Back

Treating earlier, spending less: cost-effectiveness and budget impact of immediate versus delayed antiretroviral therapy for HIV in Japan

Taniguchi, T.; Imahashi, M.; Sato, D.; Noda, T.

2026-08-26 hiv aids
10.64898/2026.08.23.26361165 medRxiv
Show abstract

Background. In Japan, lifelong antiretroviral therapy (ART) is funded through the physical disability (immune dysfunction) certification pathway, which requires two laboratory assessments four weeks or more apart. This statutory pathway, rather than clinical need, contributes to a median diagnosis-to-ART interval of about 42 days. We evaluated relaxing or reforming it to permit immediate ART. Methods. We developed a stochastic individual-based microsimulation of HIV in Japan, linked to a payer-perspective cost-effectiveness analysis over a 40-year horizon after a 20-year burn-in, calibrated to national surveillance and cascade data. We compared immediate ART with one-month (primary) and two-month (secondary) delays. Costs and quality-adjusted life-years (QALYs) were discounted at 2% per year; uncertainty was assessed across 200 seeds and by probabilistic and one-way sensitivity analyses. Findings. Against the one-month delay, immediate ART averted 2,991 infections and 1,761 deaths among people with HIV over 40 years, gained 9,071 QALYs, and reduced discounted costs by JPY 54.4 billion (net monetary benefit JPY 99.7 billion). The two-month comparison saved JPY 82.5 billion (4,535 infections, 2,688 deaths averted). Immediate ART was dominant at the base case and in all 1,000 probabilistic sensitivity-analysis iterations; cumulative savings offset the early investment within 11 to 12 years, and sensitivity analyses altered only its magnitude. Interpretation. Permitting immediate ART by reforming the certification pathway was projected to reduce HIV incidence, improve population health, and save public-payer costs within the second decade, supporting consideration of statutory reform.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.