Pop opulation D Dyn ynamics Stron ongly y Influence H e HIV E - PowerPoint PPT Presentation

Pop opulation D Dyn ynamics Stron ongly y Influence H e HIV E V Estimates es: A A Descri escriptive Study Athena Pantazis U.S. Census Bureau Population Division Annual Meeting of the Population Association of America April 10-13, 2019, Austin, TX The presentation is released to inform interested parties of ongoing research and to encourage discussion of work in progress. Any views expressed are those of the authors and not necessarily those of the U. S. Census Bureau 1

Global HIV Estimates • Software from Avenir Health ( avenirhealth.org ) • Models developed and revised by the UNAIDS Reference Group on Estimates, Modelling and Projections ( http://www.epidem.org/ ) • Different methods for obtaining HIV incidence and prevalence estimates from available data • Cohort-component population projection starting in 1970 for obtaining HIV indicators, etc. 2

Objectives Descriptive analysis to explore how population dynamics shape HIV projections using the Spectrum model 3

Methods • Project a stationary population from 1970 to 2035 in Spectrum with 2 HIV curves: one with no antiretroviral (ART) interventions and another with Fast Track (95-95-95) targets attained by 2030 • Compare with a stable, growing population projection and a stable, declining population projection • Compare with projections for 4 fertility levels based on observed total fertility rate (TFR) trends in countries with high HIV burden • Compare with projections where timing for fertility decline is varied 4

Stationary, Growing, Shrinking Populations 5

Very High, High, Medium and Low Fertility: PLHIV Population 6

Very High, High, Medium and Low Fertility: New HIV Infections 7

Very High, High, Medium and Low Fertility: PMTCT Need 8

Timing of Fertility Decline: New HIV Infections 9

Timing of Fertility Decline: PMTCT Need 10

Conclusions • Identical HIV curves produce different estimates of key HIV indicators under different population dynamics. • These differences are generally within the estimated uncertainty, though they can still be quite large. • The differences are large enough to likely impact decision-making. • Analyzing trends obtained from these different estimates would lead to different conclusions about the success of interventions and the general trends of the HIV epidemic. Underlying population dynamics need to be considered when using estimated HIV indicators 11

Thank you! 12

Extra Slides 13

Structure Demographic Data Results Demographic and Epidemic Calculations • Number HIV+ Program Statistics •New Infections • Mother-to-child transmission • Child model •AIDS deaths •Adult model •Need for ART •Need for PMTCT Epidemic Patterns Surveillance, Prevalence / Survey and incidence trend Routine Testing Data

Tracking New Adult HIV Infections by CD4 Count On ART New Infections α 1 μ 1 c 1 >500 >500 λ 1 α 2 μ 2 c 2 350-499 350-499 λ 2 α 3 μ 3 c 3 250-349 250-349 λ 3 μ 4 α 4 c 4 200-249 200-249 λ 4 μ 5 α 5 c 5 100-199 100-199 λ 5 α 6 μ 6 c 6 50-99 50-99 λ 6 α 7 μ 7 c 7 <50 <50 AIDS Death µ = non-AIDS mortality rate, λ = progression rate, c = rate of initiating ART, α = mortality rate on ART