as the utility values of dialysis versus no dialysis in CKD stage 5 were compared using paired t tests; comparisons were conducted for both TTO utility values and VAS scores.

The primary analyses removed illogical responses, that is, responses in which a participant rated a more severe health state as better than a less severe health state. Sensitivity analysis included all the illogical responses to assess the validity of VAS and TTO ratings. In addition, subgroup analyses were conducted by the order in which participants viewed the health states and by gender.