Isolated Communities and the Effectiveness of Masks in Controlling Epidemic Dynamics: An Agent-Based Modeling Study
TL;DR
Within the assumptions and intervention configurations examined, the hybrid strategy provided the strongest epidemic control; however, it should be interpreted as a full-compliance counterfactual rather than as a reconstruction of observed behavior.
Abstract
This study presents an agent-based model (ABM), implemented in NetLogo, for evaluating face-mask use and mobility restrictions in an isolated mining company town in the Brazilian Amazon. Unlike aggregate formulations, the model resolves individual agents, routine-based mobility, work shifts, household organization, and location-specific interactions within eight micro-environments. Across the configurations evaluated, the no-intervention scenario reached a peak of 793 active infections (approximately 40% of the modeled population), whereas full N95-mask adherence reduced the peak to 73 cases. The idealized hybrid scenario, combining full mask adherence and full mobility-restriction adherence, produced the lowest peak (48 cases), a growth rate of 3.19 infections/day, and a flattening duration of 30 days. Quantitative validation against the first epidemic wave in Porto Trombetas showed that the critical scenario most closely reproduced the empirical curve shape (r = 0.94; NRMSE = 22.2%), although it overestimated the empirical peak by 26.3%. Within the assumptions and intervention configurations examined, the hybrid strategy provided the strongest epidemic control; however, it should be interpreted as a full-compliance counterfactual rather than as a reconstruction of observed behavior. The proposed framework is intended for reproducible comparative scenario analysis and decision support in semi-closed populations, not for deterministic epidemic forecasting.