General health questionnaire-12: factor analysis in the general population of Bucaramanga, Colombia

Abstract

BACKGROUND: The General Health Questionnaire (GHQ-12) was designed to quantify emotional symptoms (common mental disorders) in different settings. However, its factor structure is not known in Colombian population.

OBJECTIVE: To establish the factor structure and the best form of punctuation for the GHQ-12 in the general population of Bucaramanga, Colombia.

METHOD: A probabilistic sample of 2,496 adults from the general population in Bucaramanga, Colombia, completed the GHQ-12. The internal consistency and analysis factor were computed for the ordinal scoring (0-1-2-3) and for binary scoring (0-0-0-1, 0-0-1-1, and 0-1-1-1). RESULTS: The ordinal scoring GHQ-12 presented an internal consistency of 0.779 and two factors that accounted for 42.5% of the variance. In other hand, the scoring way 0-1-1-1 exhibited an internal consistency of 0.774; 0-0-1-1 scoring, of 0.708; 0-0-0-1 scoring, of 0.360. The binary scoring 0-1-1-1 and 0-0-1-1 showed three factors responsible of 50.6% and 46.3% of the variance respectively. The 0-0-0-1 scoring did not present a good factor solution.

CONCLUSIONS: Among people from the general populalation the ordinal scoring for the GHQ-12 exhibites better psychometric properties than binary ones. Except the 0-0-0-1 scoring method, reminding scorings present a high internal consistency and aceptable factor solution.