Migratory birds, the H5N1 influenza virus and the scientific method

Area of Science:

• Epidemiology
• Veterinary Science
• Statistical Inference

Background:

• The role of migratory birds and poultry trade in highly pathogenic H5N1 (avian influenza) dispersal remains controversial.
• A recent contribution suggested strict adherence to the scientific method could resolve this debate.

Purpose of the Study:

• To critically evaluate the application of the scientific method, specifically null hypothesis testing, in understanding H5N1 dispersal.
• To propose a more comprehensive approach to scientific inquiry in epidemiology.

Main Methods:

• Critique of null hypothesis testing as the sole scientific method.
• Advocacy for incorporating hypothesis discovery alongside justification.
• Comparison of null hypothesis testing with Bayesian statistical inference methods.

Main Results:

• Null hypothesis testing is identified as a potentially misleading and weak approach for complex epidemiological questions.
• Bayesian methods are presented as a preferable alternative for statistical inference in H5N1 dispersal studies.
• The analogy between poultry transport and bird migration for disease spread is criticized.

Conclusions:

• Over-reliance on null hypothesis testing provides unrealistic expectations for definitive answers and policy guidance.
• A broader understanding of the scientific method, encompassing hypothesis generation and advanced statistical techniques, is crucial for addressing H5N1 dispersal.
• More nuanced statistical approaches are needed to accurately assess the contributions of different factors, like migratory birds and poultry trade, to H5N1 spread.