PANCEA ovrvw URGE2 02 v1

Review of CE research on HIV prevention in the developing world: 

Review of CE research on HIV prevention in the developing world

Extent of CE research: 

Extent of CE research ~ 12 studies, ~ 6 current paucity versus effectiveness studies due to need for translation to HIV infections and for cost data

CE research needs: 

CE research needs

More single intervention analyses: 

More single intervention analyses Intervention types Settings -- epidemic, economic, and health system conditions Populations -- MSMs, IDUs, high risk non-SW women

Intervention combinations: 

Intervention combinations Integrated strategies, defined by intervention mix and populations reached Comprehensive costs, including synergies (e.g., shared resources or economies of scope) Combined epidemic effects

Scale: 

Scale Scaling up opportunities (e.g., higher coverage rates, new populations) and constraints (e.g., facilities, staff, target population size) Scale economies (e.g., bulk purchasing, better use of fixed resources) and diseconomies (e.g., new infrastructure, higher recruitment costs)

Slide8: 

Cost function for hypothetical CSW Peer Counseling Intervention

PANCEA Project: 

PANCEA Project Preventing AIDS: Network for Cost-Effectiveness Analysis Purpose: to understand the costs and efficiency of HIV prevention programs in developing countries; and to use this information to improve HIV prevention resource allocation.

PANCEA conceptual approach: 

PANCEA conceptual approach Measure: prevention program costs prevention program outputs Estimate: cost per unit of program output behavior or risk change HIV infections averted cost per HIV infection averted

PANCEA Aims: 

PANCEA Aims Aim 1: Country selection; effectiveness review Aim 2: Econometrics Aim 3: Accounting Aim 3.5: Case studies Aim 4: Epidemic and CE modeling Aim 5: Simulation tool (PC-based) Aim 6: Intervention + country reports, web, articles

Aim 1: Country selection;effectiveness review: 

Aim 1: Country selection; effectiveness review Select ~5 countries based on geographic and epidemic diversity, logistics. Leading candidates: Uganda, S. Africa, Mexico, Brazil, India/Thailand, Russia. Systematic review of published and unpublished literature on effectiveness, including attempts to specify quantitative links between prevention outputs and change in risk (behavioral, cofactors, transmission risk) Lead role: UCSF

Aim 2: Econometrics: 

Aim 2: Econometrics Collect survey data on facilities at which prevention programs based, including costs, outputs, and wide range of predictors of productivity. ~ 200 facilities (40 / country), 400? programs. Develop cost functions using health facility survey data and accounting data Lead role: World Bank

Aim 3: Accounting: 

Aim 3: Accounting Collect detailed cost and output data over time at prevention progams, using standard costing protocols; conduct qualitative interviews on scaling up history and potential. ~ 40 programs (8 / country) Develop cost functions for individual and types of programs; inform development of cost functions. Lead role: UCSF

Aim 3.5: Case Studies: 

Aim 3.5: Case Studies Conduct interviews with key informants on prevention program history and environment. ~ 10 sites. “3.5” because late, important addition Use to interpret and provide context for economic findings, also to refine data collection tools Lead role: Axios

Aim 4: Epidemic and CE modeling: 

Aim 4: Epidemic and CE modeling Adapt epidemic models to generic and site-specific country settings, with enhanced facility for modeling prevention programs (e.g., how population is compartmentalized). Generate epidemic outcomes (e.g., HIV infections) for different prevention mixes, yielding “pseudo-data” to derive benefit functions (to combine with cost functions). Lead role: Imperial College, World Bank

Aim 5: Simulation Tool: 

Aim 5: Simulation Tool Develop simplified version(s) of epidemic and cost models that can run fast on PC in real-time, for interactive examination of prevention options with country policy-makers Now: Investigator operated. Later: User operated? Lead role: UCSF, Imperial College, World Bank

Aim 6: Dissemination: 

Aim 6: Dissemination Intervention reports: findings for specific interventions, across countries. Country reports: findings for specific countries, across interventions. Web site Academic articles

PANCEA internal technical Links: 

PANCEA internal technical Links How to merge cost and benefit functions? Plan: Epidemic model generates pseudo-data for use by econometricians. How to use accounting data to adjust cost functions derived from econometric data? Plan: Potentially, alter variable selection or coefficients … unresolved. How to use case study data to improve quantitative work? Plan: Serve as “gold standard” -- fuller more credible picture; identify misunderstanding of standard survey questions. How to link effectiveness to epidemic modeling? Plan: formulas that link prevention outputs (e.g., condoms) to risk inputs.

PANCEA status, Feb 02: 

PANCEA status, Feb 02 Countries largely selected (need to recruit several), interventions selected Intervention effectiveness review ~ half complete Survey tools drafted; piloting Feb/March Analysis plans/technical links outlined

PANCEA methods - team: 

PANCEA methods - team UCSF - Kahn, Marseille, Padian, Bertozzi, Pitter, staff Collaborators: INSP (Mexico) - Bertozzi, Gutierrez, + others World Bank - Over + Lundberg Imperial College - Garnett + Gregson Axios Int’l - Saba + Reeler + others 2 to be identified for data collection Funding: NIH OAR, 2 year contract

PANCEA methods - sample: 

PANCEA methods - sample ~ 8 prevention interventions (VCT, SW peer, Condom SM/promotion, Schools curriculum, Rx STDs, Mass media/IEC, MTCT, needle exchange) Criteria: effective and/or potential large expenditure of funds 5-8 countries (eg, 3 Africa, 2 Asia, 2 Latin America) Criteria: range of epidemic status and prevention success, and feasible logistics.

PANCEA methods - data collection: 

PANCEA methods - data collection Intervention effectiveness (literature review). Lead role: UCSF Prevention program costs, outputs, and potential predictors of efficiency (basic survey at 200 sites, detailed accounting for subset of 40 sites). Lead role: Axios, INSP, and other in-country Pis. Program history, special features, environment (case studies, 8 sites). Lead role: Axios

PANCEA methods - analysis + modeling: 

PANCEA methods - analysis + modeling Analysis of predictors of efficiency (eg, scale, scope, competition) in producing outputs (eg, condoms, sessions) Case study observations Modeled and/or observed links between outputs and behavior change Association of behavior change and HIV infections (generic and site-specific epidemic models)

PANCEA methods - dissemination: 

PANCEA methods - dissemination Intervention and country reports Academic manuscripts Simplified simulation model for interactive demonstration with policy makers

Intended policy uses of PANCEA - 1: 

Intended policy uses of PANCEA - 1 Prevention vs. HAART: E.g., ~$12.50 / DALY vs. ~$350 / DALY, 28:1 ratio --> prefer prevention on economic criteria Prevention A vs. B: E.g., CSW $1 / DALY vs. MTCT $10-20 / DALY

Intended policy uses of PANCEA - 2: 

Intended policy uses of PANCEA - 2 Comprehensive prevention strategies: Given prevention opportunities (populations, capacity) and CE functions (e.g., economies of scale and scope), What is most efficient/effective mix of programs? What is the “shadow cost” of political/practical constraints? Mobilizing resources: Increase the size of the pie, lower pricing.

A key practical question:Who might want/might use CE data?: 

A key practical question: Who might want/might use CE data? Ministries/ministers of health, donors/suppliers, progam managers Need to adapt CE tools, assess usefulness