Measuring the mental health status of a population: a comparison of the GHQ-12 and the SF-36 (MHI-5).
C J McCabe, K J Thomas, J E Brazier, P Coleman


BACKGROUND A comparison of the performance of the 5-item mental health dimension of SF-36 (MRI-5) with that of the 12-item General Health Questionnaire (GHQ-12) in a defined, non-patient population, using standard statistical tests.

METHODS A postal survey of 3000 patients aged 16-64 years was conducted. Patients were randomly selected from the practice lists of two general practices chosen to represent populations with different socio-economic characteristics.

RESULTS Considerable evidence was found for the internal consistency of both instruments (Cronbach's alpha 0.91 and 0.84 for GHQ-12 and the MHI-5 respectively) and for their construct validity in terms of distinguishing between groups with measured health differences. Both instruments showed a significant difference in the mean scores for men and women. In contrast to the GHQ-12, no correlation was found between age and score for the MHI-5. Both instruments were equally sensitive to socio-economic characteristics and to levels of social support. The scores on the two instruments were highly correlated (Spearman rank correlation-0.73).

CONCLUSION The MHI-5 has comparable psychometric performance to the GHQ-12, and can be used to measure and compare mental health in defined populations. Operational advantages of the MHI-5 over the GHQ-12 are that it is in the public domain, is part of a general health measure (SF-36) and is shorter.