Sample Size Estimation in Veterinary Epidemiologic Research.

Area of Science:

• Veterinary Epidemiology
• Biostatistics
• Study Design

Background:

• Sample size calculations are essential for epidemiological studies to ensure objectives are met and resources are used efficiently.
• Grant applications increasingly require justification for sample sizes, including methodologies and assumptions.
• Inadequate sample sizes risk failing to detect true differences, while excessive numbers waste resources.

Purpose of the Study:

• To outline the importance of sample size calculations in veterinary epidemiology.
• To present methods for calculating sample sizes when data are not independent, a common issue in veterinary research.
• To introduce simulation-based methods as a flexible approach for complex study designs.

Main Methods:

• Discusses the inappropriateness of standard sample size formulae that assume data independence for veterinary studies with hierarchical data.
• Presents two methods to address data non-independence: design effect inflation and simulation-based approaches.
• Details the methodological approach for simulation and provides a worked example.

Main Results:

• Highlights the need for specialized sample size calculation methods in veterinary epidemiology due to data aggregation.
• Demonstrates the utility of design effects and simulation methods for handling non-independent data.
• Emphasizes simulation methods' advantage in accommodating complex study designs where formula-based solutions are unavailable.

Conclusions:

• Appropriate sample size calculation is critical for the validity and efficiency of veterinary epidemiological studies.
• Simulation-based methods offer a powerful and adaptable tool for determining sample sizes in complex veterinary research scenarios.
• The presented methods aid researchers in obtaining accurate sample size estimates, enhancing study robustness.