Cost-effectiveness of HIV drug resistance testing to inform switching to second line antiretroviral therapy in low income settings

Abstract

Background: To guide future need for cheap resistance tests for use in low income settings, we assessed cost-effectiveness of drug resistance testing as part of monitoring of people on first line ART - with switching from first to second line ART being conditional on NNRTI drug resistance mutations being identified.

Methods: An individual level simulation model of HIV transmission, progression and the effect of ART which accounts for adherence and resistance development was used to compare outcomes of various potential monitoring strategies in a typical low income setting in sub-Saharan Africa. Underlying monitoring strategies considered were based on clinical disease, CD4 count or viral load. Within each we considered a strategy in which no further measures are performed, one with a viral load measure to confirm failure, and one with both a viral load measure and a resistance test. Predicted outcomes were assessed over 2015-2025 in terms of viral suppression, first line failure, switching to second line regimen, death, HIV incidence, disability-adjusted-life-years averted and costs. Potential future low costs of resistance tests ($30) were used.

Results: The most effective strategy, in terms of DALYs averted, was one using viral load monitoring without confirmation. The incremental cost-effectiveness ratio for this strategy was $2113 (the same as that for viral load monitoring with confirmation). ART monitoring strategies which involved resistance testing did not emerge as being more effective or cost effective than strategies not using it. The slightly reduced ART costs resulting from use of resistance testing, due to less use of second line regimens, was of similar magnitude to the costs of resistance tests.

Conclusion: Use of resistance testing at the time of first line failure as part of the decision whether to switch to second line therapy was not cost-effective, even though the test was assumed to be very inexpensive.

Figure 1. Monitoring strategies and switch criteria.

Figure 2. Outcomes by monitoring strategy.

Mean over 2015–2025 (except for cost and DALYs where total over this period is given). VL - viral load.

Figure 3. Breakdown of costs by strategy - total discounted cost 2015–2025, in $million.

See Figure 1 for legend for strategies.

Figure 4. Incremental costs and DALYs averted for monitoring strategies over 10 years, compared with no monitoring, no second line.

See Figure 1 for legend for strategies.

Figure 5. One way sensitivity analyses with the following changes from the base scenario (a) poorer adherence profile (b) 20 year time horizon (c) resistance test cost of $15 instead of $30 (d) cost of second line drugs made the same as cost of 1^st line (e) initiation of ART at CD4 count below 500 cells/mm³ rather than below 350 cells/mm³ (f) if boosted PI drugs had same potency as other drugs, and risk of resistance accumulation similar to NNRTI.