Reliability of composite measures for the assessment of psoriatic arthritis.

Background: Composite measures of disease activity have been developed for use in Psoriatic Arthritis (PsA) to capture the wide spectrum of disease but there is a lack of consensus regarding which to adopt for routine practice. It is recognised that more data is required to understand the measurement properties of existing instruments and consider the impact of modifications that may improve face validity, responsiveness or feasibility. It is important to have an estimate of a measurement instrument's reliability in the setting of stable disease in order to understand measurement error and responsiveness. To our knowledge no data exists on the stability of composite measures in PsA. Objective(s): To measure test re-test reliability of composite measures of disease activity in PsA. Method(s): Clinical and patient reported outcomes to enable the calculation of composite measures were administered to 141 patients with PsA at five time points in a UK multicentre observational study. All patients fulfilled the CASPAR criteria. Twenty-nine patients with clinically stable disease and receiving no treatment intervention underwent repeat assessment by the same examiner within 2 weeks. Patients in high and low disease were included. Reliability was evaluated by intra-class correlation coefficient (ICC) and Bland Altman plots. Result(s): Of the 29 patients included 15 were male, the mean age was 52.4 years (SD 13.39), mean disease duration at T0 was 9.2yrs (SD 8.11). The mean swollen joint count was 3.4 (SD 5.1), tender joint count 11.3 (SD 15.03) and PASI 1.0 (SD1.04). The ICC (95% CI) for tender and swollen joint counts were 0.94 (0.87-0.97) and 0.91 (0.80-0.96) respectively. The ICC for PASI was 0.95 (0.90-0.98). All composite measures demonstrated high levels of test-retest reliability with ICC >0.85, table. The most reliable measure was the PADAS ICC 0.98 (95% CI 0.954-0.991). The individual ICC for each composite measures are reported in the table and Bland Altman plots