Table 3.
Predicted effect of regimen policies on intermediate outcomes, DALYs, and net DALYs
| Tenofovir, lamivudine, and efavirenz for all | Tenofovir, lamivudine, and dolutegravir dependent on viral suppression and intention to have (more) children | Tenofovir, lamivudine, and dolutegravir dependent on intention to have (more) children only | Tenofovir, lamivudine, and dolutegravir dependent on viral suppression only | Tenofovir, lamivudine, and dolutegravir for all | ||
|---|---|---|---|---|---|---|
| Proportion on efavirenz | 92% | 52%;–40% (−41 to −39; −52 to −22) | 42%; −50% (−51% to −49; −61 to −39) | 15%; −77% (−78 to −76; −92 to −47) | 0%; −92% (−93 to −91; −100 to −78) | |
| Proportion on dolutegravir | 0% | 43%; 43% (42 to 44; 22 to 57) | 54%; 54% (53 to 55; 42 to 63) | 85%; 83% (82 to 85; 47 to 97) | 98%; 98% (97 to 99; 94 to 100) | |
| Proportion on atazanavir | 8% | 5%; −3% (−4 to −2; −9 to 0) | 4%; −3% (−4 to −2; −9 to −1) | 2%; −6% (−7 to −5; −16% to 0) | 2%; −6% (−7 to −5; −16 to 0) | |
| Proportion on zidovudine* | 8% | 5%; −2% (−3 to −1; −8 to 0) | 4%; −4% (−5 to −3; −10 to −2) | 4%; −4% (−5 to −3; −13 to 0) | 2%; −6% (−7 to −5; −16 to 0) | |
| Proportion of people on ART with viral load <1000 copies per mL | ||||||
| Mean over 1 year | 84% | 85%; 0% (0 to 0; 0 to 2) | 87%; 2% (2 to 2; 0 to 5) | 85%; 1% (1 to 1; −1 to 3) | 90%; 6% (6 to 6; 1 to 11) | |
| Mean over 5 years | 84% | 86%; 2% (2 to 2; 0 to 4) | 87%; 3% (3 to 3; 0 to 7) | 88%; 4% (4 to 4; 1 to 8) | 91%; 7% (7 to 7; 1 to 14) | |
| Mean over 20 years | 82% | 87%; 5% (5 to 5; 1 to 9) | 87%; 5% (5 to 5; 1 to 11) | 91%; 9% (9 to 9; 3 to 16) | 91%; 10% (10 to 10; 2 to 18) | |
| Of people with baseline viral load >1000 copies per mL and Lys65Arg and Met184Val mutations at baseline, percentage of people on ART with viral load <1000 copies per mL | ||||||
| Mean over 1 year | 7% | 9%; 3% (2 to 4; −1 to 8) | 23%; 16% (15 to 17; 4 to 28) | 14%; 7% (6 to 8; −1 to 21) | 49%; 42% (40 to 44; 15 to 64) | |
| Mean over 5 years | 22% | 26%; 4% (3 to 5%; −1 to 10) | 35%; 13% (12 to 14; −3 to 32) | 33%; 11% (9 to 13; −1 to 25) | 55%; 32% (29 to 35; −3 to 72%) | |
| Mean over 20 years | 50% | 50%; 0% (−1 to 1; −9 to 8) | 52%; 2% (−1 to 5; −16 to 46) | 53%; 3% (1 to 5; −15 to 18) | 57%; 7% (3 to 11; −22 to 64) | |
| AIDS death rate in people on ART (per 100 person-years) | 1·70 | 1·25; −0·46 (−0·48 to −0·44; −0·94 to −0·12) | 1·08; −0·63 (−0·66 to −0·60; −1·34 to −0·14) | 0·94; −0·76 (−0·79 to −0·73; −1·51 to −0·24) | 0·72; −0·98 (−1·02 to −0·94; −2·02 to −0·24) | |
| AIDS death rate in people on ART with viral load >1000 copies per mL (per 100 person-years) | 7·05 | 6·29; −0·76 (−0·82 to −0·70; −1·41 to −0·12) | 5·71; −1·34 (−1·44 to −1·24; −2·54 to −0·45) | 5·92; −1·12 (−1·22 to −1·02; −2·22 to −0·14) | 4·77; −2·28 (−2·08 to −2·48; −4·36 to −0·48) | |
| AIDS death rate in people on ART with viral load >1000 copies per mL and CD4 count <200 cells per μL (per 100 person-years) | 18·07 | 17·22; −0·85 (−1·24 to −0·66; −2·01 to 0·91) | 16·05; −2·02 (−2·26 to 1·78; −4·15 to −0·11) | 17·05; −1·03 (−1·34 to −0·72; −3·41 to 1·45) | 14·77; −3·30 (−3·78 to −2·82; −8·43 to 0·03) | |
