Big Data Research in Neurosurgery

A Critical Look at this Popular New Study Design

Chesney S. Oravec, Mustafa Motiwala, Kevin Reed, Douglas Kondziolka, Fred G. Barker, Lattimore Michael, Paul Klimo

Research output: Contribution to journalArticle

Abstract

The use of "big data" in neurosurgical research has become increasingly popular. However, using this type of data comes with limitations. This study aimed to shed light on this new approach to clinical research. We compiled a list of commonly used databases that were not specifically created to study neurosurgical procedures, conditions, or diseases. Three North American journals were manually searched for articles published since 2000 utilizing these and other non-neurosurgery-specific databases. A number of data points per article were collected, tallied, and analyzed.A total of 324 articles were identified since 2000 with an exponential increase since 2011 (257/324, 79%). The Journal of Neurosurgery Publishing Group published the greatest total number (n = 200). The National Inpatient Sample was the most commonly used database (n = 136). The average study size was 114 841 subjects (range, 30-4 146 777). The most prevalent topics were vascular (n = 77) and neuro-oncology (n = 66). When categorizing study objective (recognizing that many papers reported more than 1 type of study objective), "Outcomes" was the most common (n = 154). The top 10 institutions by primary or senior author accounted for 45%-50% of all publications. Harvard Medical School was the top institution, using this research technique with 59 representations (31 by primary author and 28 by senior).The increasing use of data from non-neurosurgery-specific databases presents a unique challenge to the interpretation and application of the study conclusions. The limitations of these studies must be more strongly considered in designing and interpreting these studies.

Original languageEnglish (US)
Pages (from-to)728-746
Number of pages19
JournalNeurosurgery
Volume82
Issue number5
DOIs
StatePublished - May 1 2018

Fingerprint

Neurosurgery
Databases
Research
Neurosurgical Procedures
Medical Schools
Blood Vessels
Publications
Inpatients
Research Design
Outcome Assessment (Health Care)

All Science Journal Classification (ASJC) codes

  • Surgery
  • Clinical Neurology

Cite this

Big Data Research in Neurosurgery : A Critical Look at this Popular New Study Design. / Oravec, Chesney S.; Motiwala, Mustafa; Reed, Kevin; Kondziolka, Douglas; Barker, Fred G.; Michael, Lattimore; Klimo, Paul.

In: Neurosurgery, Vol. 82, No. 5, 01.05.2018, p. 728-746.

Research output: Contribution to journalArticle

Oravec, Chesney S. ; Motiwala, Mustafa ; Reed, Kevin ; Kondziolka, Douglas ; Barker, Fred G. ; Michael, Lattimore ; Klimo, Paul. / Big Data Research in Neurosurgery : A Critical Look at this Popular New Study Design. In: Neurosurgery. 2018 ; Vol. 82, No. 5. pp. 728-746.
@article{dfe38f64a4a84e0abab715611e899122,
title = "Big Data Research in Neurosurgery: A Critical Look at this Popular New Study Design",
abstract = "The use of {"}big data{"} in neurosurgical research has become increasingly popular. However, using this type of data comes with limitations. This study aimed to shed light on this new approach to clinical research. We compiled a list of commonly used databases that were not specifically created to study neurosurgical procedures, conditions, or diseases. Three North American journals were manually searched for articles published since 2000 utilizing these and other non-neurosurgery-specific databases. A number of data points per article were collected, tallied, and analyzed.A total of 324 articles were identified since 2000 with an exponential increase since 2011 (257/324, 79{\%}). The Journal of Neurosurgery Publishing Group published the greatest total number (n = 200). The National Inpatient Sample was the most commonly used database (n = 136). The average study size was 114 841 subjects (range, 30-4 146 777). The most prevalent topics were vascular (n = 77) and neuro-oncology (n = 66). When categorizing study objective (recognizing that many papers reported more than 1 type of study objective), {"}Outcomes{"} was the most common (n = 154). The top 10 institutions by primary or senior author accounted for 45{\%}-50{\%} of all publications. Harvard Medical School was the top institution, using this research technique with 59 representations (31 by primary author and 28 by senior).The increasing use of data from non-neurosurgery-specific databases presents a unique challenge to the interpretation and application of the study conclusions. The limitations of these studies must be more strongly considered in designing and interpreting these studies.",
author = "Oravec, {Chesney S.} and Mustafa Motiwala and Kevin Reed and Douglas Kondziolka and Barker, {Fred G.} and Lattimore Michael and Paul Klimo",
year = "2018",
month = "5",
day = "1",
doi = "10.1093/neuros/nyx328",
language = "English (US)",
volume = "82",
pages = "728--746",
journal = "Neurosurgery",
issn = "0148-396X",
publisher = "Lippincott Williams and Wilkins",
number = "5",

