Incorporating Health Seeking Behaviour in a Deterministic Model for Influenza
BACKGROUND: Modelling efforts during the COVID-19 pandemic highlighted the significance of health-seeking behaviour on transmission dynamics within a population. Health seeking behaviour such as testing and access of healthcare services are important considerations that impact how cases are identified through surveillance systems. In mathematical modelling, cases reported to surveillance systems are often used for retrospective and prospective analysis. METHODS: An age-stratified SIR model that incorporates case detection for influenza is described. Influenza data and case detections rates for influenza seasons (2016-2019) were estimated from Alberta Health’s administrative data. Incorporating constant and time dependent case detection rates in the model will be compared across retrospective (i.e. assess fitting using all data from each season) and prospective (i.e. performance of projections) analysis. RESULTS: Retrospective analysis showed comparable fitting results of influenza cases using both constant and time dependent case detection rates, however the final size differed. Influenza projections performed better using time dependent compared to constant case detection rates. CONCLUSION: It is important to consider implications of incorporating model assumptions that may appear to ‘fit well to data’. Based on both retrospective and prospective analysis, using a time dependent case detection rate should be considered when modelling respiratory viruses.