A systematic review of mechanistic models used to study avian influenza virus transmission and control

TL;DR

This systematic review analyzes 45 studies on mechanistic models of avian influenza transmission, highlighting their types, estimated parameters, and control strategies, emphasizing the importance of early intervention and tailored approaches.

Contribution

It provides a comprehensive synthesis of existing mechanistic models for avian influenza, identifying gaps and suggesting improvements for better epidemic control.

Findings

Population-based models are most common

Transmission rate varies widely across settings

Early intervention improves control effectiveness

Abstract

The global spread of avian influenza A viruses in domestic birds is causing dramatic economic and social losses. Various mechanistic models have been developed in an attempt to better understand avian influenza transmission and to evaluate the effectiveness of control measures. However, no comprehensive review of the mechanistic approaches used currently exists. To help fill this gap, we conducted a systematic review of mechanistic models applied to real-world epidemics to (1) describe the type of models and their epidemiological context, (2) synthetise estimated values of AIV transmission parameters and (3) review the control strategies most frequently evaluated and their outcome. Fourty-five articles qualified for inclusion, that fitted the model to data and estimated parameter values (n = 42) and/or evaluated the effectiveness of control strategies (n = 21). The majority were population-based models (n = 26), followed by individual-based models (n = 15) and a few metapopulation models (n = 4). Estimated values for the transmission rate varied substantially according to epidemiological settings, virus subtypes and epidemiological units. Other parameters such as the durations of the latent and infectious periods were more frequently assumed, limiting the insights brought by mechanistic models on these. Concerning control strategies, many models evaluated culling (n = 15), while vaccination received less attention (n = 7). According to the reviewed articles, optimal control strategies varied between virus subtypes and local conditions, and also depended on the objective. For instance, vaccination was optimal when the objective was to limit the overall number of culled flocks, while pre-emptive culling was preferred for reducing the epidemic size and duration. Earlier implementation of interventions consistently improved the efficacy of control strategies, highlighting the need for effective surveillance and epidemic preparedness. Potential improvements of mechanistic models include explicitly accounting for various transmission routes, and distinguishing poultry populations according to species and farm type. To provide insights to policy makers in a timely manner, aspects about the evaluation of control strategies that could deserve further attention include: economic evaluation, combination of strategies including vaccination, the use of optimization algorithm instead of comparing a limited set of scenarios, and real-time evaluation.