The researchers coupled an existing extended back-calculation model with a renewal process framework inside a Bayesian statistical structure. The team analyzed annual counts of HIV diagnoses and AIDS cases among Japanese males across four decades, using penalized B-splines to model both the incidence of HIV and the diagnosis rate. The Bayesian approach allowed uncertainty to propagate through the estimation, and made it possible to produce separate estimates for different geographic regions of Japan as well as for the country as a whole.

The headline result concerns the period from 2008 to 2010, when the median estimate of R(t) dipped consistently below 1. In the major metropolitan regions — Tokyo, the Kanto-Koshinetsu area, and Kinki — the picture over the last decade is even clearer: the upper bound of the credible interval for R(t) fell below 1. When even the upper limit of a credible interval sits beneath the replacement threshold, the data support a genuine decline in transmission, not merely a point estimate that happens to hover near the boundary.

The most recent estimates carry the greatest uncertainty, the authors report. They emphasize that this uncertainty does not undermine the long-term trend, but it does mean the latest figures should be treated as provisional and updated regularly as new surveillance data arrive.

The study also tracked the annual HIV diagnosis rate, which declined from shortly before the start of the COVID-19 pandemic. According to the authors, the estimated proportion of HIV-infected individuals has not continuously increased in recent years.

That said, the analysis highlights a persistent gap in Japan’s response. The estimated proportion of diagnosed individuals has not reached 95 percent — or even 90 percent — of the true total of people living with HIV.

Overall, the authors conclude that long-term control of HIV/AIDS in Japan has been successful, with the effective reproduction number indicating sustained sub-critical transmission through the COVID-19 years. But they frame the recent uncertainty in R(t) as a warning rather than a victory lap: without regular re-estimation, supported by reformed testing and treatment strategies, Japan cannot confirm that the decline will hold.

The study’s contribution is methodological as much as substantive — a Bayesian extended back-calculation that turns four decades of routine surveillance into a continuously updatable measure of epidemic control, one that can be refreshed as new data arrive rather than treated as a one-off assessment.

Subject of Research: Estimation of the effective reproduction number of HIV infections in Japan from 1985 to 2025 using Bayesian extended back-calculation

Article Title: Estimating the effective reproduction number of HIV infections in Japan from 1985 to 2025: a Bayesian extended back calculation

Article References: Fujiwara, S., & Nishiura, H. (2026). Estimating the effective reproduction number of HIV infections in Japan from 1985 to 2025: a Bayesian extended back calculation. Frontiers in Public Health, 14, 1944492. https://doi.org/10.3389/fpubh.2026.1944492

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.1944492

Keywords: HIV, AIDS, effective reproduction number, Japan, Bayesian back-calculation, men who have sex with men, mathematical modeling, epidemiology, COVID-19, testing coverage, public health, renewal equation