Improved Indicators for Subnational Unmet Antiretroviral Therapy Need in the Health System: Updates to the Naomi Model in 2023

BACKGROUND

Quantifying subnational need for antiretroviral therapy (ART) for HIV is challenging because people living with HIV (PLHIV) access health facilities in areas that may differ from their residence. We defined and demonstrated new indicators for PLHIV treatment needed to guide health system target setting and resource allocation.

SETTING

Botswana.

METHODS

We extended Naomi, a Bayesian small-area model for estimating district-level HIV indicators from national household survey and HIV service delivery data. We used model outputs for ART seeking probabilities in neighboring districts to define the "PLHIV (attending)" indicator representing the estimated number of PLHIV who would seek treatment at health facilities in a district, and "Untreated PLHIV attending" representing gaps in ART service provision. Botswana 2021 district HIV estimates were used to demonstrate new outputs and assess the sensitivity to uncertainty in district population sizes.

RESULTS

Across districts of Botswana, estimated adult ART coverage in December 2021 ranged 90%-96%. In the capital city Gaborone, there were 50,400 resident PLHIV and 64,200 receiving ART, of whom 24% (95% CI: 20 to 32) were estimated to reside in neighboring districts. Applying ART attendance probabilities gave a "PLHIV attending" denominator of 68,300 and unmet treatment need of 4100 adults (95% CI: 3000 to 5500) for Gaborone health facilities. The facility-based "PLHIV attending" denominator was less-sensitive to fluctuations in district population size assumptions.

CONCLUSIONS

New indicators provided more consistent targets for HIV service provision, but are limited by ART data quality. This challenge will increase as treatment coverage reaches high levels and treatment gaps are smaller.

Prediction of HIV status based on socio-behavioural characteristics in East and Southern Africa

Introduction

High yield HIV testing strategies are critical to reach epidemic control in high prevalence and low-resource settings such as East and Southern Africa. In this study, we aimed to predict the HIV status of individuals living in Angola, Burundi, Ethiopia, Lesotho, Malawi, Mozambique, Namibia, Rwanda, Zambia and Zimbabwe with the highest precision and sensitivity for different policy targets and constraints based on a minimal set of socio-behavioural characteristics.

Methods

We analysed the most recent Demographic and Health Survey from these 10 countries to predict individual’s HIV status using four different algorithms (a penalized logistic regression, a generalized additive model, a support vector machine, and a gradient boosting trees). The algorithms were trained and validated on 80% of the data, and tested on the remaining 20%. We compared the predictions based on the F1 score, the harmonic mean of sensitivity and positive predictive value (PPV), and we assessed the generalization of our models by testing them against an independent left-out country. The best performing algorithm was trained on a minimal subset of variables which were identified as the most predictive, and used to 1) identify 95% of people living with HIV (PLHIV) while maximising precision and 2) identify groups of individuals by adjusting the probability threshold of being HIV positive (90% in our scenario) for achieving specific testing strategies.

Results

Overall 55,151 males and 69,626 females were included in the analysis. The gradient boosting trees algorithm performed best in predicting HIV status with a mean F1 score of 76.8% [95% confidence interval (CI) 76.0%-77.6%] for males (vs [CI 67.8%-70.6%] for SVM) and 78.8% [CI 78.2%-79.4%] for females (vs [CI 73.4%-75.8%] for SVM). Among the ten most predictive variables for each sex, nine were identical: longitude, latitude and, altitude of place of residence, current age, age of most recent partner, total lifetime number of sexual partners, years lived in current place of residence, condom use during last intercourse and, wealth index. Only age at first sex for male (ranked 10th) and Rohrer’s index for female (ranked 6th) were not similar for both sexes. Our large-scale scenario, which consisted in identifying 95% of all PLHIV, would have required testing 49.4% of males and 48.1% of females while achieving a precision of 15.4% for males and 22.7% for females. For the second scenario, only 4.6% of males and 6.0% of females would have had to be tested to find 55.7% of all males and 50.5% of all females living with HIV.

Conclusions

We trained a gradient boosting trees algorithm to find 95% of PLHIV with a precision twice higher than with general population testing by using only a limited number of socio-behavioural characteristics. We also successfully identified people at high risk of infection who may be offered pre-exposure prophylaxis or voluntary medical male circumcision. These findings can inform the implementation of new high-yield HIV tests and help develop very precise strategies based on low-resource settings constraints.

Naomi: a new modelling tool for estimating HIV epidemic indicators at the district level in sub-Saharan Africa

INTRODUCTION

HIV planning requires granular estimates for the number of people living with HIV (PLHIV), antiretroviral treatment (ART) coverage and unmet need, and new HIV infections by district, or equivalent subnational administrative level. We developed a Bayesian small-area estimation model, called Naomi, to estimate these quantities stratified by subnational administrative units, sex, and five-year age groups.

