Twenty-eight studies were included in the review (5,309 participants, range 49 to 689). Prevalence of mental disorders ranged from 14% to 70%. Prevalence of depressive disorders ranged from 7% to 92%; three studies were excluded from pooling as their prevalence rates were highly dissimilar from most other studies. Study quality was variable (nine studies scored between IV and V and three scored I).
HADS total: Sensitivity of HADS total using the threshold 10 or 11 was 0.80 (95% CI 0.75 to 0.84) and specificity was 0.74 (95% CI 0.71 to 0.77) in seven studies for screening for mental disorders. Sensitivity of HADS total using the threshold 15 was 0.87 (95% CI 0.73 to 0.95) and specificity was 0.88 (95% CI 0.84 to 0.90) in two studies for screening for depressive disorder.
HADS depression subscale: Sensitivity of the HADS depression subscale using a threshold of 5 was 0.84 (95% CI 0.79 to 0.88) and specificity was 0.50 (95% CI 0.46 to 0.55) in two studies for screening for mental disorder. Sensitivity of the HADS depression subscale using the threshold 7 was 0.86 (95% CI 0.76 to 0.93) and specificity was 0.81 (95% CI 0.78 to 0.84) in four studies for screening for depressive disorder.
HADS anxiety subscale: Sensitivity of the HADS anxiety subscale using the threshold 7 or 8 was 0.73 (95% CI 0.68 to 0.77) and specificity was 0.65 (95% CI 0.61 to 0.67) in four studies for screening for mental disorder. Sensitivity of the HADS anxiety subscale using the threshold 10 or 11 was 0.63 (95% CI 0.53 to 0.73) and specificity was 0.83 (95% CI 0.80 to 0.86) in four studies for screening for depressive disorder.
There were high levels of statistical heterogeneity in many of the specificity analyses. Other results were presented in the review.