Causal inference in food safety: Methods, applications, and future prospects

Abstract

As global food supply chains become increasingly complex, food safety risks have become more difficult to predict and control. Traditional correlation-based analytical methods are inadequate for identifying causal relationships in complex systems, leading to increased uncertainty in food safety risk assessments. To reduce this uncertainty, causal inference methods offer a means to unravel the intricate causal mechanisms underlying food safety, playing a crucial role in tracing the causal chains from genotype to phenotype in foodborne pathogens, and ultimately, to the associated food safety risks. This paper reviews the application of causal inference in food safety, discussing causal inference in genetic data and causal relationship identification in risk analysis. Additionally, it provides an overview of systematic causal reasoning methods based on causal Directed Acyclic Graphs (cDAGs) and the role of causal artificial intelligence (AI) in food safety. Despite the promise that causal inference holds for food safety research, challenges remain, including confounding factors, the limitations of randomized controlled trials, and issues with reverse causality. The further development and application of causal inference methods will provide more robust tools for food safety research, advancing methodologies, applications, and future prospects in this field.