| Proportion of all HIV-positive people with a dolutegravir resistance mutation | 0% | 2·6%; 2·6% (2·3 to 2·7; 0·4 to 7·7) | 4·0%; 4·0% (3·8 to 4·2; 0·6 to 11·0) | 4·4%; 4·4% (4·1 to 4·7; 0·7 to 12·9) | 6·7%; 6·7% (6·2 to 7·0; 1·2 to 18·5) | |
| Proportion of all HIV-positive people with an efavirenz resistance mutation | 28% | 22%; −6% (−6 to −6; −12 to −1) | 20%; −8% (−8 to −8; −14 to −3) | 15%; −13% (−13 to −13; −22 to −6) | 13%; −15% (−15 to 15; −24 to −8) | |
| Adverse birth outcomes among women with HIV (percentage of pregnancies) | ||||||
| Mother-to-child transmission | 4·2% | 3·9%; −0·2% (−0·2 to −0·2; −0·8 to 0·3) | 3·8%; −0·3% (−0·3 to −0·3; −1·1 to 0·3) | 2·9%; −1·2% (−1·2 to −1·2; −2·5 to −0·3) | 2·8%; −1·4 (−1·4 to −1·4; −2·9 to −0·3) | |
| Neural tube defect due to dolutegravir† | 0% | 0·02%; 0·02% (0·02 to 0·02; 0·0 to 0·08) | 0·03%; 0·03% (0·03 to 0·03; 0·0 to 0·09%) | 0·52%; 0·52% (0·49 to 0·55; −0·12 to 1·38) | 0·60%; 0·60% (0·58 to 0·62; 0·15 to 1·51) | |
| Costs‡ (annual difference compared with tenofovir, lamivudine, and efavirenz)§ | .. | −$5·3 million (−$5·8 million to −$4·8 million; −$19·6 million to $2·0 million) | −$5·3 million (−$4·8 million to −$5·8 million; −$20·4 million to $3·1 million) | −$10·5 million (−$11·3 million to $9·7 million; −$37·4 million to $1·4 million) | −$9·7 million (−$10·6 million to −$8·8 million; −$38·0 million to $3·9 million) | |
| DALYs averted‡ (per year, compared with tenofovir, lamivudine, and efavirenz)¶ | .. | 22 300 (21 300 to 23 300; 1400 to 53 000) | 32 800 (31 300 to 34 300; 3900 to 78 200) | 39 500 (37 000 to 42 000; 7900 to 87 900) | 58 200 (55 700 to 61 300; 11 500 to 138 300) | |
| Incremental cost-effectiveness ratio‖ | Dominated | Dominated | Dominated | Reference | $44 | |
| Net DALYs (per year, compared with tenofovir, lamivudine, and efavirenz)¶ | .. | 32 900 (31 400 to 34 400; 2700 to 84 200) | 43 500 (41 700 to 45 300; 8600 to 103 100) | 60 600 (58 000 to 63 200; 11 800 to 143 400) | 77 700 (74 700 to 80 700; 20 800 to 177 600) | |
Data are mean proportion and mean difference (95% CI; 90% range reflecting variation across setting scenarios) compared with the policy of tenofovir, lamivudine, and efavirenz for all, unless otherwise indicated. DALYs and net DALYs were measured over 20 years (2018–38) and represent the mean over 3 month periods, unless otherwise stated. Costs are in US$. ART=antiretrovoral therapy. DALY=disability adjusted life-year.
*
As opposed to tenofovir disoproxil fumarate, all are on lamivudine.
†
Neural tube defects potentially caused by dolutegravir if possible signal is confirmed.
‡
In the context of an adult population of 10 million people with HIV prevalence as in table 2. Costs and DALYs are discounted at 3% per annum.
§
Values are mean and mean difference (95% CI; 90% range reflecting variation across setting scenarios).
¶
Values are n per year (95% CI; 90% range reflecting variation across setting scenarios).
‖
Policies are dominated if there is another policy with both lower cost and more DALYs averted. The reference policy is selected among non-dominated policies.