Table 2. Percentage of correct classification of VF by Alere q and DBS methods according to plasma VL categories.

Plasma VL categories	Target not detected	<40–1000 copies/ml	Specificity	1000–10,000 copies/ml	>10,000 copies/ml	Sensitivity
Alere q	22% (21/94)	14% (7/51)	19% (28/145) (95% CI 13–27%)	100% (52/52)	100% (101/101)	100% (153/153) (95% CI 97–100%)
Abbott DBS	89% (84/94)	75% (38/51)	84% (122/145) (95% CI 77–89%)	75% (39/52)	99% (97/98)	91% (136/150) (95% CI 85–95%)
Roche FVE	100% (90/90)	98% (49/50)	99% (139/140) (95% CI 95–100%)	0% (0/50)	80% (76/95)	52% (76/145) (95% CI 44–61%)

Definition of correct classification of VF by Alere q and DBS methods according to plasma VL categories using the VF threshold of 1000 copies/ml: when plasma VL < 1000 copies/ml (both target not detected and <40-1000copies/ml categories) then Alere q/DBS VL <1000 copies/ml (specificity in identifying absence of VF), when plasma VL > = 1000 copies/ml (both 1000–10,000 and >10,000 copies/ml categories) then Alere q/DBS VL > = 1000 copies/ml (sensitivity in identifying VF). Alere q showed 100% sensitivity, however there was reduced specificity in both the target not detected and <40–1000 copies/ml plasma VL categories.(22% and 14% respectively). For Abbott DBS, sensitivity was best above 10,000 copies/ml, however sensitivity was 75% in the 1000–10,000 copies/ml category, and specificity was 75% in the <40–1000 copies/ml category. Roche FVE demonstrated the opposite performance with 0% sensitivity in the 1000–10,000 copies/ml category but had the best specificity. Correct classification was best for all methods at >10,000 copies/ml.