Increasing the number of voluntary medical male circumcisions using triangulated data in the Lakes Region, Tanzania

WHAT WAS THE PROBLEM?

Male circumcision (MC) confers a significant level of protection against heterosexual HIV acquisition among men. Since 2009, voluntary medical male circumcision (VMMC) has been a key HIV prevention intervention in Tanzania, which has a generalized HIV epidemic. With nationwide MC prevalence nearing 80 percent among males aged 15-29 years (Tanzania PHIA, conducted 2016–2017), the Government of Tanzania (GoT) appears to be on-track to reach its target of 90 percent of MC prevalence among 10-29 year-olds by 2020. In the past, VMMC program planning was often based on regional male population and MC prevalence estimates, missing significant geographic variability within regions and leading to inefficiencies in program planning and implementation. Prior to program implementation, local community members’ and health workers’ knowledge often was not sought. However, as coverage increases, the population of uncircumcised men shrinks and it becomes increasingly difficult to reach eligible uncircumcised men with VMMC. Thus, to inform targeted VMMC program planning and implementation require innovative approaches in order to reach the remaining uncircumcised men in Tanzania.

WHAT IS THE SOLUTION?

Recognizing the increasing difficulty of reaching uncircumcised men with VMMC, in late 2016, IntraHealth began triangulating quantitative (program, survey, census, and geographic) and qualitative data (key stakeholder interviews) to drive strategic planning of targeted VMMC service delivery in the five Lakes Region (i.e., Mwanza, Geita, Shinyanga, Mara, and Simiyu), which are supported through PEPFAR.

IntraHealth generated ward-level MC estimates using census male population projections, MC prevalence estimates, and MC program performance data. This information was overlaid using geographic information system (GIS). This data coupled with information from community experts, enabled the identification of focused and accessible geographical areas with large numbers of uncircumcised men. With this information, comprehensive, localized strategies were developed for VMMC service delivery.

WHAT WAS THE IMPACT?

Program performance increased by more than 200 percent once this approach was implemented across regions supported by IntraHealth, as well as achievement against targets (Table 1). This success was sustained during 2018, when IntraHealth performed 305,718 MCs, contributing to 34 percent of all MCs conducted during a record-setting year of 905,313 MCs in Tanzania.

Table 1. Annual VMMCs performance, targets, and percent achievement for IntraHealth — Mwanza, Geita, Shinyanga, Mara, and Simiyu, FY2016–FY2019Q2 shows an increase in both performance and percentage of target achievement with the introduction of this solution in 2017.

capture2.JPG

This strategy also resulted in more efficient resource utilization. For example, in 2017, IntraHealth observed that 61 percent of uncircumcised men were located in 40 percent of the wards in the five IntraHealth-supported regions in Tanzania, allowing it to allocate resources accordingly. Additionally, this approach has contributed to a reduction in seasonal variation in VMMCs performed, a change that improves the ability of IntraHealth to predict staffing needs and avoid VMMC kit stockouts.

Because of the success of this approach, IntraHealth has seen a reduction of the VMMC unit expenditure (UE) to $31.12 USD per client in 2017. This is significantly lower than the national estimate (US$50) and IntraHealth’s for the previous year (US$39.36).

HOW DOES IS WORK?

This approach involved a number of steps, including:

1.       Access available data sources. Key pieces of information needed include:

  • Population projections disaggregated by age and sex by ward level (approximately equal to seven to ten villages) from the National Bureau of Statistics (NBS)

  • MC prevalence estimates (e.g., from population-based surveys such as PHIA or DHS, or smaller regional surveys that assess MC prevalence, as available)

  • Number of VMMCs performed in the area/region (from DATIM, national databases, population-based surveys, or WHO)

  • Geo-coded shape files with physical structures and geographic features such as ward boundaries, road networks, forests/vegetation, water bodies (obtained from NBS)

2.       Map generation. Once the shape files are gathered, import them into geographic information system software (e.g., ArcGIS Pro Esri 2017) in layers – starting with the ward boundaries shape file. The population and service delivery information listed above is then summarized into table format in excel sheets and imported into the GIS software overlaid on top of the shape files. This will produce visual maps of areas with high concentrations of uncircumcised men and facility locations where demand creation and VMMC scale-up can be prioritized.

Example of a map generated with triangulated data.

Example of a map generated with triangulated data.

3.       Stakeholder engagement and information sharing. The maps are presented at a local community-based stakeholder meeting to gain insights into the local context and to confirm accuracy of the information generated from the data triangulation activity. Useful information gathered includes religious, ethnic, and tribal composition of the areas identified in the maps; confirmation of physical structures; information about additional buildings that can serve as outreach sites; and other logistical information (e.g., road condition, availability of water/electricity for sterilization of instruments, and lodging for service providers). Additionally, information about areas where men spend most of their time during the day time while away from home is critical.

4.       Program planning. With this available information, the implementing partner management team can develop a program plan for targeted outreach campaigns for VMMC services. This data is used both at high level planning as well as day-to-day for service delivery.

As a specific example of this approach, in 2018, IntraHealth began implementing VMMC in two priority areas, Tarime and Musoma district councils (DCs), in Mara region. At the regional level, Mara has one of the highest MC prevalence in the country, with less than 10 percent of the male population aged 15-29 years estimated to be uncircumcised. As is common in traditionally circumcising areas, VMMC was perceived by local influential leaders as a threat to traditional practices. Thus, overcoming community resistance to buy-in from regional and district governments to initiate a VMMC program in these two priority districts was challenging. In fact, initially IntraHealth was prohibited from offering the service in Mara.

To address this challenge, IntraHealth turned to this data triangulation strategy to make the case for VMMC in Mara. During October–December 2017, ward-level MC prevalence estimates were created for Tarime and Musoma DCs using census projections, survey estimates, and VMMC program data. This granular information was mapped in ArcGIS to identify areas where MC coverage might be lower than expected in these DCs. The maps were presented to local community leaders and health staff to confirm their accuracy and provide additional context. The pockets identified as low MC prevalence corresponded to communities with a high proportion of Sukuma and Luo tribes, both of which are immigrants to Mara from traditionally non-circumcising tribes. The maps were then geo-coded with relevant features (e.g., roads, government health facilities) to plan VMMC services in these areas by establishing static sites and planning outreach camps. Over the ensuing three quarters of 2018, IntraHealth made significant progress in Mara; during January–September 2018, 13,565 men were circumcised in Mara, representing a significant achievement in an area with an estimated ~417,000 men  aged 15-29 and a high (~90 percent) MC prevalence.

SCALE-UP

The entire IntraHealth VMMC program scaled this approach after initial success was evident in 2017. IntraHealth led the process, with support from the CDC-Tanzania office. It has been important for IntraHealth’s strategy development because it allowed them to improve efficiency, as remaining uncircumcised men become more difficult to reach in areas that achieve high VMMC coverage. 

Triangulating data improves efficiency in planning and monitoring for high-impact large-scale interventions at a minimum cost. The approach can be adopted by other implementing partners to improve VMMC program performance. Furthermore, other targeted high-impact interventions should consider using interactive GIS maps to make strategic decisions.

COST OF SOLUTION

Cost of solution:  The cost incurred to implement this solution was mainly for staff training, which cost around US$ 3,000 and ArcGIS Pro Esri 2017 annual license cost $100 per user (total US$ 400).  All the data and shape-files required for this exercise were available for free from various sources.

RESOURCES

GIS Basics (ppt)

Introduction to ArcGIS (pdf)

Scale-up of VMMC in Tanzania Poster (pdf)

Shapefiles (zipped)