Quantifying and Realizing the Benefits of Targeting for Pandemic Response

Abstract

To respond to pandemics such as COVID-19, policy makers have relied on interventions that target specific population groups or activities. Such targeting is potentially contentious, so rigorously quantifying its benefits and downsides is critical for designing effective and equitable pandemic control policies. We propose a flexible modeling framework and a set of associated algorithms that compute optimally targeted, time-dependent interventions that coordinate across two dimensions of heterogeneity: population group characteristics and the specific activities that individuals engage in during the normal course of a day. We showcase a complete implementation in a case study focused on the Île-de-France region of France, based on commonly available hospitalization, community mobility, social contacts and economic data. We find that optimized dual-targeted policies have a simple and explainable structure, imposing less confinement on group-activity pairs that generate a relatively high economic value prorated by activity-specific social contacts. When compared to confinements based on uniform or less granular targeting, dual-targeted policies generate substantial complementarities that lead to Pareto improvements, reducing the number of deaths and the economic losses overall and reducing the time in confinement foreach population group. Since dual-targeted policies could lead to increased discrepancies in the confinements faced by distinct groups, we also quantify the impact of requirements that explicitly limit such disparities, and find that satisfactory intermediate trade-offs may be achievable through limited targeting.

Pandemic lockdown, isolation, and exit policies based on machine learning predictions

“…( 2020 ) and Camelo et al. ( 2021 ) start with SEIR and, respectively, split infected individuals into clinical/subclinical and confirmed/unconfirmed through testing, the former being further divided based on symptoms. These approaches are similar to ours in spirit but suggest that eventually symptomatic individuals may initially be asymptomatic, which necessitates adding the “A”‐type compartments.…”

“…( 2020 ) study spatial targeting and Camelo et al. ( 2021 ) study dual targeting based on risk groups and their activities. Such studies are complementary to ours.…”

Optimal targeting of interventions uses estimated risk of infectiousness to control a pandemic with minimal collateral damage

“…The data is split using a risk-scoring algorithm that divides each notified contact into one of 14 bins based on duration and proximity of exposure. The risk of infection is measured by recording the fraction of contacts within each bin that report a positive test after being notified 5 . This yields a piecewise constant (not-yet normalized) risk function, f C1 (q) where q now represents quantile within contacts instead of within the population.…”

“…Several studies have investigated the use of optimal control to most effectively contain pandemics [1, 2, 3, 4], but they focused on the optimal timecourse of population-wide interventions, where all options are costly. For this reason, Camelo et al analyzed the approach of targeting social distancing based on age and activity, with the goal of reducing the chance of infection amongst those most likely to experience severe disease [5].…”

Pandemic Lock-down, Isolation, and Exit Policies Based on Machine Learning Predictions

“…For instance, Kaplan (2020) considers the standard SIR model and Acemoglu et al (2020) consider connected replicas of SIR, while both briefly mention SEIR. Birge et al (2020), Camelo et al (2021) start with SEIR and, respectively, split infected individuals into clinical/sub-clinical and confirmed/unconfirmed through testing, the former further divided based on symptoms. These approaches are similar to ours in spirit but suggest that eventually symptomatic individuals may initially be asymptomatic, which necessitates adding the "A"-type compartments.…”