Table 2.
Attribute of Viral Load Testing Approach | Values for Factor Considered |
Maximum Fully Loaded Cost of Viral Load Test for Reduced Net DALYs Compared With Base Scenario |
||
---|---|---|---|---|
(1) Delay in result | No delay | $23 | ||
3 mo delaya | $22 | |||
(2) Sensitivity/specificity relative to plasma for 1000 copies/mL cutoff | Sensitivity/Specificity | |||
58/96 | 84/89a | $21 | $22 | |
60/99 | 89/93 | $20 | $21 | |
69/97 | 90/85 | $20 | $22 | |
70/79 | 95/90 | $19 | $21 | |
74/75 | 96/65 | $17 | $15 | |
78/93 | 96/77 | $20 | $17 | |
79/71 | 99/69 | $15 | $20 | |
80/84 | 100/100 | $19 | $23 | |
82/96 | $21 | |||
(3) VL threshold used to define first-line ART failure (copies/mL) (cutoff for qualitative assay) | 200 | $20 | ||
500 | $23 | |||
1000a | $22 | |||
3000 | $19 | |||
5000 | $19 | |||
(4) Proportion of tests failed or result lostb | 0.80 | $5 | ||
0.50 | $16 | |||
0.30 | $19 | |||
0.15a | $22 | |||
0.00 | $24 | |||
(5) Probability of differentiated care if VL < 1000 being implemented | 0.5 | $14 | ||
0.6 | $16 | |||
0.7 | $19 | |||
0.8a | $22 | |||
0.9 | $24 | |||
(6) Difference in coverage of population with VL testing | 70% | $20 | ||
75%a | $22 | |||
80% | $23 | |||
85% | $25 | |||
90% | $26 | |||
95% | $26 | |||
100% | $28 | |||
(7) Probability of switch to second-line ART per 3 mo once failure definition met | 0.05 | $17 | ||
0.10 | $20 | |||
0.15 | $20 | |||
0.20a | $22 | |||
0.25 | $22 | |||
0.30 | $23 | |||
0.40 | $23 | |||
0.60 | $23 | |||
1.00 | $25 | |||
(8) Probability of ART interruption/loss to follow up (per 3 mo) | 0.030 | $3 | ||
0.025 | $14 | |||
0.020a | $22 | |||
0.015 | $29 | |||
0.010 | $41 | |||
0.005 | $48 |
Abbreviations: ART, antiretroviral therapy; DALYs, disability adjusted life years; VL, viral load.
a Base scenario (VL cost $22).
b Assumed that there is a 90% chance that the cost of the test is nevertheless incurred.