A new web application for determining sample size in freedom-from-disease testing with imperfect tests.

Area of Science:

• Veterinary epidemiology
• Biostatistics
• Disease surveillance

Background:

• Veterinary surveillance requires precise sample size calculations for freedom-from-disease testing.
• Imperfect test sensitivity and specificity introduce challenges in study design.
• Existing methods often rely on approximations or simulations.

Purpose of the Study:

• To develop an exact algorithm for sample size determination in veterinary surveillance.
• To provide a practical, user-friendly tool for calculating sample sizes.
• To incorporate uncertainty in test sensitivity and specificity into sample size calculations.

Main Methods:

• Revisiting exact sample size calculations using binomial and hypergeometric distributions.
• Developing a new algorithm implemented as an R Shiny application with a graphical user interface.
• Modeling sensitivity and specificity as Beta-distributed variables with known hyperparameters.

Main Results:

• Exact sample size calculations are feasible and more accurate than approximations.
• The new algorithm provides a practical tool for determining sample sizes in real-world scenarios.
• The approach accounts for uncertainty in test performance characteristics.

Conclusions:

• The developed algorithm offers an improvement for sample size calculations in veterinary surveillance.
• The R Shiny application provides accessible and accurate tools for researchers and students.
• This work enhances the statistical rigor of freedom-from-disease testing in veterinary epidemiology.