Differences in Breastfeeding Duration by Maternal HIV Status: A Pooled Analysis of Nationally Representative Surveys in Sub-Saharan Africa.

BACKGROUND: Breastfeeding improves child survival but is a source of mother-to-child HIV transmission among women with unsuppressed HIV infection. Estimated HIV incidence in children is sensitive to breastfeeding duration among mothers living with HIV (MLHIV). Breastfeeding duration may vary according to maternal HIV status. SETTING: Sub-Saharan Africa. METHODS: We analyzed pooled data from nationally representative household surveys conducted during 2003-2019 that included HIV testing and elicited breastfeeding practices. We fitted survival models of breastfeeding duration by country, year, and maternal HIV status for 4 sub-Saharan African regions (Eastern, Central, Southern, and Western). RESULTS: Data were obtained from 65 surveys in 31 countries. In 2010, breastfeeding in the first month of life ("initial breastfeeding") among MLHIV ranged from 69.1% (95% credible interval: 68-79.9) in Southern Africa to 93.4% (92.7-98.0) in Western Africa. Median breastfeeding duration among MLHIV was the shortest in Southern Africa at 15.6 (14.2-16.3) months and the longest in Eastern Africa at 22.0 (21.7-22.5) months. By comparison, HIV-negative mothers were more likely to breastfeed initially (91.0%-98.7% across regions) and for longer duration (median 18.3-24.6 months across regions). Initial breastfeeding and median breastfeeding duration decreased during 2005-2015 in most regions and did not increase in any region regardless of maternal HIV status. CONCLUSIONS: MLHIV in sub-Saharan Africa are less likely to breastfeed initially and stop breastfeeding sooner than HIV-negative mothers. Since 2020, UNAIDS-supported HIV estimates have accounted for this shorter breastfeeding exposure among HIV-exposed children. MLHIV need support to enable optimal breastfeeding practices and to adhere to antiretroviral therapy for HIV treatment and prevention of postnatal mother-to-child transmission.

Updates to Spectrum's case surveillance and vital registration tool for HIV estimates and projections.

INTRODUCTION: The Case Surveillance and Vital Registration (CSAVR) model within Spectrum estimates HIV incidence trends from surveillance data on numbers of new HIV diagnoses and HIV-related deaths. This article describes developments of the CSAVR tool to more flexibly model diagnosis rates over time, estimate incidence patterns by sex and age group and by key population group. METHODS: We modelled HIV diagnosis rate trends as a mixture of three factors, including temporal and opportunistic infection components. The tool was expanded to estimate incidence rate ratios by sex and age for countries with disaggregated reporting of new HIV diagnoses and AIDS deaths, and to account for information on key populations such as men who have sex with men (MSM), males who inject drugs (MWID), female sex workers (FSW) and females who inject drugs (FWID). We used a Bayesian framework to calibrate the tool in 71 high-income or low-HIV burden countries. RESULTS: Across countries, an estimated median 89% (interquartile range [IQR]: 78%-96%) of HIV-positive adults knew their status in 2019. Mean CD4 counts at diagnosis were stable over time, with a median of 456 cells/µl (IQR: 391-508) across countries in 2019. In European countries reporting new HIV diagnoses among key populations, median estimated proportions of males that are MSM and MWID was 1.3% (IQR: 0.9%-2.0%) and 0.56% (IQR: 0.51%-0.64%), respectively. The median estimated proportions of females that are FSW and FWID were 0.36% (IQR: 0.27%-0.45%) and 0.14 (IQR: 0.13%-0.15%), respectively. HIV incidence per 100 person-years increased among MSM, with median estimates reaching 0.43 (IQR: 0.29-1.73) in 2019, but remained stable in MWID, FSW and FWID, at around 0.12 (IQR: 0.04-1.9), 0.09 (IQR: 0.06-0.69) and 0.13% (IQR: 0.08%-0.91%) in 2019, respectively. Knowledge of HIV status among HIV-positive adults gradually increased since the early 1990s to exceed 75% in more than 75% of countries in 2019 among each key population. CONCLUSIONS: CSAVR offers an approach to using routine surveillance and vital registration data to estimate and project trends in both HIV incidence and knowledge of HIV status.

