Materials and methods

This prospective cohort study enrolled 54 consecutive patients aged 18–65 years who received a new diagnosis of pSjD in a tertiary hospital’s rheumatology outpatient clinic. All participants fulfilled the 2016 American College of Rheumatology/European Alliance of Associations for Rheumatology (ACR/EULAR) classification criteria [15].

Patients were reassessed after a mean follow-up of 21.36 ± 4.79 months. Of the initial cohort, 40 patients (74.1%) attended the follow-up assessment. Of these, 25 patients (46.3%) attended the follow-up visit as scheduled, whereas 15 patients (27.8%) were contacted by telephone and subsequently invited for an in-person follow-up visit. Baseline characteristics of patients who completed follow-up and those who were lost to follow-up were compared to assess potential selection bias.

Reporting guideline: This prospective observational cohort study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement for cohort studies. The completed STROBE checklist was prepared and submitted as Supplementary File 1 [16].

This study was approved by the Selçuk University Rectorate Local Ethics Committee (approval no: 2024/296; meeting date: 04 June 2024; official document date and number: 05 June 2024, E-70632468-050.04-768772). Written informed consent was obtained from all participants. The study was conducted in accordance with the Declaration of Helsinki, October 2024 version, and Good Clinical Practice principles.

Baseline demographic and clinical characteristics were recorded using a structured questionnaire, including age, sex, height, weight, smoking status, symptom duration, employment, and educational level. It should be noted that “newly diagnosed” refers to the time of diagnosis rather than symptom onset; therefore, symptom duration was recorded to better characterize disease stage.

Systemic disease activity was evaluated at baseline and at follow-up using the EULAR Sjögren’s Syndrome Disease Activity Index (ESSDAI), which assesses 12 organ domains: constitutional, lymphadenopathy, glandular, articular, cutaneous, pulmonary, renal, peripheral and central nervous system, muscular, hematological, and biological involvement. Total scores range from 0 to 123 and are categorized as low (< 5), moderate [5–13], or high (≥ 14) [17].

Patient-reported symptoms were assessed using the EULAR Sjögren’s Syndrome Patient Reported Index (ESSPRI), which evaluates dryness, pain, and fatigue on 0–10 numerical rating scales, and the total score is calculated as the mean of these domains [18].

Medication adherence was assessed at the follow-up visit using the 5-item Compliance Questionnaire for Rheumatology (CQR-5), a brief self-report instrument developed to identify patients at risk of suboptimal adherence [19]. The Turkish version of the CQR-5 has demonstrated good validity and reliability in patients with rheumatoid arthritis [20]; however, its use in patients with pSjD has not been specifically validated and should be interpreted with caution. Items are rated on a Likert scale, and according to recommended scoring procedures, patients were categorized into low- and high-adherence groups.

Patients’ beliefs regarding their prescribed treatments were evaluated at the same follow-up time point using the Beliefs about Medicines Questionnaire (BMQ), which includes specific necessity and concern subscales. This instrument was originally developed by Horne and colleagues to assess cognitive representations of medication and has been widely used and adapted in various patient populations [21]. Higher necessity scores indicate stronger beliefs in the need for medication, whereas higher concern scores reflect apprehension about potential adverse effects. The necessity-concern differential (NCD) was calculated by subtracting the concern score from the necessity score. As adherence and beliefs were assessed at the same time point, analyses reflect associations rather than temporal or predictive relationships.

Questionnaire-based assessments: The questionnaire-based assessments were planned, administered, and reported with reference to current recommendations for survey and questionnaire-based research. All patient-reported instruments were administered according to their published instructions and using validated versions where available. The timing of each assessment was prespecified: ESSPRI, HADS, fatigue severity assessment, HAQ, and SF-12 were evaluated at baseline and follow-up, whereas medication adherence using the CQR-5 and beliefs about medicines using the BMQ were assessed at the follow-up visit. The CQR-5 was used to classify patients into low- and high-adherence groups as a brief self-report adherence screening instrument. Although the Turkish version of the CQR-5 has demonstrated validity and reliability in rheumatoid arthritis, it has not been specifically validated in primary Sjögren disease; therefore, CQR-5-based adherence classification was interpreted cautiously. The BMQ was used to assess patients’ perceived necessity of treatment and concerns about medicines, and the necessity-concern differential was calculated by subtracting the concern score from the necessity score. Because adherence and medication beliefs were measured at the same follow-up time point, these analyses were considered cross-sectional within the follow-up assessment and were interpreted as associations rather than temporal or causal relationships. The questionnaires and scales were used in accordance with their published instructions, validated versions, and applicable academic-use or licensing requirements; no full questionnaire forms or copyrighted item content are reproduced in this manuscript [22].

Patients were managed according to routine clinical practice. Detailed information regarding specific medications, number of treatments, treatment modifications, or adverse effects was not systematically collected as part of the study protocol; therefore, medication adherence was evaluated independently of specific treatment regimens.

Psychological symptoms were assessed using the Hospital Anxiety and Depression Scale (HADS). Functional status was evaluated with the Health Assessment Questionnaire (HAQ), and health-related quality of life was measured using the 12-item Short Form Health Survey (SF-12). All measures were recorded at baseline and at follow-up.

All statistical analyses were performed using IBM SPSS Statistics version 22 (IBM Corp., Armonk, NY, USA). The Shapiro–Wilk test and visual inspection of histograms and Q–Q plots were used to assess the normality of continuous variables. Normally distributed variables are presented as mean ± standard deviation, whereas non-normally distributed variables are reported as median with interquartile range. Categorical variables are expressed as numbers and percentages.

Between-group comparisons were performed according to the distribution and type of variables. The independent-samples t-test or Mann–Whitney U test was used for continuous variables, and the chi-square test or Fisher’s exact test was used for categorical variables, as appropriate. Within-group changes from baseline to follow-up were assessed using paired-samples t-tests for normally distributed variables and Wilcoxon signed-rank tests for non-normally distributed variables. In addition to p values, numerical differences between groups, direction of change, odds ratios, 95% confidence intervals, and model-fit indices were considered when interpreting the findings.

Binary logistic regression analysis was performed to explore factors associated with high medication adherence. Because of the relatively small number of patients and events, the regression analysis was intentionally restricted and considered exploratory. The necessity-concern differential was entered as the main explanatory variable because it represented the principal perceptual construct of interest and showed the clearest between-group separation in univariable analyses. No multivariable overfitted model was constructed. Model fit was evaluated using the omnibus test, Cox and Snell R2, Nagelkerke R2, and overall classification accuracy. All tests were two-sided. A p value < 0.05 was considered statistically significant; however, statistical significance was not interpreted as evidence of causality.