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

Background HIV transmission among men having sex with men remains a concerning issue in Japan. Therefore, the purpose of the present study was to estimate the effective reproduction number R ( t ) of HIV by coupling a renewal process model with an existing extended back-calculation model in a Bayesian framework. Methods We analyzed the annual number of HIV diagnoses and AIDS cases among Japanese males from 1985 through 2025. Using a Bayesian back-calculation with penalized B -splines for the incidence and the diagnosis rate of HIV with a renewal equation, we estimated R ( t ) for Japan and different geographic regions within the country. Results From 2008 to 2010, we observed a consistent median value of R ( t ) < 1, with the most recent estimate of R ( t ) being the most uncertain. R ( t ) < 1 was apparent in major metropolitan areas, including Tokyo, Kanto–Koshinetsu, and Kinki, where even the upper bound of the 95% credible interval fell below 1 in the early-to-mid 2010s. The annual HIV diagnosis rate declined from shortly before the start of the COVID-19 pandemic, and the estimated proportion of HIV-infected individuals has increased only slowly in recent years, with the proportion of diagnosed individuals not reaching 95% or even 90% of the true total. Conclusion Our general R ( t ) value indicates that the control of HIV/AIDS in Japan has been successful over the long term, including during the COVID-19 pandemic. However, the uncertainty of recent R ( t ) values suggests the need for regular updates of the estimate; reforming the testing and treatment strategies in Japan is vital to accomplishing this goal.