I have a casual question. When a self-report outcomes measurement tool is being validated, is it reasonable for the subjects to complete only one tool? Meaning, in the below study, all subjects completed a legacy tool, the DASH. (Legacy tools were the first self-report tools used prior to computer adaptive testing.) The researchers then carved out the responses within the DASH and created a QuickDASH score. I don't quite understand how validation can actually happen because we really don't know if a patient completed the DASH, and was then required to complete the QuickDASH 20 minutes later if the patient would truly exactly match the reponse indicated in the DASH.
See what you think...
Here's the abstract for quick viewing.
To examine the agreement of scores between the Disability of Arm, Shoulder and Hand (DASH) and QuickDASH Questionnaires in patients with distal radius fractures (DRF) and their score's concurrent validity with PRWE scores.
Validity Study SETTING: Hand and upper limb clinic PARTICIPANTS: One hundred and seventy-seven patients with Distal Radius Fractures over the age of 18 were included in this study.
N/A MAIN OUTCOME MEASURES: Measurements of the DASH, QuickDASH, and Patient Reported Wrist Evaluation (PRWE) were taken at baseline and 1-year follow-up. QuickDASH scores were extracted from the DASH scores. Agreement analysis of the DASH and QuickDASH were evaluated using Bland-Altman's technique. Item difficultly analysis was performed to examine the distribution of QuickDASH items amongst DASH items. Responsiveness of the DASH, QuickDASH, and PRWE were also evaluated by calculating standardized response means.
QuickDASH scores were higher than DASH scores, particularly at baseline. A mean difference of 3.8 and 1.2 points were observed at baseline and 1-year follow-up, respectively. The limits of agreement (LOA) were wide at baseline with a range of 24.8 points at baseline, but decreased to 12.5 points at 1-year follow-up. Item difficulty analysis revealed that QuickDASH items were not evenly distributed at baseline. Finally, the responsiveness of the DASH, QuickDASH and PRWE were similar from baseline to 1-year follow-up (standardized response mean of 2.13, 2.17, and 2.19, respectively).
When changing from the DASH to the QuickDASH in the context of DRF, a systematic bias of higher scores on the QuickDASH should be considered by the user. However, the QuickDASH still demonstrated good concurrent validity and responsiveness.
Arch Phys Med Rehabil. 2016 Dec 27. pii: S0003-9993(16)31330-2. doi: 10.1016/j.apmr.2016.11.023. [Epub ahead of print]