A practical approach to calculate sample size for herd prevalence surveys

Area of Science:

• Veterinary epidemiology
• Diagnostic test evaluation

Background:

• Herd-level prevalence studies are essential for disease control.
• Diagnostic tests often have imperfect sensitivity and specificity, potentially biasing results.
• Accurate study design is critical for reliable disease estimation.

Purpose of the Study:

• To present an adapted formula for improved planning of herd-level prevalence studies using imperfect diagnostic tests.
• To demonstrate the impact of test imperfections on prevalence estimation and confidence ranges.
• To provide examples using bovine paratuberculosis to illustrate design trade-offs.

Main Methods:

• Development of an adapted formula to account for test sensitivity and specificity.
• Integration of the formula with existing software for practical application.
• Analysis of trade-offs between herd and animal sample sizes for achieving desired study precision.

Main Results:

• The adapted formula allows for more accurate planning of prevalence studies.
• Failure to account for test imperfections leads to biased prevalence estimates and underestimated confidence ranges.
• Examples illustrate the balance between sampling intensity (animals per herd) and herd inclusion.

Conclusions:

• Accounting for diagnostic test imperfections is vital for robust herd-level prevalence studies.
• The presented approach enhances study planning and improves the reliability of disease prevalence data.
• This method is applicable to various infectious diseases, including bovine paratuberculosis.