Participation dynamics in population-based longitudinal HIV surveillance in rural South Africa

Tuberculin skin test surveys and the Annual Risk of Tuberculous Infection in school children in Northern KwaZulu-Natal

Tuberculin skin test surveys in primary school children can be used to quantify Mycobacterium tuberculosis transmission at community level. KwaZulu-Natal province, South Africa, is home to 11.5 million people and suffers a burden of tuberculosis disease that is among the highest in the world. The last tuberculin survey in the province was undertaken in 1979. We performed a tuberculin skin test survey nested within a demographic and health household surveillance programme in Northern KwaZulu-Natal. We enrolled children aged between six and eight years of age attending primary schools in this community. Mixture analysis was used to determine tuberculin skin test thresholds and the Annual Risk of Tuberculous Infection derived from age at testing and infection prevalence. The Community Infection Ratio, a measure of the relative importance of within-household and community transmission, was calculated from data on tuberculin positivity disaggregated by household tuberculosis contact. Between June and December 2013, we obtained tuberculin skin test results on 1240 children. Mixture analysis proved unstable, suggesting two potential thresholds for test positivity. Using a threshold of ≥10mm or treating all non zero reactions as positive yielded estimates of the Annual Risk of Tuberculous Infection of 1.7% (1.4-2.1%) or 2.4% (2.0-3.0%). Using the same thresholds and including children reported to be receiving TB treatment as cases, resulted in estimates of 2.0% (1.6-2.5%) or 2.7% (2.2-3.3%). The Community Infection Ratio was 0.58 (0.33-1.01). The force of infection in this community is lower than that observed in Western Cape province, South Africa, but higher than that observed in community settings in most other parts of the world. Children in this community are commonly infected with Mycobacterium tuberculosis outside the home. Interventions to interrupt transmission are urgently needed.

Population impacts of conditional financial incentives and a male-targeted digital decision support application on the HIV treatment cascade in rural KwaZulu Natal: findings from the HITS cluster randomized clinical trial

INTRODUCTION

In South Africa, the HIV care cascade remains suboptimal. We investigated the impact of small conditional financial incentives (CFIs) and male-targeted HIV-specific decision-support application (EPIC-HIV) on the HIV care cascade.

METHODS

In 2018, in uMkhanyakude district, 45 communities were randomly assigned to one of four arms: (i) CFI for home-based HIV testing and linkage to care within 6 weeks (R50 [US$3] food voucher each); (ii) EPIC-HIV which are based on self-determination theory; (iii) both CFI and EPIC-HIV; and (iv) standard of care. EPIC-HIV consisted of two components: EPIC-HIV 1, provided to men through a tablet before home-based HIV testing, and EPIC-HIV 2, offered 1 month later to men who tested positive but had not yet linked to care. Linking HITS trial data to national antiretroviral treatment (ART) programme data and HIV surveillance programme data, we estimated HIV status awareness after the HITS trial implementation, ART status 3 month after the trial and viral load suppression 1 year later. Analysis included all known individuals living with HIV in the study area including those who did not participated in the HITS trial.

RESULTS

Among the 33,778 residents in the study area, 2763 men and 7266 women were identified as living with HIV by the end of the intervention period and included in the analysis. After the intervention, awareness of HIV-positive status was higher in the CFI arms compared to non-CFI arms (men: 793/908 [87.3%] vs. 1574/1855 [84.9%], RR = 1.03 [95% CI: 0.99-1.07]; women: 2259/2421 [93.3%] vs. 4439/4845 [91.6%], RR = 1.02 [95% CI: 1.00-1.04]). Three months after the intervention, no differences were found for linkage to ART between arms. One year after the intervention, only 1829 viral test results were retrieved. Viral suppression was higher but not significant in the EPIC-HIV intervention arms among men (65/99 [65.7%] vs. 182/308 [59.1%], RR = 1.11 [95% CI: 0.88-1.40]).

CONCLUSIONS

Small CFIs can contribute to achieve the first step of the HIV care cascade. However, neither CFIs nor EPIC-HIV was sufficient to increase the number of people on ART. Additional evidence is needed to confirm the impact of EPIC-HIV on viral suppression.

Mortality in rural coastal Kenya measured using the Kilifi Health and Demographic Surveillance System: a 16-year descriptive analysis

Background: The Kilifi Health and Demographic Surveillance System (KHDSS) was established in 2000 to define the incidence and prevalence of local diseases and evaluate the impact of community-based interventions. KHDSS morbidity data have been reported comprehensively but mortality has not been described. This analysis describes mortality in the KHDSS over 16 years. Methods: We calculated mortality rates from 2003-2018 in four intervals of equal duration and assessed differences in mortality across these intervals by age and sex. We calculated the period survival function and median survival using the Kaplan-Meier method and mean life expectancies using abridged life tables. We estimated trend and seasonality by decomposing a time series of monthly mortality rates. We used choropleth maps and random-effects Poisson regression to investigate geographical heterogeneity. Results: Mortality declined by 36% overall between 2003-2018 and by 59% in children aged <5 years. Most of the decline occurred between 2003 and 2006. Among adults, the greatest decline (49%) was observed in those aged 15-54 years. Life expectancy at birth increased by 12 years. Females outlived males by 6 years. Seasonality was only evident in the 1-4 year age group in the first four years. Geographical variation in mortality was ±10% of the median value and did not change over time. Conclusions: Between 2003 and 2018, mortality among children and young adults has improved substantially. The steep decline in 2003-2006 followed by a much slower reduction thereafter suggests improvements in health and wellbeing have plateaued in the last 12 years. However, there is substantial inequality in mortality experience by geographical location.

Predictors of migration in an HIV hyper-endemic rural South African community: evidence from a population-based cohort (2005-2017)

Globally, South Africa hosts the highest number of people living with HIV (PLHIV) and the unique legacy of internal labour migration continues to be a major driver of the regional epidemic, interrupting treatment-as-prevention efforts. The study examined levels, trends, and predictors of migration in rural KwaZulu-Natal Province, South Africa, using population-based surveillance data from 2005 through 2017. We followed 69 604 adult participants aged 15-49 years and recorded their migration events (i.e., out-migration from the surveillance area) in 423 038 person-years over 525 397 observations. Multiple failure Cox-regression models were used to measure the risk of migration by socio-demographic factors: age, sex, educational status, marital status, HIV, and community antiretroviral therapy (ART) coverage. Overall, 69% of the population cohort experienced at least one migration event during the follow-up period. The average incidence rate of migration was 9.96 events and 13.23 events per 100 person-years in women and men, respectively. Migration rates declined from 2005 to 2008 then peaked in 2012 for both women and men. Adjusting for other covariates, the risk of migration was 3.4-times higher among young women aged 20-24 years compared to those aged ≥ 40 years (adjusted Hazard Ratio [aHR] = 3.37, 95% Confidence Interval [CI]: 3:19-3.57), and 2.9-times higher among young men aged 20-24 years compared to those aged ≥ 40 years (aHR = 2.86, 95% CI:2.69-3.04). There was a 9% and 27% decrease in risk of migration among both women (aHR = 0.91, 95% CI: 0.83 - 0.99) and men (aHR = 0.73, 95% CI 0.66 - 0.82) respectively per every 1% increase in community ART coverage. Young unmarried women including those living with HIV, migrated at a magnitude similar to that of their male counterparts, and lowered as ART coverage increased over time, reflecting the role of improved HIV services across space in reducing out-migration. A deeper understanding of the characteristics of a migrating population provides critical information towards identifying and addressing gaps in the HIV prevention and care continuum in an era of high mobility.

Evaluating the impact of DREAMS on HIV incidence among adolescent girls and young women: A population-based cohort study in Kenya and South Africa

BACKGROUND

Through a multisectoral approach, the DREAMS Partnership aimed to reduce HIV incidence among adolescent girls and young women (AGYW) by 40% over 2 years in high-burden districts across sub-Saharan Africa. DREAMS promotes a combination package of evidence-based interventions to reduce individual, family, partner, and community-based drivers of young women's heightened HIV risk. We evaluated the impact of DREAMS on HIV incidence among AGYW and young men in 2 settings.

