USING MATHEMATICAL MODELING FOR POLICY AND STRATEGIC PLANNING A case study of VMMC Scale-Up in Eastern and Southern Africa INTEREST 2018 Kigali, Rwanda Emmanuel Njeuhmeli, MD, MPH, MBA Senior Medical Advisor USAID Swaziland 1
Contents Introduction to mathematical modeling Barriers and enablers to evidence use Case study: Voluntary medical male circumcision (VMMC) modeling DMPPT, version 1: Supporting global advocacy DMPPT, version 2: Country-specific programming DMPPT 2, online: User-friendly program monitoring Lessons learned Framework for success 2
Introduction 3
Introduction Today: Mathematical modeling is a critical tool for HIV program planning and strategic decision making o Examine intervention impacts over a range of timescales, settings, or subpopulations o Estimate outcomes that are otherwise difficult to measure, such as trends in HIV incidence or AIDS-related deaths o Quantify long-term gains in cost-savings or epidemiological impact Disconnects between modelers and policymakers can make it difficult to translate model results from page to policy Summarize challenges to ensuring model results are used in policymaking Describe success of VMMC program in using modeling to influence VMMC policy Propose framework to guide researchers looking to produce data for decision-making 4
The Challenge: Moving Models from Page to Policy Modelers Policymakers Need data, viable methodology, time, funding Need simplicity, replicability, effectiveness, support Communication 5
Barriers and enablers to evidence use, based on literature review Barriers 1. Lack of consideration of the end-user 2. Lack of understanding on the users end Enablers 1. Collaboration with policymakers particularly early, continued collaboration 2. Consideration to the local context in which results will eventually be applied 3. Up-front acknowledgment of assumptions and limitations 4. Sufficient commitment of time to a thorough analysis 6
Case study: Voluntary Medical Male Circumcision (VMMC) modeling 7
Case Study: The Decision Maker s Program Planning Tool (DMPPT) o A series of VMMC modeling exercises has had a significant impact on global VMMC policies from 2007 to date DMPPT Advocacy - Desk exercise in 2009-2011 - Public data from 13 countries o Three iterations of the model demonstrate modeling s wide-ranging potential and uses: As a tool for advocacy, data for strategic planning, and program monitoring o Success in policy reach could serve as a blueprint for others looking to apply modeling to inform policy DMPPT 2.0 DMPPT 2 Online Strategic Planning - Applications in 10 countries - Stakeholder engagement - Customized to each country Monitoring - Simplified, web-based version - Accessible data: VMMC targets, performance, coverage, and impact projections 8
UNAIDS-WHO Joint Strategic Action Framework for Acceleration of VMMC Scale-Up: 2012 to 2016 Microsoft-Excel-based modeling tool populated with publicly-available data Demonstrated that VMMC scale-up would result in substantial reductions in HIV infections and lower health system costs Results: The 80% target was incorporated into the WHO-UNAIDS Joint Strategic Action Framework for VMMC in December 2011 All countries included VMMC scale-up in their national strategic plans PEPFAR significantly increased its VMMC budget. 9
DMPPT 2.0: Modeling and Country Engagement for Strategic Planning A second model (DMPPT 2) was created to enable analyses of impact and cost-effectiveness by age group and subnational region starting in 2013. This time, emphasized in-country engagement from start to finish. Worked closely with VMMC stakeholders in Malawi, South Africa, Swaziland, Tanzania, Uganda, Kenya, Namibia, Lesotho, Mozambique and Zimbabwe. Customized model inputs/analyses based on stakeholder inputs. All ten countries incorporated the outputs from the modelling exercise into their program planning. Uganda Kenya Tanzania Malawi Mozambique Namibia Swaziland Lesotho South Africa 10
Age prioritization: Logic-tree applied to address the value of age-specific targeting Efficiency of VMMC Number of VMMCs per HIV infection averted after 15 yrs Immediacy of impact Focus of modeling work HIV incidence reduction after 5 yrs Magnitude of impact HIV incidence reduction after 15 yrs Costeffectiveness Age strategy Cost per HIV infection averted (15 yrs) Programmatic considerations
Age prioritization across countries INDICATOR LESOTHO MALAWI MOZAMBIQUE NAMIBIA TANZANIA SWAZILAND SOUTH AFRICA UGANDA VMMC/IA 20 34 20 34 20 34 15 29 20 34 15 34 20 34 20 34 Immediacy of impact 20 29 20 34 15 29 15 29 20 34 20 29 20 34 20 34 Magnitude of impact 15 24 15 24 10 24 15 24 10 24 15 29 15 24 10 19 Costeffectiveness 15 34 15 34 15 34 15 34 15 34 15 34 15 34 15 34 New country age target 10-29 10 34 10-29 10-29 10 34 10 34* 10-34 10 34 PEPFAR age prioritization 15 29 15 29 15 29 15 29 10 29 15 29 15 29 15 29 *50% coverage for neonates, 80% among males ages 10 29, and 55% among males ages 30 34
