Validity of 30-item sinhala version of general health questionnaires

Abstract

The General Health Questionnaire (GHQ) is widely used to detect psychiatric disorders. The Sinhala translation of GHQ 30 has been widely used for research purposes in Sri Lanka, but details pertaining to the validation procedure have not been published. The objective of the study is to determine the criterion validity of the Sinhala version of GHQ-30. This was a descriptive cross-sectional study, including 374 patients within the age range of 18 to 75 years, attending the Out Patient Department of the Colombo North Teaching Hospital during the period between June 2009 and September 2010. The Sinhala version of GHQ-30 was given to be completed by the participants. Structured clinical examination was done blindly to the GHQ score by the Consultant Psychiatrist based on the ICD 10 classification of Depression/Anxiety as the reference standard. Threshold values for GHQ-30 were determined by applying Receiver Operation Characteristic (ROC) curves. The results show that according to the ROC curve, the optimal cutoff level was 5/6 of GHQ-30, with sensitivity of 67.5%, specificity of 80%. Subgroup analysis revealed that all the subgroups had an optimal cutoff level of 5/6 except for females which was 4/5. The GHQ-30 misclassified 26.1% of subjects either with false positives or negatives when the cutoff level was 5/6. With regard to stratum specific likelihood ratios, a value of > 1 was obtained only with a total score of 7 and above and for a cutoff level set at 6/7, sensitivity and specificity were 64% and 84% respectively with a misclassification rate of 25.9%. In conclusion, GHQ-30 is a useful method for assessing psychiatric disorders in the Sinhala speaking Sri Lankan population.