METHODS AND FINDINGS

We directly estimated HIV incidence rates among open population-based cohorts participating in demographic and HIV serological surveys from 2006 to 2018 annually in uMkhanyakude (KwaZulu-Natal, South Africa) and over 6 rounds from 2010 to 2019 in Gem (Siaya, Kenya). We compared HIV incidence among AGYW aged 15 to 24 years before DREAMS and up to 3 years after DREAMS implementation began in 2016. We investigated the timing of any change in HIV incidence and whether the rate of any change accelerated during DREAMS implementation. Comparable analyses were also conducted for young men (20 to 29/34 years). In uMkhanyakude, between 5,000 and 6,000 AGYW were eligible for the serological survey each year, an average of 85% were contacted, and consent rates varied from 37% to 67%. During 26,395 person-years (py), HIV incidence was lower during DREAMS implementation (2016 to 2018) than in the previous 5-year period among 15- to 19-year-old females (4.5 new infections per 100 py as compared with 2.8; age-adjusted rate ratio (aRR) = 0.62, 95% confidence interval [CI] 0.48 to 0.82), and lower among 20- to 24-year-olds (7.1/100 py as compared with 5.8; aRR = 0.82, 95% CI 0.65 to 1.04). Declines preceded DREAMS introduction, beginning from 2012 to 2013 among the younger and 2014 for the older women, with no evidence of more rapid decline during DREAMS implementation. In Gem, between 8,515 and 11,428 AGYW were eligible each survey round, an average of 34% were contacted and offered an HIV test, and consent rates ranged from 84% to 99%. During 10,382 py, declines in HIV incidence among 15- to 19-year-olds began before DREAMS and did not change after DREAMS introduction. Among 20- to 24-year-olds in Gem, HIV incidence estimates were lower during DREAMS implementation (0.64/100 py) compared with the pre-DREAMS period (0.94/100 py), with no statistical evidence of a decline (aRR = 0.69, 95% CI 0.53 to 2.18). Among young men, declines in HIV incidence were greater than those observed among AGYW and also began prior to DREAMS investments. Study limitations include low study power in Kenya and the introduction of other interventions such as universal treatment for HIV during the study period.

CONCLUSIONS

Substantial declines in HIV incidence among AGYW were observed, but most began before DREAMS introduction and did not accelerate in the first 3 years of DREAMS implementation. Like the declines observed among young men, they are likely driven by earlier and ongoing investments in HIV testing and treatment. Longer-term implementation and evaluation are needed to assess the impact of such a complex HIV prevention intervention and to help accelerate reductions in HIV incidence among young women.

It ain't what you do, it's the way that you do it: The pitfalls of using routine data to measure early infant HIV diagnosis in HIV-exposed infants

BACKGROUND

Early infant HIV diagnosis (EID) is critical to ensuring timely diagnosis of HIV-exposed infants, and treatment in those found to be infected. However estimates of coverage vary considerably, depending on data sources used. We used 4 methods to estimate coverage among a historical cohort of HIV-exposed infants in rural South Africa, between 2010-2016.

METHODS

We estimated the proportion of infants ever tested (methods 1-3) and tested by 7 weeks of age (1-4) as follows: (1) infants born to women identified as HIV-positive in demographic surveillance were linked to those with ≥1 EID result in routine laboratory surveillance; (2) the number of infants with ≥1 EID result in laboratory surveillance divided by the estimated number of HIV-exposed infants, calculated as total live births multiplied by antenatal HIV seroprevalence; (3) the number of infants with ≥1 EID result in routine laboratory surveillance, divided by the number of HIV-exposed infants as estimated by the district health service; (4) from documentation in infants' Road-to-Health-booklets.

RESULTS

The proportion ever tested was 43%, 88% and 138% for methods 1-3, and by 7 weeks of age was 25%, 49%, 86% and 46% for methods 1-4 respectively.

CONCLUSIONS

The four methods, applied to a range of routine data sources, resulted in estimates varying considerably, and the true coverage of EID remains unclear. Our findings highlight the importance of developing unique patient identifiers, improving training of healthcare providers using reporting systems, and ensuring the accuracy of healthcare records, to ensure the best possible health outcomes for HIV-exposed infants.

Association of medical male circumcision and sexually transmitted infections in a population-based study using targeted maximum likelihood estimation

Background

Epidemiological theory and many empirical studies support the hypothesis that there is a protective effect of male circumcision against some sexually transmitted infections (STIs). However, there is a paucity of randomized control trials (RCTs) to test this hypothesis in the South African population. Due to the infeasibility of conducting RCTs, estimating marginal or average treatment effects with observational data increases interest. Using targeted maximum likelihood estimation (TMLE), a doubly robust estimation technique, we aim to provide evidence of an association between medical male circumcision (MMC) and two STI outcomes.

Methods

HIV and HSV-2 status were the two primary outcomes for this study. We investigated the associations between MMC and these STI outcomes, using cross-sectional data from the HIV Incidence Provincial Surveillance System (HIPSS) study in KwaZulu-Natal, South Africa. HIV antibodies were tested from the blood samples collected in the study. For HSV-2, serum samples were tested for HSV-2 antibodies via an ELISA-based anti-HSV-2 IgG. We estimated marginal prevalence ratios (PR) using TMLE and compared estimates with those from propensity score full matching (PSFM) and inverse probability of treatment weighting (IPTW).

Results

From a total 2850 male participants included in the analytic sample, the overall weighted prevalence of HIV was 32.4% (n = 941) and HSV-2 was 53.2% (n = 1529). TMLE estimates suggest that MMC was associated with 31% lower HIV prevalence (PR: 0.690; 95% CI: 0.614, 0.777) and 21.1% lower HSV-2 prevalence (PR: 0.789; 95% CI: 0.734, 0.848). The propensity score analyses also provided evidence of association of MMC with lower prevalence of HIV and HSV-2. For PSFM: HIV (PR: 0.689; 95% CI: 0.537, 0.885), and HSV-2 (PR: 0.832; 95% CI: 0.709, 0.975). For IPTW: HIV (PR: 0.708; 95% CI: 0.572, 0.875), and HSV-2 (PR: 0.837; 95% CI: 0.738, 0.949).

Conclusion

Using a TMLE approach, we present further evidence of a protective association of MMC against HIV and HSV-2 in this hyper-endemic South African setting. TMLE has the potential to enhance the evidence base for recommendations that embrace the effect of public health interventions on health or disease outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-021-11705-9.

Is Interview Length Associated With Blood Test Participation? Evidence From Three Population-Based HIV Impact Assessment Surveys Conducted From 2016 to 2017

BACKGROUND

High response rates in surveys are critical to ensuring that findings are unbiased and representative of the target population. Questionnaire length affects response rates, with long interviews associated with partially complete surveys, higher item nonresponse ("don't know" and "refuse" responses), and willingness to participate in future surveys. Our aim is to determine the impact of questionnaire length on blood test participation in population-based HIV surveys.

METHODS

Data are from population-based HIV impact assessments conducted in Zambia, Eswatini, and Lesotho in 2016-2017. The population-based HIV impact assessments consist of an interview followed by a blood draw. Consent for blood draw was obtained before the interview in Eswatini and after the interview in Zambia and Lesotho. Interview length was measured by the survey tablet as the time to complete the survey (interview duration) and the number of questions answered by the participant (questionnaire length). We assessed the effects of questionnaire length and interview duration on blood test participation using logistic regression.

RESULTS

Across all 3 surveys, the median interview duration was 16 minutes and the median number of questions was 77. In adjusted analyses, there was a negative impact of interview duration on blood draw consent for individuals with unknown status in Lesotho and a positive relationship between questionnaire length and blood draw consent in Zambia for those with HIV-negative and unknown status.

CONCLUSION

Although interview length is an important consideration to reduce respondent burden, a longer questionnaire does not necessarily result in lower consent rates for blood testing.

Age patterns of HIV incidence in eastern and southern Africa: a modelling analysis of observational population-based cohort studies

BACKGROUND

As the HIV epidemic in sub-Saharan Africa matures, evidence about the age distribution of new HIV infections and how this distribution has changed over the epidemic is needed to guide HIV prevention. We aimed to assess trends in age-specific HIV incidence in six population-based cohort studies in eastern and southern Africa, reporting changes in mean age at infection, age distribution of new infections, and birth cohort cumulative incidence.

METHODS

We used a Bayesian model to reconstruct age-specific HIV incidence from repeated observations of individuals' HIV serostatus and survival collected among population HIV cohorts in rural Malawi, South Africa, Tanzania, Uganda, and Zimbabwe, in a collaborative analysis of the ALPHA network. We modelled HIV incidence rates by age, time, and sex using smoothing splines functions. We estimated incidence trends separately by sex and study. We used estimated incidence and prevalence results for 2000-17, standardised to study population distribution, to estimate mean age at infection and proportion of new infections by age. We also estimated cumulative incidence (lifetime risk of infection) by birth cohort.

FINDINGS

Age-specific incidence declined at all ages, although the timing and pattern of decline varied by study. The mean age at infection was higher in men (cohort mean 27·8-34·6 years) than in women (24·8-29·6 years). Between 2000 and 2017, the mean age at infection per cohort increased slightly: 0·5 to 2·8 years among men and -0·2 to 2·5 years among women. Across studies, between 38% and 63% (cohort medians) of the infections in women were among those aged 15-24 years and between 30% and 63% of infections in men were in those aged 20-29 years. Lifetime risk of HIV declined for successive birth cohorts.

INTERPRETATION

HIV incidence declined in all age groups and shifted slightly to older ages. Disproportionate new HIV infections occur among women aged 15-24 years and men aged 20-29 years, supporting focused prevention in these groups. However, 40-60% of infections were outside these ages, emphasising the importance of providing appropriate HIV prevention to adults of all ages.

