|
Subgroup (Number of studies)
|
Summary diagnostic odds ratio (95% CI)a
|
Chi2 square test of heterogeneity
|
P value for heterogeneityb
|
|---|
|
Study design
|
|
Cross-sectional (10)
|
403.18 (120.05–1354.05)
|
36.66
|
< 0.001
|
|
Case-control (8)
|
73.86 (20.40–267.48)
|
34.01
|
< 0.001
|
|
Target sequence
|
|
IS6110 (11)
|
144.74 (51.24–408.86)
|
39.39
|
< 0.001
|
|
Other target (7)
|
297.17 (30.22–2921.73)
|
66.27
|
< 0.001
|
|
Specimen type
|
|
Respiratory (11)
|
598.12 (269.12–1329.32)
|
19.09
|
0.039
|
|
Non-respiratory (5)
|
12.39 (6.67–22.73)
|
3.57
|
0.468
|
|
Both (2)
|
202.47 (64.68–633.83)
|
0.00
|
0.944
|
|
Region of study
|
|
TB high-burden country (8)
|
281.86 (37.69–2107.75)
|
90.46
|
< 0.001
|
|
Other country (10)
|
160.73 (72.80–354.83)
|
15.17
|
0.086
|
|
Quality of study
|
|
High-quality (7)
|
926.97 (303.59–2830.38)
|
12.83
|
0.046
|
|
Medium-quality (8)
|
76.77 (22.98–256.50)
|
26.65
|
< 0.001
|
|
Low-quality (3)
|
72.35 (4.47–1170.04)
|
19.07
|
< 0.001
|
- aRandom effects model
- bchi-square or Fisher’s exact test for heterogeneity; high-quality study: a study that had no domain with a high risk of bias and no domain with high applicability concerns; medium/moderate-quality study: a study that had domain with a unclear risk of bias or domain with unclear applicability concerns; low-quality study: a study that had a domain with a high risk of bias and domain with high applicability concerns
