Abstract
Bio-psycho-social screening to predict patients with a high risk of prolonged recovery is advocated by professional and insurance groups. However, a suitable brief, accurate, and psychometrically sound questionnaire for general musculoskeletal patients is lacking. This study developed such a tool by modifying and shortening the 21-item Örebro Musculoskeletal Pain Questionnaire (ÖMPQ) to a 15 item Screening Questionnaire (ÖMSQ-15). Two itemreduction methods were investigated: concept-retention: subjective qualitative, and factor analysis-statistical based. An a-priori minimum of 15-items was determined from the Spearman-Brown prophesy where internal consistency (α) would exceed 0.80. A data set derived from generalsymptomatic musculoskeletal compensation patients (n = 194) was reanalysed with both ÖMSQ-15 versions to assess psychometric and practical characteristics at baseline and predictive performance at 6 months, then compared with the full length version. The concept-retention versionperformed best with higher correlation (Pearson's r = 0.99 p less than 0.01) whilst reliability regressed marginally (r = 0.95 p less than 0.01, ICC 2,1). Predictive performance improvedat an 86 ÖMSQ-15-points (57%) cut-off value where convergent validity gave higher subsequent positivelikelihood ratios (3.9-4.8) and correlated highly (r = 0.73 p less than 0.01) with recovery time. Divergent validity showed mean scores as significantly different (p less than 0.001) between patients with positive and negative outcome traits. Factor structure (maximum likelihood extraction) gave a coherent four-factor model that accounted for 56% of variance. Practicality improved markedly through reduced missing responses (4.9%) and a 21% reduction in scoring time and completion time (4.42 + 2.39 min). A qualitative conceptretention version rather than a statistically driven approach produced improved validity, reliability and practicality in the final tool. Screening questionnaires, particularly the Örebro tool, are now advocated by various professional groups and statutory state government insurance authorities as a means of providing early recognition of potential prolonged recovery and long term claimants. However, the most commonly advocated and recommended tool: the Örebro Musculoskeletal Pain Questionnaire (ÖMPQ),has been modified from the original low back tool without validation of the changes and subsequent interpretation within the target populations. Further more, it is not a pain questionnaire but a screening tool that provides cut-offs for levels of risk determination. In its current non-validated form, the advocated ÖMPQ gives rise to significant missing responses, poor interpretation and erroneous scores that can be both time consuming and difficult to calculate and interpret. The shortening of the tool to a validated 15 item screening questionnaire reduces missing and erroneous responses which improves practicality, patient acceptance, predictive capacity and the ease of clinical use.