Co-Identifying Policy-Relevant Modelling Questions: A Case Study of the Human Papillomavirus (HPV) Vaccine Introduction in Mozambique

1.

Mathematical models hold the potential to generate valuable evidence for shaping vaccination policies. However, maximizing their impact requires a deeper understanding of how modelling efforts can be aligned with the real-world priorities of policymakers and health officials. This study explores how structured engagement with stakeholders can help co-identify decision-relevant questions that are amenable to quantitative modelling. The focus is the human papillomavirus (HPV) vaccination programme in Mozambique.

We conducted semi-structured interviews with stakeholders involved in the HPV vaccine programme to identify key knowledge gaps in their decision-making context, i.e., practice. These were translated into research questions that informed the application of a mathematical model. An evidence brief was developed to synthesize and contextualize findings, and follow-up interviews were conducted to reflect on the utility of the evidence. Qualitative data were analysed inductively to identify emergent themes.

Stakeholders identified four priority questions: optimal vaccine delivery strategy, distributional impact, vaccine economics, and comparison with other prevention methods. They emphasized the value of tailored evidence—particularly at the provincial level—for informing financial planning, resource allocation, and advocacy. The approach facilitated collaboration between researchers and stakeholders, helped uncover previously untapped data sources, and improved the policy relevance of the modelling outputs.

This study demonstrates how co-identifying modelling questions with decision-makers can help ensure that evidence generated through mathematical models is context-specific, and policy-relevant. This type of engagement enabled clearer alignment between model development and decision-making needs—offering lessons for future applications of modelling in public health policy.