FUNDING

Bill & Melinda Gates Foundation.

Home-Based Intervention to Test and Start (HITS): a community-randomized controlled trial to increase HIV testing uptake among men in rural South Africa

Introduction

The uptake of HIV testing and linkage to care remains low among men, contributing to high HIV incidence in women in South Africa. We conducted the “Home‐Based Intervention to Test and Start” (HITS) in a 2x2 factorial cluster randomized controlled trial in one of the World’s largest ongoing HIV cohorts in rural South Africa aimed at enhancing both intrinsic and extrinsic motivations for HIV testing.

Methods

Between February and December 2018, in the uMkhanyakude district of KwaZulu‐Natal, we randomly assigned 45 communities (clusters) (n = 13,838 residents) to one of the four arms: (i) financial incentives for home‐based HIV testing and linkage to care (R50 [$3] food voucher each); (ii) male‐targeted HIV‐specific decision support application, called EPIC‐HIV; (iii) both financial incentives and male‐targeted HIV‐specific decision support application and (iv) standard of care (SoC). EPIC‐HIV was developed to encourage and serve as an intrinsic motivator for HIV testing and linkage to care, and individually offered to men via a tablet device. Financial incentives were offered to both men and women. Here we report the effect of the interventions on uptake of home‐based HIV testing among men. Intention‐to‐treat (ITT) analysis was performed using modified Poisson regression with adjustment for clustering of standard errors at the cluster levels.

Results

Among all 13,838 men ≥ 15 years living in the 45 communities, the overall population coverage during a single round of home‐based HIV testing was 20.7%. The uptake of HIV testing was 27.5% (683/2481) in the financial incentives arm, 17.1% (433/2534) in the EPIC‐HIV arm, 26.8% (568/2120) in the arm receiving both interventions and 17.8% in the SoC arm. The probability of HIV testing increased substantially by 55% in the financial incentives arm (risk ratio (RR)=1.55, 95% CI: 1.31 to 1.82, p < 0.001) and 51% in the arm receiving both interventions (RR = 1.51, 95% CI: 1.21 to 1.87 p < 0.001), compared to men in the SoC arm. The probability of HIV testing did not significantly differ in the EPIC‐HIV arm (RR = 0.96, 95% CI: 0.76 to 1.20, p = 0.70).

Conclusions

The provision of a small financial incentive acted as a powerful extrinsic motivator substantially increasing the uptake of home‐based HIV testing among men in rural South Africa. In contrast, the counselling and testing application which was designed to encourage and serve as an intrinsic motivator to test for HIV did not increase the uptake of home‐based testing.

BMI and All-Cause Mortality in a Population-Based Cohort in Rural South Africa

OBJECTIVE

This study evaluates the association between BMI and all-cause and cause-specific mortality in South Africa.

METHODS

Prospective, population-based observational cohort data from rural South Africa were analyzed. BMI was measured in 2010. Demographic characteristics were recorded and deaths were verified with verbal autopsy interview. The InterVA-5 tool was used to assign causes of death. HIV testing was conducted annually. Cox proportional hazards models were fit to estimate the effect of BMI on all-cause and cause-specific mortality, accounting for the competing risk of death from other causes. Models were adjusted for sociodemographic characteristics and HIV status, and inverse probability weighting for survey nonparticipation was used.

RESULTS

The cohort consisted of 9,728 individuals. In adjusted models, those with BMI of 25.0 to 29.9 kg/m² or 30.0 to 34.9 kg/m² had a lower hazard of death (adjusted hazard ratio: 0.80; 95% CI: 0.69-0.92 and adjusted hazard ratio: 0.75; 95% CI: 0.60-0.93, respectively) compared with those with BMI of 18.5 to 24.9 kg/m² .

CONCLUSIONS

Individuals in South Africa who meet clinically defined criteria for overweight or obesity had a lower risk of all-cause mortality than those with a normal BMI. These findings were stronger for women and communicable conditions.

HIV incidence declines in a rural South African population: a G-imputation approach for inference

BACKGROUND

Ad hoc assumptions about the unobserved infection event, which is known only to occur between the latest-negative and earliest-positive test dates, can lead to biased HIV incidence rate estimates. Using a G-imputation approach, we infer the infection dates from covariate data to estimate the HIV incidence rate in a hyper-endemic South African setting.

METHODS

A large demographic surveillance system has annually tested a cohort of HIV-uninfected participants living in the KwaZulu-Natal province. Using this data, we estimated a cumulative baseline hazard function and the effects of time-dependent covariates on the interval censored infection dates. For each HIV-positive participant in the cohort, we derived a cumulative distribution function and sampled multiple infection dates conditional on the unique covariate values. We right censored the data at the imputed dates, calculated the annual HIV incidence rate per 100 person-years, and used Rubin's rules to obtain the 95% confidence intervals.

RESULTS

A total of 20,011 uninfected individuals with a repeat HIV test participated in the incidence cohort between 2005 and 2018. We observed 2,603 infections per 58,769 person-years of follow-up among women and 845 infections per 41,178 person-years of follow-up among men. Conditional on age and circumcision status (men only), the female HIV incidence rate declined by 25%, from 5.0 to 3.7 infections per 100 person-years between 2014 and 2018. During this period, the HIV incidence rate among men declined from 2.1 to 1.1 infections per 100 person-years-a reduction of 49%. We observed similar reductions in male and female HIV incidence conditional on condom-use, marital status, urban residential status, migration history, and the HIV prevalence in the surrounding community.

CONCLUSION

We have followed participants in one of the world's largest and longest running HIV cohorts to estimate long-term trends in the population-wide incidence of infection. Using a G-imputation approach, we present further evidence for HIV incidence rate declines in this hyper-endemic South African setting.

The state of the HIV epidemic in rural KwaZulu-Natal, South Africa: a novel application of disease metrics to assess trajectories and highlight areas for intervention

BACKGROUND

South Africa is at the epicentre of the HIV pandemic, with the world's highest number of new infections and the largest treatment programme. Using metrics proposed by the Joint United Nations Programme on AIDS (UNAIDS), we evaluate progress toward epidemic control and highlight areas for intervention in a hyperendemic South African setting.

METHODS

The Africa Health Research Institute (AHRI) maintains a comprehensive population-based surveillance system in the Hlabisa sub-district of KwaZulu-Natal. Between 2005 and 2017, we tested 39 735 participants (aged 15-49 years) for HIV and followed 22 758 HIV-negative and 13 460 HIV-positive participants to identify new infections and all-cause AIDS-related deaths, respectively. Using these data, we estimated the percentage reduction in incidence, the absolute incidence rate, the incidence-mortality ratio and the incidence-prevalence ratio over place and time.

RESULTS

We observed a 62% reduction in the number of new infections among men between 2012 and 2017 and a 34% reduction among women between 2014 and 2017. Among men, the incidence-mortality ratio peaked at 4.1 in 2013 and declined to 3.1 in 2017, and among women it fell from a high of 6.4 in 2014 to 4.3 in 2017. Between 2012 and 2017, the female-incidence/male-prevalence ratio declined from 0.24 to 0.13 and the male-incidence/female-prevalence ratio from 0.05 to 0.02.

CONCLUSIONS

Using data from a population-based cohort study, we report impressive progress toward HIV epidemic control in a severely affected South African setting. However, overall progress is off track for 2020 targets set by the UNAIDS. Spatial estimates of the metrics, which demonstrate remarkable heterogeneity over place and time, indicate areas that could benefit from additional or optimized HIV prevention services.

Recent levels and trends in HIV incidence rates among adolescent girls and young women in ten high-prevalence African countries: a systematic review and meta-analysis

BACKGROUND

The roll-out of antiretroviral therapy (ART) has changed contexts of HIV risk, but the influence on HIV incidence among young women is not clear. We aimed to summarise direct estimates of HIV incidence among adolescent girls and young women since ART and before large investments in targeted prevention for those in sub-Saharan Africa.

METHODS

We did a systematic review and meta-analysis. We searched MEDLINE, Embase, Web of Science, Global Health, and CINAHL for studies reporting HIV incidence data from serological samples collected among females aged 15-24 years in ten countries (Kenya, Lesotho, Malawi, Mozambique, South Africa, Swaziland, Tanzania, Uganda, Zambia, and Zimbabwe) that were selected for DREAMS investment in 2015. We only included articles published in English. Our main outcome was to summarise recent levels and trends in HIV incidence estimates collected between 2005 and 2015, published or received from study authors, by age and sex, and pooled by region.

