Background Morning stiffness is a frequently used clinical and research outcome measure and is important to patients1, but was omitted from the RA core set because of poor measurement properties2. Current stiffness assessment is inconsistent with patient's perspectives of the symptom3,4 and has not been developed according to PROM development guidelines5. The appropriate content of a new RA stiffness PROM was previously explored and developed through qualitative interview3 and focus group studies. Draft items were subsequently tested and refined with patients during cognitive interviews, resulting in 39 draft items for inclusion in a PROM. Here we report a quantitative assessment to create the smallest and most internally consistent set of items for a developmentally valid stiffness PROM.
Objectives To develop the content and structure of a new RA stiffness PROM.
Methods A postal questionnaire pack was sent to patients with RA based in the South-West of England. It contained 45 items assessing stiffness (39 draft items and 6 items currently used in stiffness assessment), individual items capturing pain (VAS), fatigue (NRS), patient global assessment (VAS), questionnaires capturing disability (MHAQ), and patient-reported disease activity (PDAS26), and basic demographic information. Initial investigation identified items with poor response rates, distributions or correlations for removal. A series of principal component analyses were undertaken with the remaining items, balancing Cronbach's alpha for internal consistency, stability of the component structure (assessed by multiple analyses using random 50% samples of the respondents (bootstrapping)) and parsimony. Based on the statistical results and aided by expert judgement, the smallest number of informative items were retained.
Results 277 patients (91 male) aged 23–97 years with disease durations 1–45 years participated in the study (42.9% response rate). Seven of the 45 items were removed during initial item investigation. The remaining 38 items demonstrated high Cronbach's alpha (>0.9). During successive rounds of analytical refinement, 17 items were removed. After round 5, a 3-component structure emerged which remained consistent for a further 13 rounds of testing item removal, demonstrating stability. These components captured “stiffness severity”, “physical impact” and “psychosocial impact”. The overall Cronbach's alpha of the final 21 items was 0.95 indicating a homogenous set of items and bootstrapping further demonstrated the stability of the structure.
Conclusions A new 21 item, 3-component RA stiffness PROM has been developed based on qualitative work3 with RA patients to enhance content validity. Further testing is now required to assess the validity, reliability and sensitivity to change of the new RA stiffness PROM.
Hewlett et al, 2005;
Felson et al, 1993;
Halls et al, 2014;
Orbai et al, 2014;
Choy et al, 2008; 2015
Disclosure of Interest None declared