[Paper Review] Social Dilemma of Non-Pharmaceutical Interventions

Alina Glaubitz, Feng Fu|arXiv (Cornell University)|Apr 11, 2024

Zoonotic diseases and public healthMedicine3 citations

TL;DR

This paper models the dynamic, context-dependent adoption of non-pharmaceutical interventions (NPIs) like face covering and social distancing using evolutionary game theory with bounded rationality. It reveals that oscillatory infection patterns emerge due to behavioral feedback loops, and that less strict NPIs like face covering can reduce total infections compared to perfect compliance, especially when effectiveness and cost perceptions are balanced.

ABSTRACT

In fighting infectious diseases posing a global health threat, ranging from influenza to Zika, non-pharmaceutical interventions (NPI), such as social distancing and face covering, remain mitigation measures public health can resort to. However, the success of NPI lies in sufficiently high levels of collective compliance, otherwise giving rise to waves of infection incidences that are not only driven by pathogen evolution but also changing vigilance in the population. Here we show that compliance with each NPI measure can be highly dynamic and context-dependent during an ongoing epidemic, where individuals may prefer one to another or even do nothing, leading to intricate temporal switching behavior of NPI adoptions. By characterizing dynamic regimes through the perceived costs of NPI measures and their effectiveness in particular regarding face covering and social distancing, our work offers new insights into overcoming barriers in NPI adoptions.

Motivation & Objective

• To understand how individual behavioral decisions on NPI compliance dynamically shift during an epidemic, especially between social distancing and face covering.
• To model the social dilemma in NPI adoption where individual incentives conflict with collective health outcomes.
• To analyze how perceived costs and effectiveness of NPIs shape infection trajectories and herd immunity thresholds.
• To compare outcomes under bounded rationality versus perfect compliance, focusing on total infection burden.

Proposed method

• Uses a modified SIR model (SIR-SD-FC) incorporating social distancing (SD) and face covering (FC) as behavioral strategies.
• Applies evolutionary game theory with replicator dynamics to model strategic switching between NPIs based on perceived costs and benefits.
• Introduces bounded rationality and loss aversion to simulate realistic human decision-making under uncertainty.
• Models the Swiss cheese approach by combining multiple imperfect NPIs to achieve layered protection.
• Employs numerical simulations to analyze total infection fractions $R(\infty)$ across parameter space for cost, effectiveness, and recovery rate.
• Compares outcomes under bounded rationality to perfect compliance to isolate behavioral effects on epidemic control.

Experimental results

Research questions

• RQ1 How does the dynamic, context-dependent switching between social distancing and face covering affect the total number of infections over time?
• RQ2 What role do perceived costs and effectiveness of NPIs play in shaping oscillatory infection patterns during an epidemic?
• RQ3 Why does bounded rationality lead to lower total infections than perfect compliance in some parameter regimes?
• RQ4 Under what conditions does face covering alone induce a bifurcation in infection dynamics?
• RQ5 How do herd immunity thresholds shift under bounded rationality compared to optimal behavior?

Key findings

• The SIR-SD-FC model with bounded rationality exhibits oscillatory infection dynamics due to feedback between behavior and transmission, unlike the smooth decline in perfect compliance models.
• Total infections $R(\infty)$ are often lower under bounded rationality than under perfect compliance, due to earlier achievement of herd immunity at lower $I_{\text{HI}}$.
• Face covering introduces a jump discontinuity in $R(\infty)$ when it becomes ineffective, indicating a sharp transition in population behavior and infection burden.
• For ineffective or expensive face covering, $R(\infty)$ increases to the SIR baseline, showing that poorly perceived NPIs fail to reduce transmission.
• When face covering is effective and low-cost, it reduces $I_{\text{HI}}$ and $R(\infty)$, especially when combined with social distancing in a Swiss cheese protection model.
• The model reveals that oscillations in $R(\infty)$ arise from a trade-off: reducing perceived costs lowers waves initially but can trigger new waves once thresholds are crossed, increasing $I_{\text{HI}}$.