FINDINGS

51 studies were identified from nine of the ten DREAMS countries; no eligible studies from Lesotho were identified. Directly observed HIV incidence rates were lowest among females aged 13-19 years in Kumi, Uganda (0·38 cases per 100 person-years); and directly observed HIV incidence rates were highest in KwaZulu-Natal, South Africa (7·79 per 100 person-years among females aged 15-19 years, and 8·63 in those aged 20-24 years), among fishing communities in Uganda (12·40 per 100 person-years in females aged 15-19 years and 4·70 in those aged 20-24 years), and among female sex workers aged 18-24 years in South Africa (13·20 per 100 person-years) and Zimbabwe (10·80). In pooled rates from the general population studies, the greatest sex differentials were in the youngest age groups-ie, females aged 15-19 years compared with male peers in both southern African (pooled relative risk 5·94, 95% CI 3·39-10·44) and eastern African countries (3·22, 1·51-6·87), and not significantly different among those aged 25-29 years in either region. Incidence often peaked earlier (during teenage years) among high-risk groups compared with general populations. Since 2005, HIV incidence among adolescent girls and young women declined in Rakai (Uganda) and Manicaland (Zimbabwe), and also declined among female sex workers in Kenya, but not in the highest-risk communities in South Africa and Uganda.

INTERPRETATION

Few sources of direct estimates of HIV incidence exist in high-burden countries and trend analyses with disaggregated data for age and sex are rare but indicate recent declines among adolescent girls and young women. In some of the highest-risk settings, however, little evidence exists to suggest ART availability and other efforts slowed transmission by 2016. Despite wide geographical diversity in absolute levels of incidence in adolescent girls and young women, risk relative to males persisted in all settings, with the greatest sex differentials in the youngest age groups. To end new infections among the growing population of adolescents in sub-Saharan Africa, prevention programmes must address gender inequalities driving excessive risk among adolescent girls.

FUNDING

This work was conducted as part of a planning grant funded by the Bill & Melinda Gates Foundation.

Longitudinal Trends in the Prevalence of Detectable HIV Viremia: Population-Based Evidence From Rural KwaZulu-Natal, South Africa

Background

The prevalence of detectable viremia has previously been used to infer the potential for ongoing human immunodeficiency virus (HIV) transmission. To date, no study has evaluated the longitudinal change in the prevalence of detectable viremia within the HIV-positive community (PDV+) and the entire population (PDVP) using data from a sub-Saharan African setting.

Methods

In 2011, 2013, and 2014, we obtained 6752 HIV-positive and 15415 HIV-negative test results from a population-based surveillance system in the KwaZulu-Natal province of South Africa. We quantified the PDV+ as the proportion of the 6752 HIV-positive results with a viral load >1550 copies/mL and the PDVP as the proportion of the 6752 HIV-positive and 15415 HIV-negative results with a viral load >1550 copies/mL.

Results

Between 2011 and 2014, the PDV+ decreased by 16.5 percentage points (pp) for women (from 71.8% to 55.3%) and 10.6 pp for men (from 77.8% to 67.2%). However, a steady rise in the overall HIV prevalence, from 26.7% to 32.4%, offset the declines in the PDV+ for both sexes. For women, the PDVP decreased by only 2.1 pp, from 21.3% to 19.2%, but for men, the PDVP actually increased by 1.6 pp, from 14.6% to 16.2%, over the survey period.

Conclusions

The PDV+, which is currently being tracked under the UNAIDS 90-90-90 targets, may not be an accurate indicator of the potential for ongoing HIV transmission. There is a critical need for countries to monitor and report the prevalence of detectable viremia among all adults, irrespective of HIV status.

Space-time clustering of recently-diagnosed tuberculosis and impact of ART scale-up: Evidence from an HIV hyper-endemic rural South African population

In HIV hyperendemic sub-Saharan African communities, particularly in southern Africa, the likelihood of achieving the Sustainable Development Goal of ending the tuberculosis (TB) epidemic by 2030 is low, due to lack of cost-effective and practical interventions in population settings. We used one of Africa’s largest population-based prospective cohorts from rural KwaZulu-Natal Province, South Africa, to measure the spatial variations in the prevalence of recently-diagnosed TB disease, and to quantify the impact of community coverage of antiretroviral therapy (ART) on recently-diagnosed TB disease. We collected data on TB disease episodes from a population-based sample of 41,812 adult individuals between 2009 and 2015. Spatial clusters (‘hotspots’) of recently-diagnosed TB were identified using a space-time scan statistic. Multilevel logistic regression models were fitted to investigate the relationship between community ART coverage and recently-diagnosed TB. Spatial clusters of recently-diagnosed TB were identified in a region characterized by a high prevalence of HIV and population movement. Every percentage increase in ART coverage was associated with a 2% decrease in the odds of recently-diagnosed TB (aOR = 0.98, 95% CI:0.97–0.99). We identified for the first time the clear occurrence of recently-diagnosed TB hotspots, and quantified potential benefit of increased community ART coverage in lowering tuberculosis, highlighting the need to prioritize the expansion of such effective population interventions targeting high-risk areas.

Identifying 'corridors of HIV transmission' in a severely affected rural South African population: a case for a shift toward targeted prevention strategies

Background

In the context of a severe generalized African HIV epidemic, the value of geographically targeted prevention interventions has only recently been given serious consideration. However, to date no study has performed a population-based analysis of the micro-geographical clustering of HIV incident infections, limiting the evidential support for such a strategy.

Methods

We followed 17 984 HIV-uninfected individuals aged 15–54 in a population-based cohort in rural KwaZulu-Natal, South Africa, and observed individual HIV sero-conversions between 2004 and 2014. We geo-located all individuals to an exact homestead of residence (accuracy <2 m). We then employed a two-dimensional Gaussian kernel of radius 3 km to produce robust estimates of HIV incidence which vary across continuous geographical space. We also applied Tango's flexibly shaped spatial scan statistic to identify irregularly shaped clusters of high HIV incidence.

Results

Between 2004 and 2014, we observed a total of 2 311 HIV sero-conversions over 70 534 person-years of observation, at an overall incidence of 3.3 [95% confidence interval (CI), 3.1-3.4] per 100 person-years. Three large irregularly-shaped clusters of new HIV infections (relative risk = 1.6, 1.7 and 2.3) were identified in two adjacent peri-urban communities near the National Road (P = 0.001, 0.015) as well as in a rural node bordering a recent coal mine development (P = 0.020), respectively. Together the clusters had a significantly higher age-sex standardized incidence of 5.1 (95% CI, 4.7-5.6) per 100 person-years compared with a standardized incidence of 3.0 per 100 person-years (95% CI, 2.9-3.2) in the remainder of the study area. Though these clusters comprise just 6.8% of the study area, they account for one out of every four sero-conversions observed over the study period.

Conclusions

Our study has revealed clear ‘corridors of transmission’ in this typical rural, hyper-endemic population. Even in a severely affected rural African population, an approach that seeks to provide preventive interventions to the most vulnerable geographies could be more effective and cost-effective in reducing the overall rate of new HIV infections. There is an urgent need to develop and test such interventions as part of an overall combination prevention approach.

Estimating trends in the incidence rate with interval censored data and time-dependent covariates

We propose a multiple imputation method for estimating the incidence rate with interval censored data and time-dependent (and/or time-independent) covariates. The method has two stages. First, we use a semi-parametric G-transformation model to estimate the cumulative baseline hazard function and the effects of the time-dependent (and/or time-independent covariates) on the interval censored infection times. Second, we derive the participant's unique cumulative distribution function and impute infection times conditional on the covariate values. To assess performance, we simulated infection times from a Cox proportional hazards model and induced interval censoring by varying the testing rate, e.g., participants test 100%, 75%, 50% of the time, etc. We then compared the incidence rate estimates from our G-imputation approach with single random-point and mid-point imputation. By comparison, our G-imputation approach gave more accurate incidence rate estimates and appropriate standard errors for models with time-independent covariates only, time-dependent covariates only, and a mixture of time-dependent and time-independent covariates across various testing rates. We demonstrate, for the first time, a multiple imputation approach for incidence rate estimation with interval censored data and time-dependent (and/or time-independent) covariates.

Effect of ART scale-up and female migration intensity on risk of HIV acquisition: results from a population-based cohort in KwaZulu-Natal, South Africa

BACKGROUND

Despite increased antiretroviral therapy (ART) coverage, the incidence of HIV infection among women in rural South Africa remains high. While many socio-demographic and behavioral factors have been identified, the effect of female migration intensity on the risk of HIV acquisition before and after ART scale-up has not been evaluated in the country.

METHODS

We followed 13,315 female participants aged 15-49 who were HIV-uninfected at baseline and recorded their migration events between 2004 and 2015. Using a Cox proportional hazard model, we estimated the time to HIV acquisition among the women, adjusting for annual migration intensity (high: ≥2 events/year, moderate = 1 event/year, and low = 0 event/year) before and after ART scale-up in 2010.

RESULTS

1998 (15%) new HIV-infection events were recorded during the observation period. Overall, high migration intensity was associated with an increased HIV acquisition risk among women when compared with low migration intensity (HR = 2.88, 95% CI: 1.56-5.53). Among those with high migration intensity, the risk of HIV acquisition was significantly lower in the post-ART period compared to the pre-ART period, after controlling for key socio-demographic and behavioural covariates (aHR = 0.18, 95% CI 0.04-0.83).

