Forty-two studies with a total of 6,314 participants were included in the review. The sample size of the included studies ranged from 40 to 715, and the prevalence of depression was between 5.6 and 51.6%.
Thirty-two of the included studies investigated the GDS used in its original language (English), while the remaining 10 using translated versions.
In 26 studies, researchers assessing the reference standard were blinded to the results of the GDS.
Diagnostic accuracy of the GDS-15.
Not including modified Mandarin and Cantonese versions, the sensitivity ranged from 0.600 to 0.940 and the specificity from 0.570 to 0.870. For all GDS-15 studies, the weighted mean sensitivity was 0.805 and the weighted mean specificity 0.750 (21 studies).
Diagnostic accuracy of the GDS-30.
The sensitivity ranged from 0.340 to 1.000 and the specificity from 0.629 to 0.964. For all GDS-30 studies, the weighted mean sensitivity was 0.753 and the weighted mean specificity 0.770 (33 studies).
Comparison of the GDS-15 and GDS-30.
Using studies reporting data for the GDS-15 and GDS-30 in identical samples (9 studies), sensitivity was significantly higher for the GDS-30 than for the GDS-15 while specificity was significantly higher for the GDS-15 than for the GDS-30. The mean overall misclassification rate was lower for the GDS-15.
Comparison of the GDS-15 and GDS-30 with other screening instruments.
Direct comparisons of the Centre for Epidemiological Studies Depression scale (CES-D) with the GDS-15 (4 studies) showed a significantly lower mean sensitivity and a significantly higher mean specificity for the CES-D; results were similar for comparisons with the GDS-30 (6 studies). Direct comparisons of
the Yale-1-question screen with the GDS-15 (4 studies) and GDS-30 (3 studies) showed significantly lower mean sensitivity and specificity for the Yale-1-question screen in both cases.