The Posterior Shoulder Instability Questionnaire: internal consistency, content and criterion validity, responsiveness, and reliability of a new tool for the assessment of posterior shoulder instability.

ABSTRACT

BACKGROUND: Posterior shoulder instability (PSI) is an increasingly recognized cause of shoulder dysfunction particularly in young active patients and certain athlete populations. When evaluating the efficacy of treatment for PSI, specific outcome measures for this population are essential. The aim of the current research was to describe the development and evaluation of a patient reported outcome measure specific for PSI. METHODS: A retrospective cohort study design of patients with PSI was used to develop and evaluate the "Posterior Shoulder Instability Questionnaire (PSI-Q)". Items for PSI-Q were generated through an expert focus group and existing questionnaires. Preliminary data analysis identified redundancy of items and resulted in the PSI-Q being refined. The final PSI-Q was evaluated on 128 patients with PSI with a structural lesion requiring surgical intervention. Participants were excluded in the absence of a posterior glenohumeral joint lesion. Internal consistency (Cronbach α and corrected item-total correlation), content validity, criterion validity, responsiveness, and test-retest reliability (intraclass correlation coefficient) were examined. Content validity, criterion validity and responsiveness were compared with the Melbourne Instability Shoulder Scale (MISS) and the Western Ontario Shoulder Instability Index (WOSI). The minimum detectable change score (MDC) was calculated. RESULTS: The Cronbach α for the total scale preintervention and postintervention was high (α = 0.97). All five domains (pain, instability/weakness/stiffness, function, occupation and sport, and quality of life and satisfaction) demonstrated acceptable internal consistency for each subsection and the overall score of the scale (α > 0.70). The corrected-item total correlation for each domain was within an acceptable range. The responsiveness of the PSI-Q questionnaire was excellent (effect size, 2.06; standard response mean, 1.34) and was higher than the MISS and WOSI. There were no relevant floor effects and 1 ceiling effect. Reliability was excellent (intraclass correlation coefficient(1,1) = 0.93) and the calculated MDC was 10.9 points. DISCUSSION: This study designed and validated a questionnaire specific for measuring symptoms and function in people with structural PSI requiring surgery. The PSI-Q demonstrates good measurement properties and provides an MDC that is useful for researchers and clinicians. In structural PSI, the PSI-Q has a higher responsiveness and more accurately reflects a patient's overall perceived shoulder status compared to current patient reported outcomes for shoulder instability. The psychometric properties of the PSI-Q are still to be determined in a nonsurgical population.