Validity of the Malaise Inventory in general population samples
- Cite this article as:
- Rodgers, B., Pickles, A., Power, C. et al. Soc Psychiatry Psychiatr Epidemiol (1999) 34: 333. doi:10.1007/s001270050153
- 257 Views
Background: The Malaise Inventory is a commonly used self-completion scale for assessing psychiatric morbidity. There is some evidence that it may represent two separate psychological and somatic sub-scales rather than a single underlying factor of distress. This paper provides further information on the factor structure of the Inventory and on the reliability and validity of the total scale and two sub-scales. Methods: Two general population samples completed the full Inventory: over 11,000 subjects from the National Child Development Study at ages 23 and 33, and 544 mothers of adolescents included in the Isle of Wight epidemiological surveys. Results: The internal consistency of the full 24-item scale and the 15-item psychological sub-scale were found to be acceptable, but the eight-item somatic sub-scale was less reliable. Factor analysis of all 24 items identified a first main general factor and a second more purely psychological factor. Receiver operating characteristic (ROC) analysis indicated that the validity of the scale held for men and women separately and for different socio-economic groups, by reference to external criteria covering current or recent psychiatric morbidity and service use, and that the psychological sub-scale had no greater validity than the full scale. Conclusions: This study did not support the separate scoring of a somatic sub-scale of the Malaise Inventory. Use of the 15-item psychological sub-scale can be justified on the grounds of reduced time and cost for completion, with little loss of reliability or validity, but this approach would not significantly enhance the properties of the Inventory by comparison with the full 24-item scale. Inclusion of somatic items may be more problematic when the full scale is used to compare particular sub-populations with different propensities for physical morbidity, such as different age groups, and in these circumstances it would be a sensible precaution to utilise the 15-item psychological sub-scale.