CONCLUSIONS

Women who migrated frequently after ART scale-up had a significantly reduced HIV acquisition risk compared to those before its implementation. While this reduction is encouraging, women who migrate frequently remain at high risk of HIV acquisition. In the era of ART, there remains a critical need for public health interventions to reduce the risk of HIV acquisition in this highly vulnerable population.

Moderate-to-High Levels of Pretreatment HIV Drug Resistance in KwaZulu-Natal Province, South Africa

There is evidence of increasing levels of pretreatment HIV drug resistance (PDR) in Southern Africa. We used data from two large population-based HIV surveillance studies to estimate prevalence of PDR in KwaZulu-Natal, the province with the highest HIV prevalence in South Africa. Sanger sequencing was performed on samples obtained from a longitudinal HIV surveillance program (study A, 2013-2014) and the HIV Incidence Provincial Surveillance System (study B, 2014-2015). Sequences were included for adult HIV positive participants (age ≥15 years for study A, age 15-49 years for study B) with no documented prior exposure to antiretroviral therapy (ART). Overall and drug class-specific PDR was estimated using the World Health Organization 2009 surveillance drug resistance mutation (SDRM) list, and phylogenetic analysis was performed to establish evidence of drug resistance transmission linkage. A total of 1,845 sequences were analyzed (611 study A; 1,234 study B). An overall PDR prevalence of 9.2% [95% confidence interval (CI) 7.0-11.7] was observed for study A and 11.0% (95% CI 8.9-13.2) for study B. In study B, the prevalence of non-nucleoside reverse-transcriptase inhibitor (NNRTI) PDR exceeded 10% for sequences collected in 2014 (10.2%, 95% CI 7.5-12.9). The most prevalent SDRMs were K103NS (7.5%), M184VI (2.4%), and V106AM (1.4%). There was no evidence of large transmission chains of drug-resistant virus. High level NNRTI PDR (>10%) suggests a need to modify the standard first-line ART regimen and to focus attention on improving the quality of HIV prevention, treatment, and care.

High HIV incidence and low uptake of HIV prevention services: The context of risk for young male adults prior to DREAMS in rural KwaZulu-Natal, South Africa

BACKGROUND

Young men are less likely than young women to engage with HIV prevention and care, and their HIV-related mortality is higher. We describe HIV incidence and uptake of HIV services in men 20-29 years(y) in rural KwaZulu-Natal, South Africa, before the roll-out of DREAMS.

METHODS

We used data from a population-based demographic and HIV surveillance cohort. HIV incidence was estimated from anonymised testing in an annual serosurvey. Service uptake was assessed in 2011 and 2015, through two self-reported outcomes: 1) HIV testing in the past 12 months(m); 2) voluntary medical male circumcision(VMMC). Logistic regression was used to estimate odds ratios(OR) and 95% confidence intervals(CI) for factors associated with each outcome.

RESULTS

HIV incidence in 2011-2015 was 2.6/100 person-years (95%CI = 2.0-3.4) and 4.2 (95%CI = 3.1-5.6) among men 20-24y and 25-29y, respectively, with no significant change from 2006-2010. N = 1311 and N = 1221 young men participated in the 2011 and 2015 surveys, respectively. In both years, <50% reported testing for HIV in the past 12m. In 2011, only 5% reported VMMC, but coverage in 2015 increased to 40% and 20% in men 20-24y and 25-29y, respectively. HIV testing was positively associated with higher education and mobility. Testing uptake was higher in men reporting >1 partner in the past 12m, or condom use at last sex, but lower in those reporting a casual partner (adjusted (a)OR = 0.53, 95%CI = 0.37-0.75). VMMC uptake was associated with survey year and higher education. Men aged 25-29y and those who were employed (aOR = 0.66; 95%CI = 0.49-0.89) were less likely to report VMMC.

CONCLUSIONS

HIV incidence in men 20-29y was very high, and pre-exposure prophylaxis (PrEP) should be considered in this population. Uptake of services was low. VMMC coverage increased dramatically from 2011 to 2015, especially among younger men, suggesting a demand for this service. Interventions designed with and for young men are urgently needed.

Accounting for non-response bias using participation incentives and survey design: An application using gift vouchers

Standard corrections for missing data rely on the strong and generally untestable assumption of missing at random. Heckman-type selection models relax this assumption, but have been criticized because they typically require a selection variable which predicts non-response but not the outcome of interest, and can impose bivariate normality. In this paper we illustrate an application using a copula methodology which does not rely on bivariate normality. We implement this approach in data on HIV testing at a demographic surveillance site in rural South Africa which are affected by non-response. Randomized incentives are the ideal selection variable, particularly when implemented ex ante to deal with potential missing data. However, elements of survey design may also provide a credible method of correcting for non-response bias ex post. For example, although not explicitly randomized, allocation of food gift vouchers during our survey was plausibly exogenous and substantially raised participation, as did effective survey interviewers. Based on models with receipt of a voucher and interviewer identity as selection variables, our results imply that 37% of women in the population under study are HIV positive, compared to imputation-based estimates of 28%. For men, confidence intervals are too wide to reject the absence of non-response bias. Consistent results obtained when comparing different selection variables and error structures strengthen these conclusions. Our application illustrates the feasibility of the selection model approach when combined with survey metadata.

Who are the male sexual partners of adolescent girls and young women? Comparative analysis of population data in three settings prior to DREAMS roll-out

BACKGROUND

The DREAMS (Determined Resilient Empowered AIDS-free, Mentored and Safe) Partnership aims to reduce HIV incidence among adolescent girls and young women (AGYW,15-24y) with a core package of evidence-based interventions. Some interventions, including voluntary HIV counselling and testing and circumcision, will be targeted at the male sexual partners of AGYW. A priority of DREAMS is to characterise the male partners for effective targeting.

METHODS

Using population-based data (2010-2015) in three DREAMS impact evaluation settings in Kenya and South Africa, we describe the demographic characteristics and sexual behaviour of male partners reported by AGYW, and the characteristics of males who report sexual activity with AGYW.

RESULTS

In all settings, over 90% of recent male partners reported by AGYW were aged <35 years. Median ages of spousal and non-spousal partners were 29 and 23 years respectively in uMkhanyakude (rural South Africa) and 21 and 20 years respectively in Nairobi (urban Kenya). Most males reporting an AGYW partner had never been married (89%) and many were in school (39%). Most male partners reported only 1 AGYW partner in the past year; in Gem (rural Kenya) and Nairobi 25%-29% reported 2+(AGYW or older female) partners. Concurrent partners were reported by 16% of male partners in Gem and 3-4% in uMkhanyakude. Two thirds of male partners in Gem reported testing for HIV in the past 6 months and under half in uMkhanyakude reported testing for HIV in the past year. Almost all (96%) partners in Nairobi were circumcised, compared to 45% in Gem and 43% in uMkhanyakude.

CONCLUSIONS

With almost all AGYW's sexual partners aged 15-34 years, this is an appropriate target group for DREAMS interventions. Encouraging young men to reduce their number of partners and concurrency, and uptake prevention and treatment services such as HIV testing, circumcision and ART is crucial in the effort to reduce HIV among both AGYW and young men.

Community-based HIV prevalence in KwaZulu-Natal, South Africa: results of a cross-sectional household survey

BACKGROUND

In high HIV burden settings, maximising the coverage of prevention strategies is crucial to achieving epidemic control. However, little is known about the reach and effect of these strategies in some communities.

METHODS

We did a cross-sectional community survey in the adjacent Greater Edendale and Vulindlela areas in the uMgungundlovu district, KwaZulu-Natal, South Africa. Using a multistage cluster sampling method, we randomly selected enumeration areas, households, and individuals. One household member (aged 15-49 years) selected at random was invited for survey participation. After obtaining consent, questionnaires were administered to obtain sociodemographic, psychosocial, and behavioural information, and exposure to HIV prevention and treatment programmes. Clinical samples were collected for laboratory measurements. Statistical analyses were done accounting for multilevel sampling and weighted to represent the population. A multivariable logistic regression model assessed factors associated with HIV infection.

FINDINGS

Between June 11, 2014, and June 22, 2015, we enrolled 9812 individuals. The population-weighted HIV prevalence was 36·3% (95% CI 34·8-37·8, 3969 of 9812); 44·1% (42·3-45·9, 2955 of 6265) in women and 28·0% (25·9-30·1, 1014 of 3547) in men (p<0·0001). HIV prevalence in women aged 15-24 years was 22·3% (20·2-24·4, 567 of 2224) compared with 7·6% (6·0-9·3, 124 of 1472; p<0·0001) in men of the same age. Prevalence peaked at 66·4% (61·7-71·2, 517 of 760) in women aged 35-39 years and 59·6% (53·0-66·3, 183 of 320) in men aged 40-44 years. Consistent condom use in the last 12 months was 26·5% (24·1-28·8, 593 of 2356) in men and 22·7% (20·9-24·4, 994 of 4350) in women (p=0·0033); 35·7% (33·4-37·9, 1695 of 5447) of women's male partners and 31·9% (29·5-34·3, 1102 of 3547) of men were medically circumcised (p<0·0001), and 45·6% (42·9-48·2, 1251 of 2955) of women and 36·7% (32·3-41·2, 341 of 1014) of men reported antiretroviral therapy (ART) use (p=0·0003). HIV viral suppression was achieved in 54·8% (52·0-57·5, 1574 of 2955) of women and 41·9% (37·1-46·7, 401 of 1014) of men (p<0·0001), and 87·2% (84·6-89·8, 1086 of 1251) of women and 83·9% (78·5-89·3, 284 of 341; p=0·3670) of men on ART. Age, incomplete secondary schooling, being single, having more than one lifetime sex partner (women), sexually transmitted infections, and not being medically circumcised were associated with HIV-positive status.

