Multi-item scales for measuring global self-rated health - Investigation of construct validity using structural equations models

The authors used multigroup confirmatory factor analysis and structural equations models to examine the construct validity and item functioning of the five-item General Health (GH) scale from the SF-36 in Danes over 16 years of age (n=4,084). They included four criteria variables for physical and mental health. Items GH2-GH5 had low response rates among the elderly, probably due to the compact layout of these items in the questionnaire. The authors found differential item functioning for several items, indicating some degree of multidimensionality in the GH scale. Thus, GH1 had stronger associations with age, physical functioning, and chronic diseases than predicted by the one-factor model. However, psychometrical problems were mostly found in the youngest age group. If the problem of missing values is solved by layout changes or interview administration, the GH scale appears to be a valid measure of self-rated health in elderly populations.