Skip to main content

Table 3 Yield of different algorithms in TB detection and utilization of tests

From: Comparing tuberculosis symptom screening to chest X-ray with artificial intelligence in an active case finding campaign in Northeast Nigeria

Algorithm

Sensitivity

Specificity

PPV

NPV

People with TB missed

Tests used

Abnormality ≥ 0.30

95.3%

29.8%

16.9%

97.7%

4

738

Abnormality ≥ 0.50

89.4%

62.8%

26.4%

97.5%

9

424

Any cough

62.4%

37.4%

12.9%

86.9%

32

639

Cough ≥ 2 weeks

40%

61.5%

13.5%

87.3%

51

394

Any symptom

90.6%

11.4%

13.3%

89.1%

8

906

Cough OR fever

67.1%

29.7%

12.5%

85.8%

28

715

Abnormality ≥ 0.30 OR any cough

100%

5.9%

13.7%

100%

0

966

Abnormality ≥ 0.30 OR cough ≥ 2 weeks

100%

13.4%

14.7%

100%

0

896

Abnormality ≥ 0.30 OR any symptom

100%

0%

13%

-

0

1021

Abnormality ≥ 0.50 OR any cough

100%

20.5%

15.8%

100%

0

829

Abnormality ≥ 0.50 OR cough ≥ 2 weeks

97.7%

36%

18.6%

99%

2

682

Abnormality ≥ 0.50 OR any symptom

100%

5.7%

13.7%

100%

0

968

  1. PPV positive predictive value, NPV negative predictive value