Impact of Inadequate Methods and Data Analysis on Reproducibility

Area of Science:

• Oncology
• Hematology
• Biomedical Research

Background:

• Erythropoiesis-stimulating agents (ESAs) treat anemia in cancer patients.
• Concerns exist that ESAs may stimulate Epo receptors on tumor cells (Epo receptor-cancer hypothesis).
• Conflicting data in published articles necessitates quality assessment.

Purpose of the Study:

• To investigate the relationship between research quality and reproducibility in Epo receptor-cancer articles.
• To identify specific quality deficits contributing to conflicting findings.

Main Methods:

• Conducted comprehensive literature searches to identify relevant articles.
• Analyzed 280 identified articles for quality parameters and reproducibility issues.
• Assessed incidence of faults in controls, reagent/method validation, and statistical analysis.

Main Results:

• Identified numerous conflicts within and between articles.
• High incidence of quality faults: 90% lacked adequate controls, 87% had inadequate validation, 84% used improper statistics, 81% had questionable data interpretation.
• Observed false-positive/negative data varying with reagents and methods.

Conclusions:

• Low quality of evidence in Epo receptor-cancer articles likely explains poor reproducibility.
• Deficits in research methodology and statistical analysis are significant contributors to conflicting results.
• Improving research quality is essential for reliable scientific conclusions in this field.