Using a household-structured branching process to analyse contact tracing in the SARS-CoV-2 pandemic

Abstract

We explore strategies of contact tracing, case isolation and quarantine of exposed contacts to control the SARS-CoV-2 epidemic using a branching process model with household structure. This structure reflects higher transmission risks among household members than among non-household members. We explore strategic implementation choices that make use of household structure, and investigate strategies including two-step tracing, backwards tracing, smartphone tracing and tracing upon symptom report rather than test results. The primary model outcome is the effect of contact tracing, in combination with different levels of physical distancing, on the growth rate of the epidemic. Furthermore, we investigate epidemic extinction times to indicate the time period over which interventions must be sustained. We consider effects of non-uptake of isolation/quarantine, non-adherence, and declining recall of contacts over time. Our results find that, compared to self-isolation of cases without contact tracing, a contact tracing strategy designed to take advantage of household structure allows for some relaxation of physical distancing measures but cannot completely control the epidemic absent of other measures. Even assuming no imported cases and sustainment of moderate physical distancing, testing and tracing efforts, the time to bring the epidemic to extinction could be in the order of months to years. This article is part of the theme issue ‘Modelling that shaped the early COVID-19 pandemic response in the UK’.

Public perceptions and interactions with UK COVID-19 Test, Trace and Isolate policies, and implications for pandemic infectious disease modelling

“…TTI intervention effectiveness is sensitive to the proportion of symptomatic cases who go on to isolate and test [7, 70, 16]. These actions depend upon recognition of which symptoms are intended to prompt this behaviour and are a key parameter in modelling of symptomatic testing strategies.…”

“…Children have different main symptoms to adults (headache, rhinorrhoea and fatigue) [69], and it is possible this is also a point of confusion. TTI intervention effectiveness is sensitive to the proportion of symptomatic cases who go on to isolate and test [7,70,16]. These actions depend upon recognition of which symptoms are intended to prompt this behaviour and are a key parameter in modelling of symptomatic testing strategies.…”

Modelling that shaped the early COVID-19 pandemic response in the UK

“…hospitalizations, confirmed cases [14]) used to monitor epidemic trends and, when informing likely infection events, to infer the timing of transmission between cases (e.g. known infector-infectee pairs, geographical spread with travel history to outbreak locations [15]); demographic data, used to define the population at risk; census and household data, used to characterize household transmission (see [16,17]); and social contact data, essential for predicting the impact of social distancing measures [18].…”

A Review of Using Mathematical Modeling to Improve Our Understanding of Bacteriophage, Bacteria, and Eukaryotic Interactions

“…By analyzing two possible outcomes (survival versus extinction; Lashari and Trapman, 2018 ) at each branch point, this kind of model can, for example, provide information on fluctuations in population size and the differential effects of control mechanisms (e.g., transcriptional regulation) on individuals over time ( Athreya, 2006 ; Meehan et al, 2020 ) based on the different routes individuals follow (and such information is not accessible to deterministic models). For this reason, branching processes have historically been used to study long term evolution, reproduction and extinction, for example in the study of the spread of epidemics ( Lashari and Trapman, 2018 ; Fyles et al, 2021 ). An example of a study of stochasticity in infection was produced by Wood et al (2014) , where the heterogeneity in susceptibility to F. tularensis infection of individual phagocytic hosts within a population was investigated ( Wood et al, 2014 ).…”