INTERPRETATION

The HIV burden in specific age groups, the suboptimal differential coverage, and uptake of HIV prevention strategies justifies a location-based approach to surveillance with finer disaggregation by age and sex. Intensified and customised approaches to seek, identify, and link individuals to HIV services are crucial to achieving epidemic control in this community.

FUNDING

The President's Emergency Plan for AIDS Relief through the Centers for Disease Control and Prevention.

Evaluating the impact of the DREAMS partnership to reduce HIV incidence among adolescent girls and young women in four settings: a study protocol

Background

HIV risk remains unacceptably high among adolescent girls and young women (AGYW) in southern and eastern Africa, reflecting structural and social inequities that drive new infections. In 2015, PEPFAR (the United States President’s Emergency Plan for AIDS Relief) with private-sector partners launched the DREAMS Partnership, an ambitious package of interventions in 10 sub-Saharan African countries. DREAMS aims to reduce HIV incidence by 40% among AGYW over two years by addressing multiple causes of AGYW vulnerability. This protocol outlines an impact evaluation of DREAMS in four settings.

Methods

To achieve an impact evaluation that is credible and timely, we describe a mix of methods that build on longitudinal data available in existing surveillance sites prior to DREAMS roll-out. In three long-running surveillance sites (in rural and urban Kenya and rural South Africa), the evaluation will measure: (1) population-level changes over time in HIV incidence and socio-economic, behavioural and health outcomes among AGYW and young men (before, during, after DREAMS); and (2) causal pathways linking uptake of DREAMS interventions to ‘mediators’ of change such as empowerment, through to behavioural and health outcomes, using nested cohort studies with samples of ~ 1000–1500 AGYW selected randomly from the general population and followed for two years. In Zimbabwe, where DREAMS includes an offer of pre-exposure HIV prophylaxis (PrEP), cohorts of young women who sell sex will be followed for two years to measure the impact of ‘DREAMS+PrEP’ on HIV incidence among young women at highest risk of HIV. In all four settings, process evaluation and qualitative studies will monitor the delivery and context of DREAMS implementation. The primary evaluation outcome is HIV incidence, and secondary outcomes include indicators of sexual behavior change, and social and biological protection.

Discussion

DREAMS is, to date, the most ambitious effort to scale-up combinations or ‘packages’ of multi-sectoral interventions for HIV prevention. Evidence of its effectiveness in reducing HIV incidence among AGYW, and demonstrating which aspects of the lives of AGYW were changed, will offer valuable lessons for replication.

Electronic supplementary material

The online version of this article (10.1186/s12889-018-5789-7) contains supplementary material, which is available to authorized users.

Do gifts increase consent to home-based HIV testing? A difference-in-differences study in rural KwaZulu-Natal, South Africa

Background

Despite the importance of HIV testing for controlling the HIV epidemic, testing rates remain low. Efforts to scale up testing coverage and frequency in hard-to-reach and at-risk populations commonly focus on home-based HIV testing. This study evaluates the effect of a gift (a US$5 food voucher for families) on consent rates for home-based HIV testing.

Methods

We use data on 18 478 individuals (6 418 men and 12 060 women) who were successfully contacted to participate in the 2009 and 2010 population-based HIV surveillance carried out by the Wellcome Trust's Africa Health Research Institute in rural KwaZulu-Natal, South Africa. Of 18 478 potential participants contacted in both years, 35% (6 518) consented to test in 2009, and 41% (7 533) consented to test in 2010. Our quasi-experimental difference-in-differences approach controls for unobserved confounding in estimating the causal effect of the intervention on HIV-testing consent rates.

Results

Allocation of the gift to a family in 2010 increased the probability of family members consenting to test in the same year by 25 percentage points [95% confidence interval (CI) 21-30 percentage points; P  < 0.001]. The intervention effect persisted, slightly attenuated, in the year following the intervention (2011).

Conclusions

In HIV hyperendemic settings, a gift can be highly effective at increasing consent rates for home-based HIV testing. Given the importance of HIV testing for treatment uptake and individual health, as well as for HIV treatment-as-prevention strategies and for monitoring the population impact of the HIV response, gifts should be considered as a supportive intervention for HIV-testing initiatives where consent rates have been low.

Incidence rate estimation, periodic testing and the limitations of the mid-point imputation approach

Background

It is common to use the mid-point between the latest-negative and earliest-positive test dates as the date of the infection event. However, the accuracy of the mid-point method has yet to be systematically quantified for incidence studies once participants start to miss their scheduled test dates.

Methods

We used a simulation-based approach to generate an infectious disease epidemic for an incidence cohort with a high (80-100%), moderate (60-79.9%), low (40-59.9%) and poor (30-39.9%) testing rate. Next, we imputed a mid-point and random-point value between the participant's latest-negative and earliest-positive test dates. We then compared the incidence rate derived from these imputed values with the true incidence rate generated from the simulation model.

Results

The mid-point incidence rate estimates erroneously declined towards the end of the observation period once the testing rate dropped below 80%. This decline was in error of approximately 9%, 27% and 41% for a moderate, low and poor testing rate, respectively. The random-point method did not introduce any systematic bias in the incidence rate estimate, even for testing rates as low as 30%.

Conclusions

The mid-point assumption of the infection date is unjustified and should not be used to calculate the incidence rate once participants start to miss the scheduled test dates. Under these conditions, we show an artefactual decline in the incidence rate towards the end of the observation period. Alternatively, the single random-point method is straightforward to implement and produces estimates very close to the true incidence rate.

HIV Prevention Efforts and Incidence of HIV in Uganda

BACKGROUND

To assess the effect of a combination strategy for prevention of human immunodeficiency virus (HIV) on the incidence of HIV infection, we analyzed the association between the incidence of HIV and the scale-up of antiretroviral therapy (ART) and medical male circumcision in Rakai, Uganda. Changes in population-level viral-load suppression and sexual behaviors were also examined.

METHODS

Between 1999 and 2016, data were collected from 30 communities with the use of 12 surveys in the Rakai Community Cohort Study, an open, population-based cohort of persons 15 to 49 years of age. We assessed trends in the incidence of HIV on the basis of observed seroconversion data, participant-reported use of ART, participant-reported male circumcision, viral-load suppression, and sexual behaviors.

RESULTS

In total, 33,937 study participants contributed 103,011 person-visits. A total of 17,870 persons who were initially HIV-negative were followed for 94,427 person-years; among these persons, 931 seroconversions were observed. ART was introduced in 2004, and by 2016, ART coverage was 69% (72% among women vs. 61% among men, P<0.001). HIV viral-load suppression among all HIV-positive persons increased from 42% in 2009 to 75% by 2016 (P<0.001). Male circumcision coverage increased from 15% in 1999 to 59% by 2016 (P<0.001). The percentage of adolescents 15 to 19 years of age who reported never having initiated sex (i.e., delayed sexual debut) increased from 30% in 1999 to 55% in 2016 (P<0.001). By 2016, the mean incidence of HIV infection had declined by 42% relative to the period before 2006 (i.e., before the scale-up of the combination strategy for HIV prevention) - from 1.17 cases per 100 person-years to 0.66 cases per 100 person-years (adjusted incidence rate ratio, 0.58; 95% confidence interval [CI], 0.45 to 0.76); declines were greater among men (adjusted incidence rate ratio, 0.46; 95% CI, 0.29 to 0.73) than among women (adjusted incidence rate ratio, 0.68; 95% CI, 0.50 to 0.94).

CONCLUSIONS

In this longitudinal study, the incidence of HIV infection declined significantly with the scale-up of a combination strategy for HIV prevention, which provides empirical evidence that interventions for HIV prevention can have a population-level effect. However, additional efforts are needed to overcome disparities according to sex and to achieve greater reductions in the incidence of HIV infection. (Funded by the National Institute of Allergy and Infectious Diseases and others.).

High percentage of undiagnosed HIV cases within a hyperendemic South African community: a population-based study

BACKGROUND

Undiagnosed HIV infections could undermine efforts to reverse the global AIDS epidemic by 2030. In this study, we estimated the percentage of HIV-positive persons who remain undiagnosed within a hyperendemic South African community.

