Measuring Care Work-Related Quality of Life: A Rasch Validation Study of the Adult Social Care Outcomes Toolkit for Workforce (ASCOT-Workforce)

Rasch analysis was applied to establish the structural validity of the ASCOT-Workforce measure of care-work related quality of life (CWRQoL). The analysis was also used to guide potential item reduction to create a short-form version, which could be used in future studies to develop preference weights for use in economic evaluation of interventions and policy. Rasch analysis was applied to the ASCOT-Workforce (13 items) from a survey of the adult social care workforce in England (n = 7233). The assumption of a single scale was based on the finding of a single factor structure for the 13-item measure from a prior study that applied exploratory factor analysis. Rasch analysis of ASCOT-Workforce (13 items) had acceptable model fit, internal consistency and met assumptions of unidimensionality and local independence. There was suboptimal distinguishability at some thresholds between response options and evidence of differential item functioning (DIF). Most issues were addressed by omitting items to create a 9-item version, although there were still issues with three items. These were retained based on policy relevance and face validity to key stakeholders. ASCOT-Workforce (13-item, 9-item) are both valid instruments of CWRQoL for those working in the adult social care workforce. The short-form (9-items) performs slightly better than the 13-item measure, although suboptimal distinguishability at category thresholds and DIF persisted for three items. Further qualitative study to explore the reasons underlying DIF, and whether re-drafting the response options could improve the response category distinguishability, are recommended.