How to incorporate social vulnerability into epidemic mathematical modelling: recommendations from an international Delphi

Abstract

Epidemic mathematical modelling plays a crucial role in understanding and responding to infectious disease epidemics. However, these models often neglect social vulnerability (SV): the social, economic, political, and health system inequalities that inform disease dynamics. Despite its importance in health outcomes, SV is not routinely included in epidemic modelling. Given the critical need to include SV but limited direction, this paper aimed to develop research recommendations to incorporate SV in epidemic mathematical modelling. Using the Delphi technique, 22 interdisciplinary experts from 12 countries were surveyed to reach consensus on research recommendations. Three rounds of online surveys were completed, consisting of free-text and seven-point Likert scale questions. Descriptive statistics and inductive qualitative analyses were conducted. Consensus was reached on 27 recommendations across seven themes: collaboration, design, data selection, data sources, relationship dynamics, reporting, and calibration and sensitivity. Experts also identified 92 indicators of SV with access to sanitation (n = 14, 6.1 %), access to healthcare (n = 12, 5.3 %), and household density and composition (n = 12, 5.3 %) as the most frequently cited. Given the recent focus on the social determinants of pandemic resilience, this study provides both process and technical recommendations to incorporate SV into epidemic modelling. SV's inclusion provides a more holistic view of the real world and calls attention to communities at risk. This supports forecasting accuracy and the success of policy and programmatic interventions.