Measurement properties of the decisional conflict scale in people living with chronic non-cancer pain: a secondary analysis of the DECIDE-PAIN pan-Canadian survey.

Objectives: The Decisional Conflict Scale is a candidate instrument for the Core Outcome Measurement Set of the OMERACT Shared Decision-Making working group. However, its validity among chronic non-cancer pain populations is lacking. We explored measurement properties of the English version of the 16-item Decisional Conflict Scale in this population.

Methods: Informed by the COnsensus-based Standards for the selection of health status Measurement INstruments guidelines, we conducted a secondary analysis of a panCanadian cross-sectional survey. We assessed the distribution of score including ceiling and floor effects, readability, internal consistency, structural validity of an hypothesized five-factor model, and convergent and discriminant validities. We tested measurement invariances regarding age, sex, health literacy, and education.

Results: We collected data from 1103 respondents. We found ceiling effect at the item-level, while no ceiling and floor effects at the subscale and total score levels. Readability scores were below the sixth-grade reading level. All results of internal consistency were above the a priori threshold. The confirmatory factorial analysis resulted in a good model fit. We found good convergent, but poor discriminant validities. We confirmed measurement invariances for age, sex, health literacy, and education.

Conclusion: This secondary analysis of a panCanadian survey provides evidence of the measurement properties of the Decisional Conflict Scale in people living with chronic non-cancer pain. We propose that this instrument should be analyzed using its total score rather than its subscales due to a lack of discriminant validity. Further research is necessary to enhance its validity, possibly by re-examining its factor structure.