Estimating undiagnosed HIV infections by age group in Japan: an extended age-dependent back-calculation

Abstract

Understanding the number of undiagnosed HIV-infected individuals by age is essential for improving the test-and-treat strategy. We developed an extended back-calculation by age group to investigate the situation in Japan, describing the data-generating process of AIDS cases and HIV diagnoses as a function of age and time. We considered the incubation period as a function of both age and time since infection, and estimated the number of new HIV infections and annual diagnosis rate by age and time. The diagnosed proportion of HIV infections at the end of 2022 was estimated to be 93.2 % (95 % CI: 90.2, 95.8) in their 20s, 90.4 % (95 % CI: 87.0, 93.7) in their 40s, 90.3 % (95 % CI: 86.9, 93.5) in their 50s or older, and 89.4 % (95 % CI: 85.1, 93.2) in their 30s. The annual rate of diagnosis of people in their 40s decreased from 16.9 % in 2015-2019 to 14.8 % in 2020-22. Despite increasing trend in diagnostic rate, the estimate for those in their 50s was as small as 13.6 % (95 % CI: 8.5, 19.4) in 2020-2022. We identified a difficulty in diagnosing HIV-infected individuals aged 40 and older. The absolute number of infections is greater among those in their 30s than 40s, but the AIDS incidence is the opposite, suggesting that older individuals would require more customized (and easy to access) opportunities for diagnosis.

Keywords: Antiretroviral therapy; Ascertainment; Retrovirus; Test-and-treat; Treatment as prevention; U=U.

Fig. 1

Estimate of annual new HIV infections, annual HIV diagnosis rate, people with a diagnosed HIV infection, and AIDS patients who are previously undiagnosed as HIV infected. (A) Yearly number of new HIV infections, by age-dependent incubation period, which are modeled using a discrete Weibull distribution. A five-step function is assumed as the yearly number of new HIV infections in the model. Age groups are categorized by colors (blue is people in their 20s, yellow 30s, pink 40s and orange 50s and over). Dots represent estimated values and ribbons the 95 % confidence interval. (B) Annual probability diagnosis rate, assuming age-dependent incubation periods, based on age-dependent incubation periods modeled by a discrete Weibull distribution. A step function for every 5 years was used as the yearly probability diagnosis rate in the model. (C) Estimate of HIV diagnosed and 95 % confidence interval by age group. Dots represent observed values, solid lines estimated values and ribbons 95 % confidence interval. (D) Estimate of AIDS patients and 95 % confidence interval by age group.

Fig. 2

Estimate of the population with undiagnosed HIV infections. The number of people with undiagnosed HIV infection is represented by dots and the 95 % confidence interval by the ribbon. We divided age groups by colors, with those in their 20s in blue, 30s in yellow, 40s in pink and 50s and over in orange.

Fig. 3

Trends in diagnosis rate, with and without AIDS cases. (A) Estimation of diagnosis rate trends in all people diagnosed with HIV (including those with AIDS) by age groups. We divided age groups by colors, with those in their 20s in blue, 30s in yellow, 40s in pink and 50s and over in orange. Dots represent estimated values and ribbons show the 95 % confidence interval. (B) Estimation of diagnosis rate trends excluding people with AIDS by age groups.

Fig. 4

Cohort estimate of annual new HIV incidence, annual diagnosis rate, undiagnosed HIV-infected cases, and the number of AIDS cases. (A) The estimated value of the yearly number of HIV infections is shown by cohort: those born from 1960 to 1969 in red, 1965 to 1974 in olive green, 1970 to 1979 in green, 1975 to 1984 in blue, and 1980 to 1989 in purple. (B) The estimated value of yearly probability of diagnosis by cohort is shown by the same colors. (C) Estimate of undiagnosed HIV infections by cohort. (D) Estimate of the diagnosed population with AIDS.

Fig. S1. Yearly number of new HIV infections by age group. Age groups are categorized by colors (blue represents those in their 20s, yellow 30s, pink 40s and orange 50s and over). Dots represent estimated values and ribbons show 95 % confidence interval.

Fig. S2. Yearly probability HIV diagnosis rate by age group. Age groups are shown by colors, where blue represents those in their 20s, yellow 30s, pink 40s and orange 50s and over. Dots represent estimated values and ribbons show the 95 % confidence interval.

Fig. S3. Estimate of the number of diagnosed HIV infections and 95 % confidence interval by age group. Age groups are categorized by colors (blue is those in their 20s, yellow 30s, pink 40s and orange 50s and over). Dots represent estimated values and ribbons show 95 % confidence interval.

Fig. S4. Estimate of AIDS patients and 95 % confidence interval by age group. Age groups are categorized by colors (blue represents those in their 20s, yellow 30s, pink 40s and orange 50s and over). Dots represent estimated values and ribbons show the 95 % confidence interval.

Fig. S5. Estimate of the population with undiagnosed HIV infections by age group. Age groups are categorized by colors (blue represents those in their 20s, yellow 30s, pink 40s and orange 50s and over). Dots represent estimated values and ribbons show the 95 % confidence interval.

Fig. S6. Trends of diagnosis rate for all HIV infections, including people with AIDS, by age group. Age groups are categorized by colors (blue represents those in their 20s, yellow 30s, pink 40s and orange 50s and over). Dots represent estimated values and ribbons show the 95 % confidence interval.

Fig. S7. Trends of diagnosis rate excluding people with AIDS, by age group. Age groups are categorized by colors (blue represents those in their 20s, yellow 30s, pink 40s and orange 50s and over). Dots represent estimated values and ribbons show the 95 % confidence interval.