A modified QuickDASH-9 provides a valid outcome instrument for upper limb function

Abstract

Background The 30-item Disabilities Arm Shoulder and Hand (DASH) questionnaire was introduced to facilitate assessment of upper limb functional limitations. To improve practicality and eliminate item redundancy a modified instrument was needed. The 11-item QuickDASH was developed to fulfil these requirements and translated into several languages. However, prospective investigations of psychometric and practical characteristics are limited. No published study investigated readability or used concurrent validation with a standardized upper limb criterion measure. The validity of the QuickDASH has been questioned as the results for factor structure are conflicting, and the English-language version has not yet had factor structure reported. A shortened 9-item version, the QuickDASH-9, that addresses these issues is proposed.

Methods This two-stage observational study assessed the psychometric and practical characteristics of the QuickDASH and the extracted QuickDASH-9. The Upper Limb Functional Index (ULFI) was the criterion standard in both stages. Stage 1, calibration, reanalyzed extracted QuickDASH and QuickDASH-9 responses from a previous prospective study, by the authors, of the 30-item DASH (n = 137). Stage 2, prospective validation, investigated the QuickDASH through repeated measures in consecutive upper limb musculoskeletal participants' consulting for physical therapy in Australia (n = 67). The QuickDASH and extracted QuickDASH-9 data from both stages was analyzed and compared for psychometric properties, practical characteristics and factor structure.

Results The proposed QuickDASH-9 had a unidimensional structure, high reliability (ICC 2:1, r = 0.92), internal consistency (alpha = 0.93) and responsiveness (ES = 1.05). It correlated highly with both the DASH (r = 0.97), QuickDASH (r = 0.99) and ULFI criterion (r = 0.85). QuickDASH-9 missing responses reduced to 3.5% from 26% in the QuickDASH. Completion and scoring time was 134 ± 56 seconds and required a computational aid. The QuickDASH demonstrated a bidimensional structure making it invalid. The QuickDASH-9 summary performance was measured on the 'Measurement of Outcome Measures' at 88% and on the 'Bot' clinimetric scale at 75%.

Conclusions The proposed QuickDASH-9 had a unidimensional structure and similar psychometric precision to the full-length DASH with improved practicality and completion time. The QuickDASH was invalid as its bidimensional structure made a single summated score inappropriate. The QuickDASH-9 offers a future direction for ongoing use of the QuickDASH concept.