P Martus1, O Jakob, U Rose, R Seibt, G Freude. 1. Institute for Biostatistics and Clinical Epidemiology, Hindenburgdamm 30, 12000 Berlin, Germany. peter.martus@charite.de
Abstract
BACKGROUND: The Work Ability Index (WAI) is a well-established instrument to measure work ability. However, the dimensionality of the WAI remains controversial. AIMS: To identify the dimensionality of the WAI and to investigate dependencies of factors and subscales. METHODS: The sample analysed in this study consisted of 371 subjects of different occupational groups (teachers, office workers, nursery school teachers and managers). The WAI was measured for all subgroups. Psychometric characteristics of the WAI were investigated using factor analyses with different numbers and different patterns of dependency among the factors. Chi-square analysis and the Comparative Fit Index were used to statistically assess fit quality. RESULTS: The group of managers had to be excluded from the analysis as their results were probably overoptimistic due to reporting bias; thus, 324 subjects entered. The one-factor model and an orthogonal two-factor model did not fit the observed correlational structures. A satisfactory fit was obtained using a two-dimensional model with correlated factors. These factors could be interpreted as subjectively estimated work ability and objective health status. Only five of seven items of the WAI could be related unambiguously to one of both factors. CONCLUSIONS: From our study, we conclude that using only the total score of the WAI is not adequate for population analysis of and assessment of work ability to individuals. Instead, the two-dimensional structure of the instrument must be taken into account.
BACKGROUND: The Work Ability Index (WAI) is a well-established instrument to measure work ability. However, the dimensionality of the WAI remains controversial. AIMS: To identify the dimensionality of the WAI and to investigate dependencies of factors and subscales. METHODS: The sample analysed in this study consisted of 371 subjects of different occupational groups (teachers, office workers, nursery school teachers and managers). The WAI was measured for all subgroups. Psychometric characteristics of the WAI were investigated using factor analyses with different numbers and different patterns of dependency among the factors. Chi-square analysis and the Comparative Fit Index were used to statistically assess fit quality. RESULTS: The group of managers had to be excluded from the analysis as their results were probably overoptimistic due to reporting bias; thus, 324 subjects entered. The one-factor model and an orthogonal two-factor model did not fit the observed correlational structures. A satisfactory fit was obtained using a two-dimensional model with correlated factors. These factors could be interpreted as subjectively estimated work ability and objective health status. Only five of seven items of the WAI could be related unambiguously to one of both factors. CONCLUSIONS: From our study, we conclude that using only the total score of the WAI is not adequate for population analysis of and assessment of work ability to individuals. Instead, the two-dimensional structure of the instrument must be taken into account.
Authors: Auli Airila; Jari Hakanen; Anne Punakallio; Sirpa Lusa; Ritva Luukkonen Journal: Int Arch Occup Environ Health Date: 2012-01-21 Impact factor: 3.015
Authors: Virve Pekurinen; Laura Willman; Marianna Virtanen; Mika Kivimäki; Jussi Vahtera; Maritta Välimäki Journal: Int J Environ Res Public Health Date: 2017-10-18 Impact factor: 3.390