Naomi

Naomi estimates district-level HIV epidemic indicators using a Bayesian small-area estimation model to provide granular counts, rates, and probabilistic uncertainty ranges for people living with HIV (PLHIV), antiretroviral treatment (ART) coverage, unmet ART need, and new HIV infections, particularly in sub-Saharan Africa.

Key Features:

• Stratified Estimation: Estimates HIV-related indicators stratified by sex and five-year age groups.
• Data Integration: Integrates subnational household survey data on HIV prevalence, ART coverage, and incidence, routine antenatal service delivery data, and health service delivery records of PLHIV receiving ART.
• Joint Calibration: Small-area regressions for HIV prevalence, ART coverage, and incidence are jointly calibrated to ensure consistency across indicators.
• Empirical Bayes Framework: Employs an empirical Bayes framework to estimate outputs and provide probabilistic uncertainty ranges for each indicator.
• Geographic and Demographic Aggregation: Aggregates model outputs into various geographic and demographic stratifications for subnational analysis.
• Incidence Modeling: Models HIV incidence based on district-level prevalence and ART coverage.
• Outputs: Produces counts and rates for each indicator along with uncertainty estimates.

Scientific Applications:

• HIV program planning: Supports district-level HIV program planning, resource allocation, and target setting.
• Identification of gaps: Characterizes the distribution of PLHIV, ART coverage disparities, and unmet treatment needs across subnational units.
• Targeted interventions: Informs targeted interventions and resource allocation by combining multiple data sources into consistent small-area estimates.

Methodology:

Models HIV incidence based on district-level prevalence and ART coverage using jointly calibrated small-area regressions and an empirical Bayes approach to produce counts, rates, and probabilistic uncertainty ranges, with outputs aggregated into geographic and demographic stratifications.