Background
Accurately predicting the age and sex distribution of HIV over time is critical for precision interventions and resource planning. Progress towards epidemic control and changes in sexual partnership patterns over time impact this distribution. Characterizing age-specific sexual partner matrices may improve prediction accuracy and better inform programming for epidemic control.
Methods
We used data from PHIA surveys in Zambia and Botswana to create partner matrices. For respondents with viral nonsuppression (NVLS=copies >1000/mL) we extracted the mean number of partners in the last 12 months, stratified by age, sex, and partner HIV status. Partner matrices in Zambia from 2021 against those from 2016 (prior NVLS data unavailable for Botswana) were compared. Intergenerational partnerships (IP), defined as sexual partner +/-10 years of age, by HIV status, age, and sex, and condomless sex (CS) among PWH with NVLS were estimated. Estimates and 95% confidence intervals (95% CI) are weighted.
Results
In Zambia mean number of serodifferent female partners increased among NVLS males in younger age groups and decreased in older males where seroconcordant female partners increased (Figure 1). Mean serodifferent same sex partners decreased among males aged 40-44 and seroconcordant increased among NVLS males 45-49. In Zambia, prevalence of IP increased 23% among males and 28% among females. Prevalence of CS with last partner among NVLS males decreased in Zambia from 48% (43-54) to 14% (11-18) and from 44% (40-47) to 15% (6-25) among NVLS females. CS among males and females with NVLS who were aware of their HIV status was similar over time, 70% (57-81) in 2016 vs 79% (62-96) in 2021 and 73% (66-80) in 2016 vs 67% (19-100) in 2021. In Botswana in 2021, prevalence of IP was 17% among males and 20% among females, similar age pattern as Zambia. Prevalence of CS within NVLS males and females was 10%. Among males and females aware of their status with NVLS CS was 9% (0-19) and 30% (12-49). Prevalence of IP among PWH was higher than among HIV-negative people in both countries regardless of sex.
Conclusions
Sexual partner profiles change within and across age/sex specific groups over time which can influence patterns of HIV acquisition risk. Differences in progress towards epidemic control in Zambia and Botswana appear to be reflected in the prevalence of IP and CS among people with a viral load high enough to transmit to their partners. Reaching status aware NVLS PWH is important for epidemic control.
Table or Figure
