TABLE 3.

Results of meta-analyses for estimated stool Xpert sensitivity and specificity^a

Comparison	Main results			Results of sensitivity analysis excluding the study that did not use culture as reference standard
	No. of studies included (no. of children included)	Pooled sensitivity (%) (95% CI); I2 statistic (95% CI)	Pooled specificity (%) (95% CI); I2 statistic (95% CI)	No. of studies included (no. of children included)	Pooled sensitivity (%) (95% CI); I2 statistic (95% CI)	Pooled specificity (%) (95% CI); I2 statistic (95% CI)
Stool Xpert against microbiological reference standard	9b (1,681)	67 (52–79); 83 (72–93)	99 (98–99); 62 (35–90)	8f (1,644)	64 (49–76); 81 (69–93)	99 (98–100); 61 (31–91)
Stool Xpert against clinical reference standard	5c (869)	22 (9.0–44); 95 (92–98)	100 (66–100); 78 (59–97)	Not applicable	Not applicable	Not applicable
Stool Xpert against microbiological reference standard for children with HIV	5d (395)	79 (68–87); 0 (0–100)	99 (94–100); 35 (0–99)	5g (379)	80 (68–88); 0 (0–100)	99 (94–100); 51 (0–100)
Stool Xpert against microbiological reference standard for HIV-negative children	7e (974)	61 (40–79); 39 (0–100)	99 (98–100); 56 (13–100)	Not applicable	Not applicable	Not applicable

^aThe I² statistic was used to quantify the effect of between-study heterogeneity.

^bFrom references 14–22.

^cFrom references 16–20.

^dFrom references 15, 17, 18, 21, and 22.

^eFrom references 14–17, 19, 20, and 22.

^fFrom references 14 and 16–22.

^gFrom references 17, 18, 21, and 22.