}

TY - JOUR

T1 - Big Data Research in Neurosurgery

T2 - A Critical Look at this Popular New Study Design

AU - Oravec, Chesney S.

AU - Motiwala, Mustafa

AU - Reed, Kevin

AU - Kondziolka, Douglas

AU - Barker, Fred G.

AU - Michael, Lattimore

AU - Klimo, Paul

PY - 2018/5/1

Y1 - 2018/5/1

N2 - The use of "big data" in neurosurgical research has become increasingly popular. However, using this type of data comes with limitations. This study aimed to shed light on this new approach to clinical research. We compiled a list of commonly used databases that were not specifically created to study neurosurgical procedures, conditions, or diseases. Three North American journals were manually searched for articles published since 2000 utilizing these and other non-neurosurgery-specific databases. A number of data points per article were collected, tallied, and analyzed.A total of 324 articles were identified since 2000 with an exponential increase since 2011 (257/324, 79%). The Journal of Neurosurgery Publishing Group published the greatest total number (n = 200). The National Inpatient Sample was the most commonly used database (n = 136). The average study size was 114 841 subjects (range, 30-4 146 777). The most prevalent topics were vascular (n = 77) and neuro-oncology (n = 66). When categorizing study objective (recognizing that many papers reported more than 1 type of study objective), "Outcomes" was the most common (n = 154). The top 10 institutions by primary or senior author accounted for 45%-50% of all publications. Harvard Medical School was the top institution, using this research technique with 59 representations (31 by primary author and 28 by senior).The increasing use of data from non-neurosurgery-specific databases presents a unique challenge to the interpretation and application of the study conclusions. The limitations of these studies must be more strongly considered in designing and interpreting these studies.

AB - The use of "big data" in neurosurgical research has become increasingly popular. However, using this type of data comes with limitations. This study aimed to shed light on this new approach to clinical research. We compiled a list of commonly used databases that were not specifically created to study neurosurgical procedures, conditions, or diseases. Three North American journals were manually searched for articles published since 2000 utilizing these and other non-neurosurgery-specific databases. A number of data points per article were collected, tallied, and analyzed.A total of 324 articles were identified since 2000 with an exponential increase since 2011 (257/324, 79%). The Journal of Neurosurgery Publishing Group published the greatest total number (n = 200). The National Inpatient Sample was the most commonly used database (n = 136). The average study size was 114 841 subjects (range, 30-4 146 777). The most prevalent topics were vascular (n = 77) and neuro-oncology (n = 66). When categorizing study objective (recognizing that many papers reported more than 1 type of study objective), "Outcomes" was the most common (n = 154). The top 10 institutions by primary or senior author accounted for 45%-50% of all publications. Harvard Medical School was the top institution, using this research technique with 59 representations (31 by primary author and 28 by senior).The increasing use of data from non-neurosurgery-specific databases presents a unique challenge to the interpretation and application of the study conclusions. The limitations of these studies must be more strongly considered in designing and interpreting these studies.

UR - http://www.scopus.com/inward/record.url?scp=85048424507&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048424507&partnerID=8YFLogxK

U2 - 10.1093/neuros/nyx328

DO - 10.1093/neuros/nyx328

M3 - Article

VL - 82

SP - 728

EP - 746

JO - Neurosurgery

JF - Neurosurgery

SN - 0148-396X

IS - 5

ER -