Detailed Description:
Study Design and Objectives
STRIDE-PsA is a prospective, multicentre observational cohort study conducted within routine NHS rheumatology care. Approximately ten NHS sites will recruit adults with psoriatic arthritis (PsA) initiating a new biologic or targeted synthetic disease-modifying antirheumatic drug (b/tsDMARD). Participants will be followed for 52 weeks with assessments at baseline and weeks 4, 12, 24, and 52. As this is a non-interventional study, treatment decisions and clinical management will remain at the discretion of the treating clinician.
The study aims to evaluate treatment effectiveness across different lines of advanced therapy and to investigate whether anti-drug antibodies (ADAs) or circulating drug levels are associated with clinical response. The underlying hypothesis is that patients may achieve comparable clinical improvement despite multiple prior therapy switches.
Study Procedures and Data Collection
Participants will be followed for 52 weeks with assessments aligned to routine NHS rheumatology care. Data collection will include clinical examinations, patient-reported outcome measures (PROMs), and blood sampling for ADA and drug level analysis.
At baseline, informed consent will be obtained and demographic and clinical history recorded. Participants will undergo routine clinical assessment to derive disease activity measures (including cDAPSA, BSA, and minimal disease activity status), complete validated PROMs (PSAID-9, HAQ-DI, PGADA), and provide routine NHS blood tests. An additional small research blood sample will be collected, where feasible alongside routine bloods.
Follow-up assessments will occur at approximately Weeks 4, 12, 24, and 52. PROMs will be collected at each follow-up time point, with repeat clinical examination and a second research blood sample at Week 12 during routine care visits. Information on suspected treatment-related adverse events identified during routine clinical review will be recorded throughout follow-up.
All procedures are designed to align with standard NHS appointments to minimise additional participant burden, and visit windows are permitted to allow flexibility in scheduling.
Recruitment and Follow-up
Potential participants will be identified during routine rheumatology appointments by the direct care team and provided with study information. Written informed consent will be obtained by trained research staff prior to any study-specific procedures. Participants may provide consent on the day of approach or after further consideration, according to preference.
Sample Size Rationale
The primary analysis will compare mean change in cDAPSA from baseline to Week 12 between patients initiating 4th/5th-line versus 2nd/3rd-line b/tsDMARD. The non-inferiority margin is set at -5.7 cDAPSA points, based on published estimates of the minimally clinically important difference, and the investigators assume SD = 12.36 for the 12-week change from published data. With one-sided α = 0.05 and 80% power, the required sample size is \~58 per group. Allowing for 10% drop out, this gives 65 per group, for a total of 130 participants across the two primary groups (equal allocation). Patients initiating first-line therapy and those on 6th-line or beyond will also be recruited as descriptive comparative cohorts. To ensure adequate precision for secondary endpoints (including patient-reported outcomes and quality-of-life measures) and to support regression analyses of treatment sequencing, the investigators plan to over-recruit, aiming for approximately 100 patients in each of the four groups (1st line, 2nd/3rd, 4th/5th, and 6th+; ≈400 total). Based on pilot data from Bath, which indicates around 80 patients have been commenced on b/tsDMARD therapy over a year, the investigators anticipate that recruitment from ten sites with a 50% consent rate will allow us to meet our recruitment target over 18 months.
Bias, Confounding, and Data Quality
Given the observational design, several strategies will be implemented to minimise bias. Screening logs will be maintained to evaluate recruitment patterns and potential selection bias. Analyses will adjust for pre-specified confounders, including study site, age, sex, baseline disease activity, disease duration, and comorbidities. Standardised electronic case report forms, validated PROMs, and site training will be used to reduce information bias.
Standardised data collection forms, training of research staff, and use of validated patient-reported outcome measures will minimise variation in data collection across sites. Loss to follow-up will be monitored, and reasons recorded where available. Sensitivity analyses will explore the impact of missing data, including multiple imputation where appropriate.
Statistical Analysis
The primary endpoint is change in cDAPSA score from baseline to Week 12. A non-inferiority margin of -5.7 points has been pre-specified based on published minimally clinically important difference thresholds. Change in cDAPSA will be analysed using analysis of covariance (ANCOVA), adjusting for baseline cDAPSA, age, sex, disease duration, relevant comorbidities, and study centre. A one-sided α = 0.05 will be used. Analyses will be conducted in both intention-to-treat and per-protocol populations.
Secondary endpoints will be analysed using regression models with covariate adjustment consistent with the primary analysis.
Achievement of minimal disease activity (MDA) at Week 12 and composite clinical response (cDAPSA low disease activity plus BSA \<1%) will be analysed using logistic regression. Continuous change in skin involvement (body surface area, BSA) will be analysed using analysis of covariance (ANCOVA), with logistic regression used for categorical skin response thresholds.
Patient-reported outcomes, including health-related quality of life (PsAID-9), physical function (HAQ-DI), and patient global assessment (PGADA), will be analysed using ANCOVA models adjusting for baseline values and predefined covariates.
Treatment persistence will be evaluated using Kaplan-Meier survival analysis and Cox proportional hazards regression. Flare outcomes, as measured by the PsA Flare questionnaire, will be analysed using logistic regression.
Anti-drug antibody (ADA) outcomes will be summarised descriptively, including prevalence, incidence, and titre distribution by therapy line. Exploratory analyses will assess associations between ADA status or titres, circulating drug levels, clinical response, treatment-related adverse events, and treatment persistence using appropriate multivariable models.
Missing data are anticipated due to incomplete follow-up. The primary analysis will assume data are missing at random and will use multiple imputation for continuous and binary Week-12 endpoints, incorporating baseline values and predefined covariates. Sensitivity analyses will include complete-case analyses and conservative assumptions for binary outcomes. For time-to-event endpoints, participants lost to follow-up will be right-censored at the last known treatment date, and event times will not be imputed.