View Full Size

• Download Figure
• Download figure as PowerPoint slide

Your editorial staff has been pondering for a couple of years how to improve our reporting of clinically relevant statistics with effect size and CIs and with less reliance on P values. I will take full ownership for taking so long to get to this. Like many, I was raised on ANOVAs and regression analysis and relied on P values for interpreting the results of my own studies and those of others. In veterinary medicine, most of our studies will be underpowered due to small sample sizes and we are therefore likely to miss clinically important effects of an intervention when the P values suggest the results are not significant.

One of my favorite examples comes from Sullivan and Feinn¹: “A commonly cited example of this problem is the Physicians Health Study of aspirin to prevent myocardial infarction (MI). In more than 22 000 subjects over an average of 5 years, aspirin was associated with a reduction in MI (although not in overall cardiovascular mortality) that was highly statistically significant: P < .00001. The study was terminated early due to the conclusive evidence, and aspirin was recommended for general prevention. However, the effect size was very small: a risk difference of 0.77% with r² = .001—an extremely small effect size. As a result of that study, many people were advised to take aspirin who would not experience benefit yet were also at risk for adverse effects. Further studies found even smaller effects, and the recommendation to use aspirin has since been modified.” Not to be insulting, but risk difference of 0.77% means that, compared to the control group, the aspirin group experienced a 0.77% lower risk of MI—< 1% in terms of clinical effect. In veterinary medicine, we run an equal or greater risk of missing a large clinical effect in a small study where the P value suggests it is not statistically significant, resulting in missed opportunities to enhance the health of patients.

What is the editorial staff doing to rectify this overreliance on P values? First, we ourselves opened our minds, got educated, and committed to improving statistical reporting in JAVMA and AJVR. This will be no small endeavor. We estimate that only 40% of manuscripts submitted have reported effect size. We attended webinars, took classes at universities, and had many meetings to discuss changing our guidelines for readers and authors. We moved into action in August 2024, when Dr. Eric Monnet’s team submitted an article to JAVMA resulting in a statistical reviewer encouraging movement away from reporting solely P values. Dr. Monnet entirely agreed with the reviewer, but like many of us, including myself, was unsure of how omitting P values would be received by readers. We had many valuable conversations and Dr. Monnet kindly agreed to help us get the ball rolling on encouraging clinical effect reporting. His article, “A mattress suture with pledgets can partially occlude intrahepatic shunts in 23 dogs: evaluation of long-term survival” by Drs. Margaret Chu, Sarah Marvel, and Eric Monnet,² is available in this issue.

Also in this issue, I encourage you to read the invited commentary³ to Dr. Monnet’s article by JAVMA Associate Editor Constance N. White, DVM, MPH, PhD. As you might glean from the title, “The perilous P value,” Dr. White takes a lighthearted approach to the somewhat dry subject matter of statistical reporting, using Dr. Monnet’s article as a framework to emphasize the importance of effect size reporting, and she provides several good references for the interested reader.

We have also updated our instructions for authors with guidelines and links to informative resources (https://avmajournals.avma.org/page/JAVMA-Original-Research). While there is no consensus on inclusion of P values in reporting, we are encouraging the inclusion of both P values and effect size to help readers understand the magnitude differences found (effect size) and if the findings are likely due to chance (P value). Statistics are not the forte of most clinicians and there are no perfect editorial procedures to ensure consistent reporting, so we will continue to monitor and modify our processes with valuable feedback from our authors and readers. A huge thank-you to Drs. Monnet and White for kicking off this important initiative!

As always, we welcome your comments and thoughts about JAVMA and AJVR.

Respectfully,

Lisa A. Fortier DVM, PhD, DACVS

Editor-in-Chief, JAVMA and AJVR

Chief Publications Officer, AVMA

lfortier@avma.org