Beyond the p-Value: Has Statistical Reporting Evolved in Surgical Infections?

Appropriate statistical methodology and transparent reporting are essential for the validity and reproducibility of research. Despite landmark guidance from the American Statistical Association (ASA), inadequate statistical reporting remains common in surgical literature. We examined whether the nature and quality of statistical reporting in the Surgical Infections journal changed between 2015 and 2023.

A retrospective review of original research articles published in Surgical Infections in 2015 compared to 2023 was performed. The methods and results sections of included articles were reviewed for: reporting of a priori significance level (α), power calculation performance and methodology, p-value reporting modalities, confidence interval use, and claims of statistical significance. Dichotomous variables were compared using Fisher exact test, continuous variables by Student's t-test, and variances by F-statistic.

A total of 102 (2015) and 86 (2023) articles met inclusion criteria. Power calculations were reported in only 11.7% of articles overall, with no significant change between years (p = 0.25). Reporting of α decreased from 79.4% to 67.4% (p = 0.07), with near-uniform convergence to p < 0.05 in 2023 (86% vs. 98%; p = 0.01). The number of unique p-value reporting modalities increased between years, with one 2023 article employing 15 distinct modalities. Confidence-interval reporting was unchanged (58.9% vs. 61.6%; p = 0.6). Claims of statistical significance decreased from 92.2% to 79.1% of articles (p = 0.01).

Statistical reporting in Surgical Infections showed little improvement from 2015 to 2023 despite evolving national guidance. Fewer authors claimed significance, and the journal demonstrated no requirement for significance for publication. We recommend standardized statistical-reporting guidelines.