Development and validity of a work functioning impairment scale based on the Rasch model among Japanese workers

OBJECTIVES: The purpose of this study was to develop a new work functioning impairment scale (WFun) and examine its validity. METHODS: The WFun was developed based on the Rasch model, which consists of seven items. We conducted a pilot study (n=1,000) using an Internet investigation and a field study (n=1,294) in a manufacturing industry, and we additionally collected data from six workplaces from other industries. This series of studies was examined with a Rasch model analyses including item fit statistics as well as hypothesis testing. Convergent validity was used to examined the association of the WFun with the Stanford Presenteeism Scale, SF-8, Work Ability Index, and several types of job disruptions. We also examined differential test functioning. RESULTS: All the items showed adequate fit (infit mean-square statistics <1.5). The item reliability was 0.98, and the item separation index was 6.37. The person reliability was 0.86, and the person separation index was 2.32. All tests for convergent validity showed significant differences. All p values derived from ANOVA were highly significant (p<0.001). No differential test function was observed between groups by age, sex, or job type or between various samples from different workplaces. The intraclass correlation of the estimated Rasch measurements from these groups was 0.99 (95% CI: 0.976-0.992). CONCLUSIONS: The WFun was confirmed to show good fit to a Rasch model and construct validity. Given that its good fit indicates specific objectivity, this tool will be useful in assessing the ability of individuals to function at work and in evaluating group levels for benchmarking.