Historical visit attendance as predictor of treatment interruption in South African HIV patients: Extension of a validated machine learning model.

Retention of antiretroviral (ART) patients is a priority for achieving HIV epidemic control in South Africa. While machine-learning methods are being increasingly utilised to identify high risk populations for suboptimal HIV service utilisation, they are limited in terms of explaining relationships between predictors. To further understand these relationships, we implemented machine learning methods optimised for predictive power and traditional statistical methods. We used routinely collected electronic medical record (EMR) data to evaluate longitudinal predictors of lost-to-follow up (LTFU) and temporal interruptions in treatment (IIT) in the first two years of treatment for ART patients in the Gauteng and North West provinces of South Africa. Of the 191,162 ART patients and 1,833,248 visits analysed, 49% experienced at least one IIT and 85% of those returned for a subsequent clinical visit. Patients iteratively transition in and out of treatment indicating that ART retention in South Africa is likely underestimated. Historical visit attendance is shown to be predictive of IIT using machine learning, log binomial regression and survival analyses. Using a previously developed categorical boosting (CatBoost) algorithm, we demonstrate that historical visit attendance alone is able to predict almost half of next missed visits. With the addition of baseline demographic and clinical features, this model is able to predict up to 60% of next missed ART visits with a sensitivity of 61.9% (95% CI: 61.5-62.3%), specificity of 66.5% (95% CI: 66.4-66.7%), and positive predictive value of 19.7% (95% CI: 19.5-19.9%). While the full usage of this model is relevant for settings where infrastructure exists to extract EMR data and run computations in real-time, historical visits attendance alone can be used to identify those at risk of disengaging from HIV care in the absence of other behavioural or observable risk factors.

Spatio-temporal estimates of HIV risk group proportions for adolescent girls and young women across 13 priority countries in sub-Saharan Africa.

The Global AIDS Strategy 2021-2026 identifies adolescent girls and young women (AGYW) as a priority population for HIV prevention, and recommends differentiating intervention portfolios geographically based on local HIV incidence and individual risk behaviours. We estimated prevalence of HIV risk behaviours and associated HIV incidence at health district level among AGYW living in 13 countries in sub-Saharan Africa. We analysed 46 geospatially-referenced national household surveys conducted between 1999-2018 across 13 high HIV burden countries in sub-Saharan Africa. Female survey respondents aged 15-29 years were classified into four risk groups (not sexually active, cohabiting, non-regular or multiple partner[s] and female sex workers [FSW]) based on reported sexual behaviour. We used a Bayesian spatio-temporal multinomial regression model to estimate the proportion of AGYW in each risk group stratified by district, year, and five-year age group. Using subnational estimates of HIV prevalence and incidence produced by countries with support from UNAIDS, we estimated new HIV infections in each risk group by district and age group. We then assessed the efficiency of prioritising interventions according to risk group. Data consisted of 274,970 female survey respondents aged 15-29. Among women aged 20-29, cohabiting (63.1%) was more common in eastern Africa than non-regular or multiple partner(s) (21.3%), while in southern countries non-regular or multiple partner(s) (58.9%) were more common than cohabiting (23.4%). Risk group proportions varied substantially across age groups (65.9% of total variation explained), countries (20.9%), and between districts within each country (11.3%), but changed little over time (0.9%). Prioritisation based on behavioural risk, in combination with location- and age-based prioritisation, reduced the proportion of population required to be reached in order to find half of all expected new infections from 19.4% to 10.6%. FSW were 1.3% of the population but 10.6% of all expected new infections. Our risk group estimates provide data for HIV programmes to set targets and implement differentiated prevention strategies outlined in the Global AIDS Strategy. Successfully implementing this approach would result in more efficiently reaching substantially more of those at risk for infections.

Modelling the impact of screening for chlamydia and gonorrhoea in youth and other high-prevalence groups in a resource-limited setting.

OBJECTIVES: Modelling the potential impact of screening for chlamydia and gonorrhoea in youth and other populations in a resource-limited setting. METHODS: We extended an agent-based model of heterosexual chlamydia and gonorrhoea transmission in South Africa to investigate the impact of screening strategies in key populations including youth, patients in HIV care, pregnant women and female sex workers (FSWs). Additionally, we compared the modelled impact of a standardised screening programme to results obtained from other published mathematical models of chlamydia screening. RESULTS: All strategies resulted in reductions in general and targeted population chlamydia and gonorrhoea transmission. Opportunistic screening of patients in youth and HIV care was shown to be the most effective, and FSW screening was shown to be the most efficient strategy. Differences between models could be attributed to differences in the modelled heterogeneity in sexual behaviour as well as differences in assumptions about immunity following chlamydia recovery. CONCLUSIONS: Taking modelling assumptions into account, opportunistic chlamydia and gonorrhoea screening of youth and those in HIV care represents a viable intervention for reducing sexually transmitted infections in the South African population.

Does HIV Exploit the Inflammatory Milieu of the Male Genital Tract for Successful Infection?

In many parts of the World, medical male circumcision (MMC) is used as standard prevention of care against HIV infection. This is based on seminal reports made over 10 years ago that removal of the foreskin provides up to 60% protection against HIV infection in males and seems currently the best antiretroviral-free prevention strategy yet against the global epidemic. We explore the potential mechanisms by which MMC protects against HIV-1 acquisition and that one of the oldest, albeit re-invented, rituals of removing a foreskin underscores the exploitative nature of HIV on the anatomy and tissue of the uncircumcised penis. Furthermore, foreskin removal also reveals how males acquire HIV, and in reality, the underlying mechanisms of MMC are not known. We argue that the normal sequelae of inflammation in the male genital tract (MGT) for protection from sexually transmitted infections (STI)-induced pathology represents a perfect immune and microbial ecosystem for HIV acquisition. The accumulation of HIV-1 target cells in foreskin tissue and within the urethra in response to STIs, both during and after resolution of infection, suggests that acquisition of HIV-1, through sexual contact, makes use of the natural immune milieu of the MGT. Understanding immunity in the MGT, the movement of HIV-1 target cells to the urethra and foreskin tissue upon encounter with microbial signals would provide more insight into viral acquisition and lay the foundation for further prevention strategies in males that would be critical to curb the epidemic in all sexual partners at risk of infection. The global female-centric focus of HIV-1 transmission and acquisition research has tended to leave gaps in our knowledge of what determines HIV-1 acquisition in men and such understanding would provide a more balanced and complete view of viral acquisition.