Containing the spread of infectious diseases in fatigued populations: a modelling perspective.

Infectious pathogens spread through complex and variable mechanisms conditioned on the contact between an infected host and a susceptible individual. Slowing down or controlling such spread can be achieved through a variety of interventions, affecting the chances of transmission given a contact, or impacting the probability of contact between an infected and a susceptible individual (e.g. with or without the use of pharmaceutical products, and targeting infected individuals or entire populations). The success of mitigation strategies however strongly relies on the collaboration of the population. To evaluate the effect of imperfect adherence (or fatigue) of populations on disease spread, we use mathematical models, ranging from simple analytical (describing a variety of disease and population characteristics) to highly detailed data-driven ones (with more practical value for policy makers). We investigate the role of a number of drivers of imperfect adherence on the efficacy of two milestone non-pharmaceutical targeted interventions: the self-isolation of diagnosed cases and the contact tracing measure (i.e. the isolation of the contacts of diagnosed cases). Through a detailed data-driven branching process model for COVID-19 spread in France, Belgium and Italy during the third winter wave of the pandemic, we find that these measures would not be sufficient to control an Omicron winter wave and that a booster campaign is crucial in a post-acute phase of the outbreak characterised by low population adherence. Through an analytical deterministic treatment of simple compartmental models, we find that heterogeneous adherence in the population could completely undermine the efficacy of the measures and that contact tracing has a limited mitigation efficacy (even when a share of infections is asymptomatic).