Pitfalls in scientific research: critical appraisal of articles published in one of the international journals in Egypt

• Main Authors: Sarah S. Nasr, Ghada M. Sherif, Amal S. Ibrahim, Rasha M. Allam
• Format: Article
• Language: English
• Published: SpringerOpen 2020-10-01
• Series: Journal of the Egyptian National Cancer Institute
• Subjects: Critical appraisal; Peer-review; Scientific articles; Medical manuscript review; Study design; Statistical methodology
• Online Access: http://link.springer.com/article/10.1186/s43046-020-00050-4

Abstract Background To identify and report flaws of Internet-published articles in the Journal of the Egyptian National Cancer Institute (JENCI), Cairo University, through a retrospective documentary study on articles published during the period from 2011 to 2016. All sections were reviewed against a collective checklist. Articles were grouped by publication year into 2 intervals: early (from 2011 to 2013) and recent (from 2014 to 2016) to identify changes in study characteristics over time. Results The study included 139 original articles. Half of the titles represented aim and 9.4% represented study design. Abstracts were concise, clear, with structured writing format in 98.6%, 93.5%, and 35.3%, respectively. Most introductions included the study aim, while 41% had a rationale. Study timing was reported in 59.0%, while the study design was reported in 25.9%. Inclusion and exclusion criteria were clearly reported in 43.1% and 40.1%, respectively. Statistical methods were mentioned in 80.6%, complete in 30.4%, and appropriate in 85.7%. Four studies reported sample size estimation. Only 52.5% and 58.3% of results were exhaustive and answer the research question, respectively. Incorrect statistical calculations occurred in 41.0%, inappropriate statistical tests or descriptive parameter selection in 26.6%, while inappropriate test application occurred in 49.1%. About 60% of discussions did not completely cover results, 31.7% fully justified the findings, 56.1% followed a logical flow, and 36.7% had contradiction within the text. Conclusions were mostly linked to aim, imprecise, and extrapolating beyond results. On comparing both periods, only a significant less misuse of statistical terms, more reporting conflict of interest, more missing references for cited texts in the recent period, and more participation of NCI over other institutes in the early period were found. Conclusion Articles published in JENCI (from 2011 to 2016) had many methodological and reporting defects and some points of strength. Using the collective checklist developed by this study, continuous training of researchers, involving epidemiologists throughout the whole research process, and applying strict journal reporting and publication rules should be encouraged.