METHODS

The data come from a population-based surveillance system located in the Umkhanyakude district of the northern KwaZulu-Natal province, South Africa. We annually tested 38 661 adults for HIV between 2005 and 2016. Using the HIV-positive test results of 12 039 (31%) participants, we then back-calculated the incidence of infection and derived the number of undiagnosed cases from this result.

RESULTS

The percentage of undiagnosed HIV cases decreased from 29.3% in 2005 to 15.8% in 2011. During this period, however, approximately 50% of the participants refused to test for HIV, which lengthened the average time from infection to diagnosis. Consequently, the percentage of undiagnosed HIV cases reversed direction and steadily increased from 16.1% to 18.9% over the 2012-2016 period.

CONCLUSIONS

Results from this hyperendemic South African setting show that the HIV testing rate is low, with long infection times, and an unsatisfactorily high percentage of undiagnosed cases. A high level of repeat HIV testing is needed to minimise the time from infection to diagnosis if the global AIDS epidemic is to be reversed within the next two decades.

Tuberculosis mortality and the male survival deficit in rural South Africa: An observational community cohort study

BACKGROUND

Women live on average five years longer than men, and the sex difference in longevity is typically lower in populations with high mortality. South Africa-a high mortality population with a large sex disparity-is an exception, but the causes of death that contribute to this difference are not well understood.

METHODS

Using data from a demographic surveillance system in rural KwaZulu-Natal (2000-2014), we estimate differences between male and female adult life expectancy by HIV status. The contribution of causes of death to these life expectancy differences are computed with demographic decomposition techniques. Cause of death information comes from verbal autopsy interviews that are interpreted with the InSilicoVA tool.

RESULTS

Adult women lived an average of 10.4 years (95% confidence Interval 9.0-11.6) longer than men. Sex differences in adult life expectancy were even larger when disaggregated by HIV status: 13.1 (95% confidence interval 10.7-15.3) and 11.2 (95% confidence interval 7.5-14.8) years among known HIV negatives and positives, respectively. Elevated male mortality from pulmonary tuberculosis (TB) and external injuries were responsible for 43% and 31% of the sex difference in life expectancy among the HIV negative population, and 81% and 16% of the difference among people living with HIV.

CONCLUSIONS

The sex differences in adult life expectancy in rural KwaZulu-Natal are exceptionally large, atypical for an African population, and largely driven by high male mortality from pulmonary TB and injuries. This is the case for both HIV positive and HIV negative men and women, signalling a need to improve the engagement of men with health services, irrespective of their HIV status.

Human immunodeficiency virus care cascade among sub-populations in Rakai, Uganda: an observational study

INTRODUCTION

To assess progress towards the UNAIDS 90-90-90 initiative targets, we examined the HIV care cascade in the population-based Rakai Community Cohort Study (RCCS) in rural Uganda and examined differences between sub-groups.

METHODS

Self-reports and clinical records were used to assess the proportion achieving each stage in the cascade. Statistical inference based on a χ2 test for categorical variables and modified Poisson regression were used to estimate prevalence risk ratios (PRRs) and 95% confidence intervals (CI) for enrolment into care and initiating antiretroviral therapy (ART).

RESULTS

From September 2013 through December 2015, 3,666 HIV-positive participants were identified in the RCCS. As of December 2015, 98% had received HIV Counseling and Testing (HCT), 74% were enrolled in HIV care, and 63% had initiated ART of whom 92% were virally suppressed after 12 months on ART. Engagement in care was lower among men than women (enrolment in care: adjPRR 0.84, 95% CI 0.77-0.91; ART initiation: adjPRR 0.75, 95% CI 0.69-0.82), persons aged 15-24 compared to those aged 30-39 (enrolment: adjPRR 0.72, 95% CI 0.63-0.82; ART: adjPRR 0.69, 95%CI 0.60-0.80), unmarried persons (enrolment: adjPRR 0.84, 95% CI 0.71-0.99; ART adjPRR 0.80, 95% CI 0.66-0.95), and new in-migrants (enrolment: adjPRR 0.75, 95% CI 0.67-0.83; ART: adjPRR 0.76, 95% CI 0.67-0.85). This cohort achieved 98-65-92 towards the UNAIDS "90-90-90" targets with an estimated 58% of the entire HIV-positive RCCS population virally suppressed.

CONCLUSIONS

This cohort achieved over 90% in both HCT and viral suppression among ART users, but only 65% in initiating ART, likely due to both an ART eligibility criterion of <500 CD4 cells/mL and suboptimal entry into care among men, younger individuals, and in-migrants. Interventions are needed to promote enrolment in HIV care, particular for hard-to-reach sub-populations.

From HIV infection to therapeutic response: a population-based longitudinal HIV cascade-of-care study in KwaZulu-Natal, South Africa

BACKGROUND

Standard approaches to estimation of losses in the HIV cascade of care are typically cross-sectional and do not include the population stages before linkage to clinical care. We used indiviual-level longitudinal cascade data, transition by transition, including population stages, both to identify the health-system losses in the cascade and to show the differences in inference between standard methods and the longitudinal approach.

METHODS

We used non-parametric survival analysis to estimate a longitudinal HIV care cascade for a large population of people with HIV residing in rural KwaZulu-Natal, South Africa. We linked data from a longitudinal population health surveillance (which is maintained by the Africa Health Research Institute) with patient records from the local public-sector HIV treatment programme (contained in an electronic clinical HIV treatment and care database, ARTemis). We followed up all people who had been newly detected as having HIV between Jan 1, 2006, and Dec 31, 2011, across six cascade stages: three population stages (first positive HIV test, HIV status knowledge, and linkage to care) and three clinical stages (eligibility for antiretroviral therapy [ART], initiation of ART, and therapeutic response). We compared our estimates to cross-sectional cascades in the same population. We estimated the cumulative incidence of reaching a particular cascade stage at a specific time with Kaplan-Meier survival analysis.

FINDINGS

Our population consisted of 5205 individuals with HIV who were followed up for 24 031 person-years. We recorded 598 deaths. 4539 individuals gained knowledge of their positive HIV status, 2818 were linked to care, 2151 became eligible for ART, 1839 began ART, and 1456 had successful responses to therapy. We used Kaplan-Meier survival analysis to adjust for censorship due to the end of data collection, and found that 8 years after testing positive in the population health surveillance, 16% had died. Among living patients, 82% knew their HIV status, 45% were linked to care, 39% were eligible for ART, 35% initiated ART, and 33% had reached therapeutic response. Median times to transition for these cascade stages were 52 months, 52 months, 20 months, 3 months, and 9 months, respectively. Compared with the population stages in the cascade, the transitions across the clinical stages were fast. Over calendar time, rates of linkage to care have decreased and patients presenting for the first time for care were, on average, healthier.

INTERPRETATION

HIV programmes should focus on linkage to care as the most important bottleneck in the cascade. Cascade estimation should be longitudinal rather than cross-sectional and start with the population stages preceding clinical care.

FUNDING

Wellcome Trust, PEPFAR.

Analyses of Sensitivity to the Missing-at-Random Assumption Using Multiple Imputation With Delta Adjustment: Application to a Tuberculosis/HIV Prevalence Survey With Incomplete HIV-Status Data

Multiple imputation with delta adjustment provides a flexible and transparent means to impute univariate missing data under general missing-not-at-random mechanisms. This facilitates the conduct of analyses assessing sensitivity to the missing-at-random (MAR) assumption. We review the delta-adjustment procedure and demonstrate how it can be used to assess sensitivity to departures from MAR, both when estimating the prevalence of a partially observed outcome and when performing parametric causal mediation analyses with a partially observed mediator. We illustrate the approach using data from 34,446 respondents to a tuberculosis and human immunodeficiency virus (HIV) prevalence survey that was conducted as part of the Zambia-South Africa TB and AIDS Reduction Study (2006-2010). In this study, information on partially observed HIV serological values was supplemented by additional information on self-reported HIV status. We present results from 2 types of sensitivity analysis: The first assumed that the degree of departure from MAR was the same for all individuals with missing HIV serological values; the second assumed that the degree of departure from MAR varied according to an individual's self-reported HIV status. Our analyses demonstrate that multiple imputation offers a principled approach by which to incorporate auxiliary information on self-reported HIV status into analyses based on partially observed HIV serological values.

From HIV infection to therapeutic response: a population-based longitudinal HIV cascade-of-care study in KwaZulu-Natal, South Africa

BACKGROUND

Standard approaches to estimation of losses in the HIV cascade of care are typically cross-sectional and do not include the population stages before linkage to clinical care. We used indiviual-level longitudinal cascade data, transition by transition, including population stages, both to identify the health-system losses in the cascade and to show the differences in inference between standard methods and the longitudinal approach.

