Table 3.
Proportion of setting scenarios in which tenofovir, lamivudine, and dolutegravir brings direct health benefit or is cost-effective, over 20 years
| Tenofovir, lamivudine, and dolutegravir brings direct health benefit (ie, DALYs averted) | Tenofovir, lamivudine, and dolutegravir is cost-effective (ie, net DALYs are averted) | ||
|---|---|---|---|
| Overall | 83% | 87% | |
| Sensitivity analyses—restricted to setting scenarios for which: | |||
| Proportion of ART-naive ART initiators with non-nucleoside reverse transcriptase inhibitor resistance in 2018 is in the range 10–30% | 89% | 94% | |
| Proportion of ART-naive ART initiators with non-nucleoside reverse transcriptase inhibitor resistance in 2018 <5% | 79% | 82% | |
| HIV prevalence 10–28% in 2018 | 86% | 91% | |
| Proportion of people on ART with viral load <1000 copies per mL 90–97% in 2018 | 71% | 77% | |
| Proportion of people on ART with viral load <1000 copies per mL 72–85% in 2018 | 93% | 95% | |
| Viral load monitoring and switching to second-line ART operating well* | 75% | 87% | |
| Viral load monitoring and switching to second-line ART operating poorly or not at all† | 89% | 87% | |
| 16–19% of women giving birth per year | 78% | 82% | |
| 5–12% of women giving birth per year | 85% | 89% | |
| 1·25 times increased risk of non-HIV death due to weight gain | 76% | 85% | |
| ≤1·05 times increased risk of non-HIV death due to weight gain | 85% | 88% | |
| Absolute additional risk of stillbirth or neonatal death due to weight gain is 0·05% | 82% | 85% | |
| Absolute additional risk of stillbirth or neonatal death due to weight gain is 0·30% | 82% | 84% | |
| Risk of neural tube defects due to dolutegravir is 0·61% | 82% | 87% | |
| Reduced impact of non-nucleoside reverse transcriptase inhibitor mutations on efavirenz activity‡ | 73% | 78% | |
| Halving of the risk of resistance mutations emerging for all drugs | 82% | 86% | |
ART=antiretroviral therapy. DALY=disability-adjusted life-year.
Viral load monitoring and switching to second-line ART operating poorly means the probability of each scheduled viral load test being done is 0·1 or 0; the probability per 3 months of a switch to second-line ART once failure criteria are met is 0 or 0·2.
Viral load monitoring and switching to second-line ART operating well means the probability of each scheduled viral load test being done is 0·85; the probability per 3 months of a switch to second-line ART once failure criteria are met is 0·5.
With this assumption, the mean odds ratio for viral load >1000 copies per mL at 1 year from the start of ART associated with pre-treatment on nucleoside reverse transcriptase inhibitor resistance is 1·75, compared with 3·3 for the overall result (appendix p 36).