Patients’ Perspectives Regarding Ankylosing Spondylitis Treatment Administration Features: Evidence from Qualitative Interviews and a Multinational Quantitative Preference Survey

Eligible patients for the qualitative interviews were aged ≥ 18 years, resided in United States (US), had a confirmed AS diagnosis (i.e., patients provided screenshots of their medical records, patient portal, or AS medication), and had a score of ≥ 4 on the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) indicating that AS is not well controlled and that a change in drug therapy may be necessary to reduce disease activity [13]. Those who were currently enrolled in a clinical trial, had a diagnosis of Alzheimer’s disease or dementia, or who were unable to speak English (i.e., interviews were conducted in English) were ineligible to participate. A total of ten patients with AS were recruited to participate in the qualitative interviews. Researchers have noted that data saturation may be attained by as little as 6–12 interviews, depending on the nature of the data being collected [14].

An experienced moderator conducted one-on-one interviews lasting approximately 45 min. A semi-structured interview guide was followed with open-ended questions to identify key attributes that influence treatment preference. Topics explored in the interview included patient experience and patient journey with AS and patient perceptions of AS treatment modes of administration. Qualitative interview data were collected between February 16 and April 1, 2022.

Patients with AS in the US, United Kingdom (UK), and one European country (Italy) were recruited to complete a cross-sectional, web-based, patient preference survey. Eligible patients were aged ≥ 18 years, resided in US, UK, or Italy, had confirmed AS diagnosis using the same verification options to confirm diagnosis used in the qualitative interviews described above, and met one or both of the following criteria: score of ≥ 4 on the BASDAI (i.e., not well-controlled disease) and/or currently taking advanced therapy. These criteria were applied to categorize participants into three groups: on advanced therapy/well-controlled disease, on advanced therapy/not well-controlled disease, and not on advanced therapy/not well-controlled disease. Data were collected between June 8, 2023 and October 5, 2023.

The survey included two preference elicitation assessments: a case 1 best–worst scaling (BWS) exercise [15], as well as six fixed-choice tasks. An object-case (case 1) BWS exercise was used to assess patient preferences for attributes representing the process of AS treatment administration (including attributes of biologic or targeted synthetic disease-modifying anti-rheumatic drugs [bDMARDs or tsDMARDs] administered via oral, subcutaneous [SC], or intravenous [IV] modes of administration). In a case 1 BWS, respondents are asked to rank the most desirable and least desirable items within a subset of a list of items in a series of questions in order to put the relative desirability of each item on a numerical scale [16, 17]. BWS yields a rank ordering of the importance of all items as well as a quantitative measure of the relative importance (RI) of each item. A total of 12 attributes related to the process of treatment administration were included as items in the BWS exercise. The attributes covered four general categories of processes reflecting treatment administration: mode of administration (e.g., oral/pill), frequency of administration (e.g., once a day), treatment accessibility/affordability (e.g., can be delivered to your home), and characteristics of the treatment (e.g., does not require refrigeration). Attributes included in the BWS exercise are shown in Table 1; an example BWS question is shown in Fig. 1. Each BWS question asked: “Assuming you are choosing a treatment for ankylosing spondylitis (AS), which of the following things about a treatment would you like the most, and which would you like the least?” This study followed the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) guidelines on conducting conjoint analysis [18‐20].

Bild vergrößern

The fixed-choice tasks were designed to directly elicit preferences regarding mode and frequency of AS treatment administration. Each fixed-choice task presented two options and asked the individual to report whether Option A is preferred, Option B is preferred, or they have no preference. In the fixed-choice tasks, the frequency of administration shown for ‘pill’ (once a day, twice a day) and ‘self-injection’ (once a month, twice a month) were varied to enable pairwise comparisons between the four frequency and mode of administration options; the instructions for each task told patients to assume that the options presented do not differ in effectiveness or side effects and that any self-injected medication requires refrigeration.

In addition to treatment preferences assessed via fixed-choice tasks and BWS exercise, the survey included other items to characterize the study sample on sociodemographic and clinical characteristics and AS treatment use.

A content and thematic analysis of the interview data was performed. Thematic analysis is a method for identifying, analyzing, and reporting patterns (themes) within data [21]. The verbatim transcripts were uploaded to a qualitative data analysis program (NVivo v12), which enabled an assignment of codes to the emerging themes and facilitated organization of the data. One researcher reviewed the first two transcripts and developed an initial coding system, specifying key themes and sub-themes. The draft code system and respective transcripts were reviewed by a second researcher. The code system was refined throughout the analysis, where applicable. Saturation was tracked and confirmed through a saturation grid, which was used to compare and tally themes elicited during each interview. Information saturation, in which no new key themes emerge with each successive interview, was attained by the last interview.

For patient characteristics, including current treatment use, means and standard deviations (SDs), medians, first and third quartiles (Q1 and Q3), minimums, and maximums were provided when performing descriptive analysis of continuous or discrete (i.e., count) data. Frequencies and percentages were provided when performing descriptive analysis of categorical data.

