Errata and Corrigenda in Neurosurgical Publications: an In-Depth Analysis and Inference.

OBJECTIVE: There has been an increasing number of reported errors in neurosurgical publications. Subsequent published correction details in the form of errata and corrigenda has not been analyzed previously. Our study aims to review the published errata and corrigenda in neurosurgical literature, and we discuss the characteristics and future implications of postpublication errors. METHODS: PubMed and Embase databases were screened using a designed search strategy for errata and corrigenda in neurosurgical articles published between 1990 and March 2021. Data including journal impact factor, number of authors and citations, country of origin, study design, level of evidence, category, severity, and timing of correction of errors were extracted for summary and analysis. RESULTS: A total of 768 included articles contributed to 918 error corrections. In 563 (73.31%) articles, the correction was acknowledged in the original record. Median journal impact factor was 3.114 (interquartile range [IQR], 2.139). Median correction time was 3 months (IQR, 5 months), with no statistically significant difference in timing of correction across different error severities (Kruskal-Wallis test, P = 0.058). A total of 398 (51.82%) studies showed level 3 evidence. Errors with minimal severity most commonly occurred in the author list 197(82.43%), with typographic error being the predominant cause. Errors with high severity most commonly occurred in the Results section. Eight errors (0.87%) prompted modification of study conclusions. CONCLUSIONS: Observations of postpublication corrections across a wide range of studies prompted more awareness of errors in the neurosurgical literature regardless of impact factors and level of evidence. More standardization in the recognition and acknowledgment of errors, with active engagements from authors, readers, editors, and publishers, is recommended.