METHODS

We used non-parametric survival analysis to estimate a longitudinal HIV care cascade for a large population of people with HIV residing in rural KwaZulu-Natal, South Africa. We linked data from a longitudinal population health surveillance (which is maintained by the Africa Health Research Institute) with patient records from the local public-sector HIV treatment programme (contained in an electronic clinical HIV treatment and care database, ARTemis). We followed up all people who had been newly detected as having HIV between Jan 1, 2006, and Dec 31, 2011, across six cascade stages: three population stages (first positive HIV test, HIV status knowledge, and linkage to care) and three clinical stages (eligibility for antiretroviral therapy [ART], initiation of ART, and therapeutic response). We compared our estimates to cross-sectional cascades in the same population. We estimated the cumulative incidence of reaching a particular cascade stage at a specific time with Kaplan-Meier survival analysis.

FINDINGS

Our population consisted of 5205 individuals with HIV who were followed up for 24 031 person-years. We recorded 598 deaths. 4539 individuals gained knowledge of their positive HIV status, 2818 were linked to care, 2151 became eligible for ART, 1839 began ART, and 1456 had successful responses to therapy. We used Kaplan-Meier survival analysis to adjust for censorship due to the end of data collection, and found that 8 years after testing positive in the population health surveillance, 16% had died. Among living patients, 82% knew their HIV status, 45% were linked to care, 39% were eligible for ART, 35% initiated ART, and 33% had reached therapeutic response. Median times to transition for these cascade stages were 52 months, 52 months, 20 months, 3 months, and 9 months, respectively. Compared with the population stages in the cascade, the transitions across the clinical stages were fast. Over calendar time, rates of linkage to care have decreased and patients presenting for the first time for care were, on average, healthier.

INTERPRETATION

HIV programmes should focus on linkage to care as the most important bottleneck in the cascade. Cascade estimation should be longitudinal rather than cross-sectional and start with the population stages preceding clinical care.

FUNDING

Wellcome Trust, PEPFAR.

Association of Medical Male Circumcision and Antiretroviral Therapy Scale-up With Community HIV Incidence in Rakai, Uganda

IMPORTANCE

Medical male circumcision (MMC) and antiretroviral therapy (ART) are proven HIV prevention interventions, but there are limited data on the population-level effect of scale-up of these interventions in sub-Saharan Africa. Such evaluation is important for planning and resource allocation.

OBJECTIVE

To examine whether increasing community MMC and ART coverage was associated with reduced community HIV incidence in Rakai District, Uganda.

DESIGN, SETTING, AND PARTICIPANTS

Using person-level data from population-based surveys conducted from 1999 through 2013 in 45 rural Rakai communities, community-level ART and MMC coverage, sociodemographics, sexual behaviors, and HIV prevalence and incidence were estimated in 3 periods: prior to the availability of ART and MMC (1999-2004), during early availability of ART and MMC (2004-2007), and during mature program scale-up (2007-2013).

EXPOSURES

Community MMC coverage in males and ART coverage in HIV-positive persons of the opposite sex based on self-reported MMC status and ART use.

MAIN OUTCOMES AND MEASURES

Adjusted incidence rate ratios (IRRs) for sex-specific community HIV incidence estimated using multivariable Poisson regression with generalized estimating equations.

RESULTS

From 1999 through 2013, 44,688 persons participated in 1 or more surveys (mean age at the first survey, 24.6 years [range, 15-49]; female, 56.5%; mean survey participation rate, 92.6% [95% CI, 92.4%-92.7%]). Median community MMC coverage increased from 19% to 39%, and median community ART coverage rose from 0% to 21% in males and from 0% to 26% in females. Median community HIV incidence declined from 1.25 to 0.84 per 100 person-years in males, and from 1.25 to 0.99 per 100 person-years in females. Among males, each 10% increase in community MMC coverage was associated with an adjusted IRR of 0.87 (95% CI, 0.82-0.93). Comparing communities with MMC coverage more than 40% (mean male community incidence, 1.03 per 100 person-years) with communities with coverage of 10% or less (mean male incidence, 1.69 per 100 person-years), the adjusted IRR was 0.61 (95% CI, 0.43-0.88). For each 10% increase in female self-reported ART coverage, there was no significant reduction in male HIV incidence (adjusted IRR, 0.95 [95% CI, 0.81-1.13]). Comparing communities with female ART coverage more than 20% (mean male incidence, 0.87 per 100 person-years) to communities with female ART coverage of 20% or less (mean male incidence, 1.17 per 100 person-years), the adjusted IRR was 0.77 (95% CI, 0.61-0.98). Neither MMC nor male ART coverage was associated with lower female community HIV incidence.

CONCLUSIONS AND RELEVANCE

In Rakai, Uganda, increasing community MMC and female ART coverage was associated with lower community HIV incidence in males. If similar associations are found elsewhere, this would support further scale-up of MMC and ART for HIV prevention in sub-Saharan Africa.

Heterogeneity of the HIV epidemic in agrarian, trading, and fishing communities in Rakai, Uganda: an observational epidemiological study

BACKGROUND

Understanding the extent to which HIV burden differs across communities and the drivers of local disparities is crucial for an effective and targeted HIV response. We assessed community-level variations in HIV prevalence, risk factors, and treatment and prevention service uptake in Rakai, Uganda.

METHODS

The Rakai Community Cohort Study (RCCS) is an open, population-based cohort of people aged 15-49 years in 40 communities. Participants are HIV tested and interviewed to obtain sociodemographic, behavioural, and health information. RCCS data from Aug 10, 2011, to May 30, 2013, were used to classify communities as agrarian (n=27), trading (n=9), or lakeside fishing sites (n=4). We mapped HIV prevalence with Bayesian methods, and characterised variability across and within community classifications. We also assessed differences in HIV risk factors and uptake of antiretroviral therapy and male circumcision between community types.

FINDINGS

17 119 individuals were included, 9215 (54%) of whom were female. 9931 participants resided in agrarian, 3318 in trading, and 3870 in fishing communities. Median HIV prevalence was higher in fishing communities (42%, range 38-43) than in trading (17%, 11-21) and agrarian communities (14%, 9-26). Antiretroviral therapy use was significantly lower in both men and women in fishing communities than in trading (age-adjusted prevalence risk ratio in men 0·64, 95% CI 0·44-0·97; women 0·53, 0·42-0·66) and agrarian communities (men 0·55, 0·42-0·72; women 0·65, 0·54-0·79), as was circumcision coverage among men (vs trading 0·48, 0·42-0·55; vs agrarian 0·64, 0·56-0·72). Self-reported risk behaviours were significantly higher in men than in women and in fishing communities than in other community types.

INTERPRETATION

Substantial heterogeneity in HIV prevalence, risk factors, and service uptake in Rakai, Uganda, emphasises the need for local surveillance and the design of targeted HIV responses. High HIV burden, risk behaviours, and low use of combination HIV prevention in fishing communities make these populations a priority for intervention.

FUNDING

National Institute of Mental Health, the National Institute of Allergy and Infectious Diseases, the National Institute of Child Health and Development, and the National Institute for Allergy and Infectious Diseases Division of Intramural Research, National Institutes of Health; the Bill & Melinda Gates Foundation; and the Johns Hopkins University Center for AIDS Research.

Adjusting HIV prevalence estimates for non-participation: an application to demographic surveillance

INTRODUCTION

HIV testing is a cornerstone of efforts to combat the HIV epidemic, and testing conducted as part of surveillance provides invaluable data on the spread of infection and the effectiveness of campaigns to reduce the transmission of HIV. However, participation in HIV testing can be low, and if respondents systematically select not to be tested because they know or suspect they are HIV positive (and fear disclosure), standard approaches to deal with missing data will fail to remove selection bias. We implemented Heckman-type selection models, which can be used to adjust for missing data that are not missing at random, and established the extent of selection bias in a population-based HIV survey in an HIV hyperendemic community in rural South Africa.

METHODS

We used data from a population-based HIV survey carried out in 2009 in rural KwaZulu-Natal, South Africa. In this survey, 5565 women (35%) and 2567 men (27%) provided blood for an HIV test. We accounted for missing data using interviewer identity as a selection variable which predicted consent to HIV testing but was unlikely to be independently associated with HIV status. Our approach involved using this selection variable to examine the HIV status of residents who would ordinarily refuse to test, except that they were allocated a persuasive interviewer. Our copula model allows for flexibility when modelling the dependence structure between HIV survey participation and HIV status.

RESULTS

For women, our selection model generated an HIV prevalence estimate of 33% (95% CI 27-40) for all people eligible to consent to HIV testing in the survey. This estimate is higher than the estimate of 24% generated when only information from respondents who participated in testing is used in the analysis, and the estimate of 27% when imputation analysis is used to predict missing data on HIV status. For men, we found an HIV prevalence of 25% (95% CI 15-35) using the selection model, compared to 16% among those who participated in testing, and 18% estimated with imputation. We provide new confidence intervals that correct for the fact that the relationship between testing and HIV status is unknown and requires estimation.

CONCLUSIONS

We confirm the feasibility and value of adopting selection models to account for missing data in population-based HIV surveys and surveillance systems. Elements of survey design, such as interviewer identity, present the opportunity to adopt this approach in routine applications. Where non-participation is high, true confidence intervals are much wider than those generated by standard approaches to dealing with missing data suggest.