Twelve studies with a total of 6,394 participants were included in the review.
In all 12 studies the index test formed part of the reference standard, indicating a potential for incorporation bias.
Primary care.
For studies using a diagnostic algorithm threshold (4 studies, n=3,053), the sensitivity ranged from 0.71 (specificity 0.92) to 0.86 (specificity 0.94), and the specificity ranged from 0.90 (sensitivity 0.83) to 0.98 (sensitivity 0.73). The pooled estimates for sensitivity and specificity were 0.77 (95% confidence interval, CI: 0.71, 0.84) and 0.94 (95% CI: 0.90, 0.97), respectively. There was significant between-study heterogeneity in the estimates of specificity. The one study using a summary score threshold reported a sensitivity and specificity of 0.88.
Secondary care.
For studies using a diagnostic algorithm threshold (4 studies, n=1,411), the sensitivity ranged from 0.43 (specificity 0.92) to 0.84 (specificity 0.92), and the specificity ranged from 0.91 (sensitivity 0.55) to 0.94 (sensitivity 0.69). For studies using a summary score threshold (3 studies, n=1,402), the sensitivity ranged from 0.54 (specificity 0.90) to 0.94 (specificity 0.91), and the specificity ranged from 0.89 (sensitivity 0.91) to 0.91 (sensitivity 0.94).