Using Rasch analysis to assess the latent construct of the Capacity to Work Index in a Swedish working population sample.

Measurements of capacity to work (CTW) in relation to common mental disorders (CMD) are needed to improve research on determinants for maintained work participation (WP). The aim of this study was to assess the construct validity of the Capacity to Work Index (C2WI) in a heterogenous sample of the Swedish working population. Cross-sectional web survey data among Swedish employees (n = 8201) was used. The construct validity was assessed in terms of the unidimensionality of the scale, response categories appropriateness, and the differential item functioning with respect to gender and age. Rasch analysis was used in both the full sample and randomly selected subsets. The Rasch model (RM) was fitted using two versions of the C2WI construct: the original 17-item scale and a 7-item scale. The 17-item scale did not discriminate as required, whereas the 7-item scale demonstrated a better fit to the RM. However, statistically significant χ2 statistics indicated non-invariant item performance across the latent continuum. A third attempt assessed both scales on a subgroup, yielding improved results, but an overall fit to the RM was not achieved. CTW in relation to CMD assessed by the C2WI did not fulfil the requirements for construct validity outlined by the RM. Real-world experiences of CTW are complex and reflect expressions of mental health in diverse work environments. Further studies are required to determine the predictive capacity of C2WI and its individual items in relation to relevant outcomes, such as maintained WP in the working population.