Abstract for: Understanding pandemic response decision pathways: a systems perspective based on learning across four EU countries

Countries in Europe, despite having shared regional coordination mechanisms, experienced important variations in COVID-19 health outcomes. This paper examines how decision pathways emerged under uncertainty in Belgium, Denmark, France, and Greece, and how system-level mechanisms shaped response options at different points in time. The aim was to understand how countries interpreted evolving signals, managed constraints, and adjusted measures within their specific national contexts and across a shared EU coordination setting. Building on 38 semi-structured interviews with stakeholders engaged in pandemic preparedness and response, insights were thematically coded using a predefined framework that integrates emergency response functions and steps in the policy cycle. Causal loop diagrams (CLDs) were constructed and validated to capture early and mid-stage dynamics that drove national response strategies against COVID-19. The CLDs reflect feedback dynamics between disease burden, policy decisions, implementation, adherence, and review mechanisms. Findings indicate that variation in the type and timing of strategies reflected differences in system capacities and decision environments. Initially, measures were largely driven by biomedical and epidemiological indicators, including saturated healthcare capacity and pandemic fear. Over time, growing socioeconomic costs of non-pharmaceutical interventions led to pandemic fatigue and reduced trust in authorities. Together with the introduction of vaccines, greater consideration for behavioral indicators led to the relaxation of measures. EU-level coordination and cross-border dynamics were important to inform and strengthen national strategies. The findings highlight how response pathways emerged from interacting system elements across the science-policy-society interface. Furthermore, they help explain why feasible options differed across settings and evolved over time, offering insights for strengthening preparedness under uncertainty. AI was used a few times only for language and style improvements