Big Data Research in Pediatric Neurosurgery: Content, Statistical Output, and Bibliometric Analysis

Area of Science:

• Pediatric Neurosurgery
• Health Informatics
• Biostatistics

Background:

• The use of large-scale datasets, or "big data," in pediatric neurosurgery research has grown significantly since 2000.
• Analysis of these publications is crucial to understand their statistical rigor and clinical impact.

Purpose of the Study:

• To characterize pediatric "big data" publications in neurosurgery from 2000 onwards.
• To evaluate the statistical output and clinical implications of these studies.
• To assess the bibliometric impact of this body of research.

Main Methods:

• A systematic search of four major North American neurosurgical journals was conducted for articles using non-neurosurgery-specific databases.
• Articles were analyzed for descriptive statistics, effect sizes (ESs), confidence intervals (CIs), and p-values.
• Bibliometric analysis using citation metrics was performed.

Main Results:

• 74 articles met the criteria, representing 1.7% of pediatric neurosurgery publications, with a notable increase after 2013.
• Healthcare Cost and Utilization Project (HCUP) databases were most common, with hydrocephalus as a frequent topic.
• Most studies reported small effect sizes (55%), high confidence interval spread (60%), and very strong p-values (50%). Only 20% of reported associations combined medium-to-large ESs, medium-to-high CIs, and strong p-values.
• Citations were concentrated, with a median of 8.5 per article and a few highly cited studies.

Conclusions:

• Big data utilization in pediatric neurosurgery literature has surged dramatically.
• A significant portion of reported associations may have limited clinical relevance for practitioners.
• Research impact is concentrated in a small number of highly cited publications.