Introduction: Healthcare reform, quality improvement, value of care, future physician reimbursement policies, and personalized patient care depend on robust data streams of demographical, clinical, and financial outcomes after neurosurgery. The neurosurgical literature lacks a comprehensive review of large national databases. In this study, we assessed the strengths and limitations of various resources for outcomes research in neurosurgery.
Methods: We conducted a systematic search of the PubMed and Web of Science databases for studies that used large, national datasets for analysis. We compiled a list of each dataset currently in use, and then supplemented the list with additional datasets that have not yet been used for neurosurgical research. 18 databases were identified in total. Variables collected included length of follow-up (30-day outcomes, longitudinal), number of records, availability of financial data, number of total citations to date, number of neurosurgical citations, and method of access.
Results: The number of unique patients contained within each dataset ranged from 7300 (N2QOD – National Neurosurgical Quality and Outcomes Database) to 180 million (MarketScan). The SEER (6063 overall citations, 72 neurosurgery citations) and NIS (2254 overall citations, 124 neurosurgery citations) databases were most frequently used for outcomes research in neurosurgery. By year of oldest citation in neurosurgery, SEER (1981) is the oldest and N2QOD (2013) is the newest. In the pediatric neurosurgery literature, KIDS (401 overall citations, 29 neurosurgery citations) is the most frequently used. The Pediatric Health Information System database (PHIS, 322 overall citations, 3 neurosurgery citations) is relatively underutilized by neurosurgeons. The method of access varied from free access for reporting institutions (PHIS) to application and small financial fee ($350 per year for 2007-2013 NIS) to substantial fees ($51,000 for multi-study access to 5 years of data in MarketScan).
Conclusions: Multiple options exist for neurosurgical outcomes research with varying lengths of follow-up, data completeness, clinical relevance to neurosurgery, and prior research utilization. With the ongoing trend of building neurosurgery-specific registries like N2QOD, large national databases will be a central tool in the future development of neurosurgery outcomes research.
Patient Care: To date, there has been no comprehensive review of national surgery outcomes databases. Here we discuss strengths and limitations of national surgical outcomes databases and provide a detailed guide for clinicians to find and use appropriate databases for targeted questions. The combination of the volume of data contained within national outcomes databases with big data techniques such as machine learning offers the opportunity to create highly predictive models for personalized patient care, value based reform, and quality improvement.
Learning Objectives: 1) Understand key strengths and limitations of major national outcomes databases
2) Recognize potential of national outcomes databases for healthcare reform, quality improvement, future reimbursement policies, and personalized patient care
3) Discuss future applications of big data and machine learning to neurosurgical outcomes research