To estimate RI for each attribute in the BWS, hierarchical Bayesian (HB) estimation was fitted to the BWS data [22]. The HB model yields a coefficient for each item in the BWS for each respondent. These coefficients were assumed to be drawn from a multivariate normal distribution across respondents in the sample (the higher level of the model). The coefficients for a respondent were used in a logit model for the task where one selects the best and worst from a group of items. The best item was chosen in a logistic regression model, so the components of the regression model are: exp(A)/(exp(A) + exp(B) + exp(C) + exp(D)). In the BWS, the worst item is also chosen in a logistic regression model, but the item that was chosen as best is dropped, negating the coefficients. Hence, the components in the regression model for choosing B worst are: exp(-B)/(exp(-B) + exp(-C) + exp(-D)). The estimated coefficients were transformed with the logit and converted into RI scores and standardized to sum to 100 across attributes for reporting. The items were then rank-ordered by RI. Further details on this analytic approach have been described previously [23].

Subgroup variables were derived from survey responses to BASDAI and current treatment. For the treatment/disease status subgroup, advanced treatment/well-controlled disease was defined as BASDAI < 4 and current use of any advanced therapy. Advanced treatment/not well-controlled disease was defined as BASDAI ≥ 4 and current use of any advanced therapy; no advanced treatment use/not well-controlled disease was defined as BASDAI ≥ 4 and no advanced therapy use. The combination of no advanced treatment and well-controlled disease was not included because patients were required to have either poor disease control or be on advanced therapy to be eligible for the study. Sex and age subgroups were also created. Age 42 was used as the cut-off point (aged < 42 years, ≥ 42 years) based on the mean age of patients with AS who had received tofacitinib in a prior randomized clinical phase 3 trial [24]. RI estimates for each attribute across the three treatment/disease control subgroups and the two age subgroups were compared via one-way analysis of variance tests (ANOVAs) and two-sample t tests, respectively. P values < 0.05, two-tailed, were considered statistically significant. Frequencies and percentages were used to summarize fixed-choice task responses for the total sample and by subgroup.

BWS analyses were performed using Sawtooth’s Lighthouse Studio v9.13.1 (Provo, UT, USA). All other analyses were performed using IBM SPSS Statistics v28.0 (Armonk, NY, USA).

Of the ten participants, the mean age was 46 years, and 60% were male and had at least a college degree. Half were currently working, and 30% were on disability or unable to work. Nearly all patients (≥ 90%) see a rheumatologist or primary care provider for their AS. At the time of the study, the mean time since diagnosis was 11 (range 1–22) years; patients reported a mean of 8 (5–20) years from AS symptom onset to diagnosis.

Illustrative verbatim quotes from patients regarding the different impacts of AS are shown in Table 2. Patients most commonly cited pain (70.0%) and symptom unpredictability (50.0%) as the most bothersome disease-specific aspects of AS. Fatigue (40.0%), limited mobility (40.0%) and stiffness (30.0%) were also noted by patients as bothersome aspects of AS. Many (40.0%) also reported experiencing moderate/severe impacts of AS on their ability to work. Attributes related to the process of treatment administration that were bothersome to patients include needing to refrigerate, needing to take daily, and maintaining insurance coverage/concerns about insurance not covering treatment, which were mentioned by one patient each. Patients reported trying a series of different treatments to find relief for their symptoms, and when they find something effective like injectables, they preferred to stay with that treatment type.

Patients mentioned a variety of potential advantages and disadvantages of different modes of administration (Fig. 2). Administering injections at home was the most frequently mentioned advantage of self-injections (50.0%), with requiring refrigeration being the most cited disadvantage of this mode of administration (30.0%). The only advantage attributed to infusions was the potential for less frequent administration, with 50.0% reporting that they would prefer infusion to be administered every 3–6 months; patients most often (80.0%) mentioned that the time and office visit required for each infusion were inconvenient. For an oral mode of administration, a variety of advantages were mentioned by patients, including being better for those with needle phobias, convenient for travel, and the ease of following the regimen (all mentioned by two patients each); the biggest disadvantage of an oral mode of administration was forgetting to take the medication, which was cited by 30.0% of patients. Patients perceived more advantages with self-injection and oral modes of administration than with infusion, for which disadvantages were primarily cited.

Bild vergrößern

A total of 222 eligible patients completed the survey, and their data were reviewed for quality. It was determined that n = 12 records needed to be excluded from analyses due to the inability to distinguish their responses from random responding on the BWS choice tasks, which resulted in a final study sample of 210 eligible patients with AS. Patient characteristics are shown in Table 3. Among 210 patients, half were male (50.0%), half were aged ≥ 42 years (50.0%), and the mean age of patients in the sample was 44.0 years. Most patients (93.3%) had poorly controlled disease (BASDAI score ≥ 4); 6.7% had well-controlled disease (BASDAI score < 4). Just over a third (37.6%) were currently on an advanced therapy.