Statistical Competence and Editorial Policies in Biomedical Publications

Area of Science:

• Biomedical Research
• Statistical Methodology
• Scientific Integrity

Background:

• Widespread misuse of statistical methods in biomedical research compromises scientific integrity and public health.
• Statistical errors stem from training gaps across the research lifecycle, from design to reporting.
• Limitations in peer-review processes are evident from increasing retractions and continued citation of flawed studies.

Purpose of the Study:

• To highlight the prevalence and impact of statistical errors in biomedical research.
• To identify key areas for improvement in statistical practices and reporting.
• To propose structural interventions for enhancing the quality of biomedical literature.

Main Methods:

• Review of common statistical errors in study design, analysis, and interpretation.
• Discussion of the role of statistical software (SPSS, R/RStudio) and artificial intelligence (AI) in statistical workflows.
• Analysis of current peer-review limitations and publisher-level interventions.

Main Results:

• Statistical errors affect clinical guidelines, treatment efficacy, and research reproducibility.
• Software choice impacts analytical outcomes; full disclosure of versions and procedures is vital.
• AI tools in statistics pose risks like hallucination and lack of transparency, requiring validation guidelines.

Conclusions:

• Enhancing statistical quality requires addressing training gaps and improving reporting transparency.
• Publisher-level interventions like statistical editors and open data policies are effective.
• Coordinated efforts from all stakeholders are essential for improving the statistical rigor of biomedical research.