Estimating and projecting the number of new HIV diagnoses and incidence in Spectrum's case surveillance and vital registration tool.

OBJECTIVE: The Joint United Nations Programme on HIV/AIDS-supported Spectrum software package is used by most countries worldwide to monitor the HIV epidemic. In Spectrum, HIV incidence trends among adults (aged 15-49 years) are derived by either fitting to seroprevalence surveillance and survey data or generating curves consistent with case surveillance and vital registration data, such as historical trends in the number of newly diagnosed infections or AIDS-related deaths. This article describes development and application of the case surveillance and vital registration (CSAVR) tool for the 2019 estimate round. METHODS: Incidence in CSAVR is either estimated directly using single logistic, double logistic, or spline functions, or indirectly via the 'r-logistic' model, which represents the (log-transformed) per-capita transmission rate using a logistic function. The propensity to get diagnosed is assumed to be monotonic, following a Gamma cumulative distribution function and proportional to mortality as a function of time since infection. Model parameters are estimated from a combination of historical surveillance data on newly reported HIV cases, mean CD4 at HIV diagnosis and estimates of AIDS-related deaths from vital registration systems. Bayesian calibration is used to identify the best fitting incidence trend and uncertainty bounds. RESULTS: We used CSAVR to estimate HIV incidence, number of new diagnoses, mean CD4 at diagnosis and the proportion undiagnosed in 31 European, Latin American, Middle Eastern, and Asian-Pacific countries. The spline model appeared to provide the best fit in most countries (45%), followed by the r-logistic (25%), double logistic (25%), and single logistic models. The proportion of HIV-positive people who knew their status increased from about 0.31 [interquartile range (IQR): 0.10-0.45] in 1990 to about 0.77 (IQR: 0.50-0.89) in 2017. The mean CD4 at diagnosis appeared to be stable, at around 410 cells/µl (IQR: 224-567) in 1990 and 373 cells/µl (IQR: 174-475) by 2017. CONCLUSION: Robust case surveillance and vital registration data are routinely available in many middle-income and high-income countries while HIV seroprevalence surveillance and survey data may be scarce. In these countries, CSAVR offers a simpler, improved approach to estimating and projecting trends in both HIV incidence and knowledge of HIV status.

Improvements in Spectrum's fit to program data tool.

OBJECTIVE: The Joint United Nations Program on HIV/AIDS-supported Spectrum software package (Glastonbury, Connecticut, USA) is used by most countries worldwide to monitor the HIV epidemic. In Spectrum, HIV incidence trends among adults (aged 15-49 years) are derived by either fitting to seroprevalence surveillance and survey data or generating curves consistent with program and vital registration data, such as historical trends in the number of newly diagnosed infections or people living with HIV and AIDS related deaths. This article describes development and application of the fit to program data (FPD) tool in Joint United Nations Program on HIV/AIDS' 2016 estimates round. METHODS: In the FPD tool, HIV incidence trends are described as a simple or double logistic function. Function parameters are estimated from historical program data on newly reported HIV cases, people living with HIV or AIDS-related deaths. Inputs can be adjusted for proportions undiagnosed or misclassified deaths. Maximum likelihood estimation or minimum chi-squared distance methods are used to identify the best fitting curve. Asymptotic properties of the estimators from these fits are used to estimate uncertainty. RESULTS: The FPD tool was used to fit incidence for 62 countries in 2016. Maximum likelihood and minimum chi-squared distance methods gave similar results. A double logistic curve adequately described observed trends in all but four countries where a simple logistic curve performed better. CONCLUSION: Robust HIV-related program and vital registration data are routinely available in many middle-income and high-income countries, whereas HIV seroprevalence surveillance and survey data may be scarce. In these countries, the FPD tool offers a simpler, improved approach to estimating HIV incidence trends.