Example: Application of the DMPPT 2.0 in Swaziland In 2013/2014, Swaziland MOH wanted to understand how to better align their VMMC targets with the varying demand observed among different age groups. After examining the results of 6 modeling scenarios, the MOH chose to scale up to 50% coverage for neonates, 80% among males ages 10-29, and 55% among males ages 30-34. These evidence-based targets were incorporated into the Male Circumcision Strategic and Operational Plan for HIV Prevention, 2014 2018. [Our] plan was chosen with the intention of balancing cost, cost-effectiveness, impact, and programmatic feasibility. Dr Simon M. Zwane, Principal Secretary of Swaziland s Ministry of Health 13
VMMC 2021: UNAIDS-WHO Framework for VMMC Takes a people-centered approach to service delivery, with appropriate packages offered to different age groups and risk profiles Calls for a sound national accountability framework and management system for expanded men s and boys health programs, with VMMC at its core Two main targets aligned with the UNAIDS fast track goals: By 2012, 90% of males aged 10-29 years will have been circumcised in priority settings, and 90% of 10-29 year-old males will have accessed age-specific health services tailored to their needs 14
DMPPT 2 Online Set-up The DMPPT 2 has been migrated into an online version. This version now operates at the district level, and generates coverage estimates, targets, and impact projections by five-year age band. The purpose of the new online version is to have a simple-to-use interface for program planners, building from the previous Excel-based tool.
DMPPT 2 online sample output Male circumcision prevalence (%) by district before VMMC program initiation and modeled current estimates of coverage
DMPPT 2 online sample output Modeled estimates of MC coverage by end 2015, by SNU Sample country 10 14 15 19 20 24 25 29 30 34 10-29 15 29 District A 56% 65% 67% 43% 27% 59% 60% District B 13% 14% 16% 18% 18% 15% 16% District C 15% 18% 18% 19% 19% 17% 18% District D 27% 31% 25% 21% 20% 27% 27% District E 46% 52% 43% 35% 29% 45% 45% District F 71% 85% 74% 51% 34% 72% 72% District G 86% 100% 92% 58% 36% 88% 89% District H 19% 24% 22% 21% 19% 21% 22% District I 32% 53% 36% 27% 24% 38% 41% District J 56% 71% 64% 50% 38% 61% 63% District K 24% 26% 21% 20% 20% 23% 23% District L 58% 65% 52% 34% 24% 54% 52% Total 50% 60% 52% 38% 28% 51% 51% Sample data only
DMPPT 2 Online Output Tables Once the user has chosen a scale-up scenario, he or she can download the corresponding output table for that scenario, which contains the following data for each subnational unit (SNU): Targets # of VMMCs required in each age group and year in order to reach the specified coverage target Coverage estimated VMMC coverage by age group for the next five years Uptake rate # of circumcisions conducted in a given age group and year divided by the number of uncircumcised men in that age group and year HIV infections averted projected HIV infections averted over the next 15 years, based on circumcisions conducted to date, as well as future scale-up circumcisions VMMC per HIV infection averted # of VMMCs required to avert one HIV infection over the next 15 years, based on the user-specified coverage target
DMPPT user statistics Total: 623 users Multi-country users: 99 Uganda: 85 Kenya: 78 Agency Users USAID 68 CDC 47 State department 9 Implementing partners, HQ staff, Ministry staff, other in-country stakeholders 499 Total 623 Namibia: 25 Tanzania: 33 Malawi: 70 Zimbabwe: 66 Mozambique: 14 South Africa: 84 Lesotho: 29 Swaziland: 40
Modeling for Policy: A Framework for Success 20
Neither the elegance of the science nor the strength of the effect predict the ease of implementation David Stanton, USAID, 2009 http://collections.plos.org/s/vmmc 21
DMPPT Lessons Learned Identify a champion (or champions) Modeling can be an effective tool to promote, plan, and monitor policy and programs: The three iterations of the modeling tool demonstrate a range of uses for modeling in HIV programming. Engage stakeholders early and often: Country stakeholders provided feedback on the DMPPT 2 model inputs, assumptions and research questions, leading to confidence in and use of the results. Get to know the policymaking process: It can be difficult to appreciate what goes into the policymaking process. With high levels of engagement, as well as coordination with global funding cycles and country processes, one can better ensure that results will feed into national strategies and plans. Plan for sustainability: At the same time, policymakers may not immediately understand what goes into the modeling process. Facilitating clear dialogue around sustainability is critical: what is needed by way of planning or support to ensure that a model is updated and applied on an ongoing basis? 22
Emmanuel Njeuhmeli, MD, MPH, MBA 23