Reporting and interpreting statistical results in veterinary medicine: Calling for change

Area of Science:

• Veterinary statistics
• Evidence-based veterinary medicine

Background:

• Accurate interpretation of statistical results is crucial for evidence-based veterinary medicine.
• Current veterinary publications often rely on P values (<0.05) for significance, overlooking limitations and alternatives.
• There is a need to improve statistical reporting and interpretation in veterinary research.

Purpose of the Study:

• To address misconceptions about P values in veterinary medicine.
• To advocate for reporting effect estimates with confidence intervals.
• To provide practical guidance for improved statistical reporting and interpretation.

Main Methods:

• Discussion of P value limitations using a hypothetical randomized controlled trial.
• Introduction of effect estimates and confidence intervals as alternatives.
• Explanation of confidence interval functions and their utility as visual tools.

Main Results:

• Effect estimates with confidence intervals enable better assessment of clinical relevance and uncertainty.
• Confidence intervals provide a more comprehensive understanding of statistical findings.
• The proposed change in reporting and interpretation is feasible and necessary.

Conclusions:

• Shifting from P values to effect estimates and confidence intervals is essential for veterinary medicine.
• This change will enhance clinical decision-making by providing clearer interpretation of research findings.
• Improved statistical reporting promotes more robust evidence-based veterinary practices.