Optimizing contraceptive distribution in Togo

The Togolese Ministry of Health (MoH) distributes contraceptives to women nationwide to promote family planning. They use two distribution strategies: open days, where contraceptives are provided at district hospitals, and mobile clinics, where health workers bring services to remote areas. The MoH combines these two strategies across all 39 districts every quarter to meet annual coverage targets.

The MoH typically drafts a budget at the national level and secures funding for the year. However, determining the optimal combination of open days and mobile clinics to implement within this budget at the district level and still meet nationwide coverage targets can be challenging. This complexity arises from Togo’s varied topography, differences in healthcare needs and population density, and the costs associated with implementation in each district.

To tackle these challenges, we developed an optimization model using linear programming to assist the Ministry of Health in optimizing distribution within budget while achieving its target coverage.

Given all the data, the optimization model produces a binary-valued table showing whether or not the MoH should implement a strategy in a specific district in a specific quarter (Figure 2).

To evaluate the performance of our model, we compared its recommendations to the manual planning conducted by the Ministry of Health (MoH) in 2022. We replicated the conditions of the program in 2022 to ensure a fair comparison. We found that the model achieves ~40% reduction in costs while meeting the same coverage targets (see Figure 3). This is attributed to the model’s flexibility, allowing it to allocate more resources when demand is high and fewer resources when demand is low.

What’s next?

After developing the optimization model, the next step is integrating it into the Ministry of Health’s planning process. This will enable them to continuously enhance budgeting and program implementation. Additionally, in the future, we can improve the model further to optimize not only for distribution points but also for the quantity of contraceptives to have at each supply location.

The model’s effectiveness depends on how well it aligns with the MoH’s plans and on-the-ground realities. To facilitate adoption, we had to engage in multiple rounds of discussions with decision-makers to (a) gather their feedback on our assumptions and (b) understand which types of analyses— comparative or otherwise —would be most relevant for their decision-making.

Although our model proved effective for planning, it needed to be iterative to have long-term viability. The model is based on several strong assumptions – such as how we computed costs or our belief that the program’s cost drops with subsequent implementations – which must be validated through real-world data collection. If the MoH acts on the recommendations from this version of the model and encounters issues—such as our assumed costs being too low—we can adjust the model with more realistic cost estimates, leading to better recommendations in the following year.

In summary, optimization modeling is a powerful and effective approach to resource optimization, but its success depends not only on the quality of the model itself but also on a deep understanding of the problem, the data, and active stakeholder engagement. Building trust in the model by incorporating stakeholder inputs and emphasizing the flexibility and iterative nature of model improvements can go a long way in ensuring the uptake and impact of the tool.