Epidemiological and Health Economic Implications of Symptom Propagation: A Mathematical Modelling Investigation

<p>Respiratory pathogens inflict a substantial burden on public health and the economy. Although the severity of symptoms caused by these pathogens can vary from asymptomatic to fatal, the factors that determine symptom severity are not fully understood. Correlations in symptom severity between infector-infectee pairs, for which evidence is accumulating, can generate large-scale clusters of severe infections that could be devastating to those most at risk, whilst also conceivably leading to chains of mild or asymptomatic infections that generate widespread immunity with minimal cost to public health. Although this effect could be harnessed to amplify the impact of interventions that reduce symptom severity, the mechanistic representation of the propagation of symptom severity within mathematical and health economic modelling of respiratory diseases is understudied.</p><p><br></p><p>We propose a novel framework for incorporating different levels of symptom propagation into models of infectious disease transmission via a single parameter, α. Varying α tunes the model from having no symptom propagation (α=0, as is typically assumed) to one where symptoms always propagate. For parameters corresponding to three respiratory pathogens — seasonal influenza, pandemic influenza and SARS-CoV-2 — we explored how symptom propagation impacted the relative epidemiological and health-economic performance of three interventions, conceptualised as vaccines with different actions: symptom-attenuating (labelled SA), infection-blocking (IB) and infection-blocking admitting only mild breakthrough infections (IB_MB).</p><p>For fixed underlying epidemiological parameters, before interventions were applied, stronger symptom propagation increased the proportion of cases that were severe. For SA and IB_MB, interventions were more effective at reducing prevalence (all infections and severe cases) for higher levels ofsymptom propagation. For IB, symptom propagation had no effect, and for seasonal influenza this intervention was most effective at reducing severe infections for all levels of symptom propagation. For pandemic influenza and SARS-CoV-2, at low intervention uptake, SA was more effective than IB; for high uptake, SA only became more effective under strong symptom propagation.</p><p><br></p><p>Overall, symptom propagation shifted the balance between competing interventions. Given the importance of determining robust public health responses, we highlight the need to gather further data on symptom propagation, with our modelling framework acting as a template for future analysis.</p><p><br></p><p><br></p><p><br></p>