A novel, scenario-based approach to comparing non-pharmaceutical intervention strategies across nations

Abstract

Comparing COVID-19 non-pharmaceutical intervention (NPI) strategies across nations is a key step in preparing for future pandemics. Conventional comparisons, which rank individual NPI effects, are limited by: 1) vastly different political, economic, and social conditions among nations, 2) NPIs typically being applied as packages of interventions, and 3) an exclusive focus on epidemiological outcomes of interventions. Here, we develop a coupled epidemiological-behavioural-macroeconomic model that allows us to transfer NPI strategies from a reference nation to a focal nation while preserving the packaged nature of NPIs, controlling for differences among nations, and quantifying epidemiological, behavioural and economic outcomes. As a demonstration, we take Germany as our focal nation during Spring 2020, and New Zealand and Switzerland as reference nations with contrasting NPI strategies. We show that, while New Zealand's more aggressive strategy would have yielded modest epidemiological gains in Germany, it would have resulted in substantially higher economic costs while dramatically reducing social contacts. In contrast, Switzerland's more lenient NPI strategy would have prolonged the first wave in Germany, but would have also have increased relative costs. Our results demonstrate that Germany's intermediate strategy was effective in quelling the first wave while mitigating both economic and social costs.

Autonomous and policy-induced behavior change during the COVID-19 pandemic: Towards understanding and modeling the interplay of behavioral adaptation

“…Notwithstanding, particularly autonomous adaptation processes should receive more attention in infectious disease models, as others have argued [149]. Based on our literature and empirical analysis, we discussed relevant avenues in behavioral-epidemiological modeling: Emerging approaches combine endogenous autonomous adaptation with representations of policy impacts, thus including both autonomous and policy-induced adaptation [135, 137]. Future work should increasingly focus on their interplay and why parameters vary over time.…”

“…Thus, it is promising that an increasing number of models includes the effects of both autonomous and policy-induced behavioral adaptation [11, 16, 135–137]. Here, the key question is how the two effects are set in relation to one another.…”

“…In [135], for example, NPIs set the boundary conditions for the number of possible contacts, which then vary in dependence of ICU occupancy. [137] characterize their relationship as additive and estimate the parameters for both effects concurrently from data. In their approach, identifiability is related to the dynamics of incidence number and change points of policy response.…”

“…As we expound further below, however, it is of key interest whether a change in behavior is the result of policy, of autonomous adaptation or both. Thus, it is promising that an increasing number of models includes the effects of both autonomous and policy-induced behavioral adaptation [11,16,[135][136][137]. Here, the key question is how the two effects are set in relation to one another.…”