Is there room for improvement in the Spanish version of the scaled General Health Questionnaire (GHQ-28)?

To develop a new scaled GHQ version (GHQ-28) for use in Primary Care in Spain. Principal components analysis of the parental version (GHQ-60) with subsequent selection of 7 items with higher loadings for the first 4 components. SITE: Three Primary Care Centres in the metropolitan area of Bilbao. A total of 202 patients coming from a prior study on frequent visits to Primary Care clinics. Comparison between the new and the authorised version of GHQ-28 in discriminating between psychiatric and non-psychiatric cases, diagnosed independently by psychiatric interview. Whereas the structure of two scales was similar to that described previously, the other two scales showed higher variability with the inclusion of new items and a combination of others previously ascribed to different measurement constructs. No significant difference (P=.63) was observed between both versions regarding their discriminant validity. The most stable GHQ-28 scales were social dysfunction and depression. The variability found among the other scales could be explained by the non-specificity of psychiatric symptoms presented in Primary Care.