Bovine Tuberculosis in Cattle in the Highlands of Cameroon: Seroprevalence Estimates and Rates of Tuberculin Skin Test Reactors at Modified Cut-Offs

Table 3

Predictive values and likelihood ratios at the ≥2 mm, ≥3 mm, and ≥4 mm cutoff points for tuberculin skin tests and lateral flow assay of cattle reactors in Cameroon.

Cutoff point	Test predictive value; % (95% CI)		Diagnostic likelihood ratio; (95% CI)
	Positive result	Negative result	LR+	LR−
(a) For SICCT-BT test against antibovine TB Ab assay
≥2 mm	34.05 (29.16–38.50)	94.41 (91.66–96.41)	2.54 (2.03–3.08)	0.29 (0.45–0.18)
≥3 mm	29.55 (25.32–33.13)	97.58 (95.42–98.79)	2.77 (2.24–3.27)	0.16 (0.32–0.08)
≥4 mm	14.67 (12.15–15.94)	100 (98.88–100)	2.87 (2.31–3.17)	0* (0.19–0)
(b) For SIT-BT test against antibovine TB Ab assay
≥4 mm	33.03 (28.13–37.61)	93.53 (90.87–95.58)	2.45 (1.94–2.99)	0.34 (0.50–0.23)

*The perfect diagnostic test would be expected to have an LR− equal to zero and an LR+ equal to infinity (producing no false negatives, but detecting all negatives and detecting all positives, and generating no false positives). The best test therefore for excluding a disease is the one with the lowest LR− and the test with the highest LR+ is the best for detecting disease [28].