The Lancet's statistical review process: areas for improvement by authors

Area of Science:

• Medical research methodology
• Statistical analysis in publications

Background:

• The Lancet implemented statistical review for manuscripts post-conventional review.
• This review aimed to identify common statistical and methodological shortcomings in submitted papers.

Purpose of the Study:

• To summarize criticisms from statistical reviewers on submitted manuscripts.
• To assess the quality of statistical methods, reporting, and inference in medical publications.

Main Methods:

• Analysis of statistical review feedback for 191 papers submitted to The Lancet between November 1990 and June 1991.
• Categorization of papers based on reviewer recommendations (acceptable, revision needed, rejection).

Main Results:

• Only 54% of papers were acceptable or acceptable after revision; 14% were recommended for rejection.
• Method and results descriptions were inadequate in ~50% of papers; abstracts and conclusions were inadequate in ~25%.
• Major inference errors, linked to analysis or design flaws, occurred in 48 papers.

Conclusions:

• A significant proportion of submitted medical papers exhibit statistical and methodological deficiencies.
• Authors need to improve the justification of conclusions based on study design and statistical analysis.
• Statistical review is crucial for ensuring the validity and reliability of published research.