METHODS

Small-area regressions for HIV prevalence, ART coverage and HIV incidence were jointly calibrated using subnational household survey data on all three indicators, routine antenatal service delivery data on HIV prevalence and ART coverage among pregnant women, and service delivery data on the number of PLHIV receiving ART. Incidence was modelled by district-level HIV prevalence and ART coverage. Model outputs of counts and rates for each indicator were aggregated to multiple geographic and demographic stratifications of interest. The model was estimated in an empirical Bayes framework, furnishing probabilistic uncertainty ranges for all output indicators. Example results were presented using data from Malawi during 2016-2018.

RESULTS

Adult HIV prevalence in September 2018 ranged from 3.2% to 17.1% across Malawi's districts and was higher in southern districts and in metropolitan areas. ART coverage was more homogenous, ranging from 75% to 82%. The largest number of PLHIV was among ages 35 to 39 for both women and men, while the most untreated PLHIV were among ages 25 to 29 for women and 30 to 34 for men. Relative uncertainty was larger for the untreated PLHIV than the number on ART or total PLHIV. Among clients receiving ART at facilities in Lilongwe city, an estimated 71% (95% CI, 61% to 79%) resided in Lilongwe city, 20% (14% to 27%) in Lilongwe district outside the metropolis, and 9% (6% to 12%) in neighbouring Dowa district. Thirty-eight percent (26% to 50%) of Lilongwe rural residents and 39% (27% to 50%) of Dowa residents received treatment at facilities in Lilongwe city.

CONCLUSIONS

The Naomi model synthesizes multiple subnational data sources to furnish estimates of key indicators for HIV programme planning, resource allocation, and target setting. Further model development to meet evolving HIV policy priorities and programme need should be accompanied by continued strengthening and understanding of routine health system data.

Hormonal contraception alters vaginal microbiota and cytokines in South African adolescents in a randomized trial

Young women in sub-Saharan Africa are disproportionally affected by HIV infection and unintended pregnancies. However, hormonal contraceptive (HC) use may influence HIV risk through changes in genital tract microbiota and inflammatory cytokines. To investigate this, 130 HIV negative adolescent females aged 15-19 years were enrolled into a substudy of UChoose, an open-label randomized crossover study (NCT02404038), comparing acceptability and contraceptive product preference as a proxy for HIV prevention delivery methods. Participants were randomized to injectable norethisterone enanthate (Net-En), combined oral contraceptives (COC) or etonorgesterol/ethinyl estradiol combined contraceptive vaginal ring (CCVR) for 16 weeks, then crossed over to another HC for 16 weeks. Cervicovaginal samples were collected at baseline, crossover and exit for characterization of the microbiota and measurement of cytokine levels; primary endpoints were cervical T cell activation, vaginal microbial diversity and cytokine concentrations. Adolescents randomized to COCs had lower vaginal microbial diversity and relative abundance of HIV risk-associated taxa compared to Net-En or CCVR. Cervicovaginal inflammatory cytokine concentrations were significantly higher in adolescents randomized to CCVR compared to COC and Net-En. This suggests that COC use may induce an optimal vaginal ecosystem by decreasing bacterial diversity and inflammatory taxa, while CCVR use is associated with genital inflammation.

Hormonal contraception and risk of STIs and bacterial vaginosis in South African adolescents: secondary analysis of a randomised trial

OBJECTIVES

Young women in sub-Saharan Africa are at high risk of STIs and unintended pregnancies, yet hormonal contraceptive (HC) use may affect STI risk. We compared the influence of three HCs on the incidence and prevalence of STIs and bacterial vaginosis (BV) in South African adolescents.

METHODS

One hundred and thirty adolescents between 15 and 19 years were randomised to the injectable norethisterone enanthate (Net-En), combined oral contraceptives (COC) (Triphasil or Nordette) or a combined contraceptive vaginal ring (CCVR; NuvaRing) for 16 weeks (clinicaltrials.gov/NCT02404038). Vaginal samples were collected at baseline and 16 weeks post contraceptive initiation for STI and BV testing.

RESULTS

In an intention-to-treat analysis, no significant differences in BV prevalence were found between study arms. The overall incidence of any STI at follow-up was high: 16.2% in the COC arm; 25.7% in the Net-En arm; and 37.1% in the CCVR arm. The incidence rate (IR) of any STI was similar between Net-En (IR 0.74 (95% CI 0.34 to 1.41)) and the oestrogen-containing contraceptives (IR 0.78 (95% CI 0.47 to 1.22)). A lower IR of Chlamydia trachomatis (incidence rate ratio (IRR) 0.68 (95% CI 0.19 to 1.99)) and Neisseria gonorrhoeae (IRR 0.25 (95% CI 0.01 to 1.35)) but a higher IR of Mycoplasma genitalium (IRR 16.0 (95% CI 2.96 to 400)), was observed in the Net-En arm compared with the oestrogen-containing contraceptives, although the overall incidence of M. genitalium was low (4.7%). In an exploratory analysis, the risk of any STI and N. gonorrhoeae was lower in the COC arm compared with CCVR. A per-protocol analysis yielded similar results.

CONCLUSION

Our results suggest that use of Net-En may be associated with increased risk of M. genitalium compared with oestrogen-containing contraceptives but not with overall STI risk. COC use may decrease STI risk relative to CCVR.