M. A. Leon1, D. W. Lahm1, P. R. Armijo1, B. Pokala1, L. Flores1, D. Oleynikov1 1University of Nebraska Medical Center,Surgery,Omaha, NE, USA
Introduction:
Surgical outcomes research is integral to the evolution of evidence-based medicine and healthcare policy. Large national databases provide a new way to perform surgical outcomes research; however, there is vast variability between existing databases. The goal of this study was to summarize the contents and characteristics of current databases that can be utilized for Minimally Invasive Surgery (MIS) and Bariatric surgery outcomes research.
Methods:
A literature review was performed to identify 130 datasets used for healthcare research. The list was then refined to include only those databases tracking MIS/Bariatric surgical outcomes. These databases were further assessed by examining published database descriptions, analyzing data dictionaries, and reviewing existing studies that utilized databases of interest. The duration of patient follow-up and availability of longitudinal outcomes tracking was determined for each database. Data dictionaries for several private databases were unable to be obtained and were excluded. Ultimately, a total of 19 outcome research variables were created (Graphic 1), and the presence or absence of each was collected for all databases. A more granular analysis was performed to determine which data endpoints were available within each outcome variable.
Results:
Out of 130 databases, eight contained a majority of the variables of interest: The Americas Hernia Society Quality Collaborative (AHSQC), the Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program (MBSAQIP), Medicare, Vizient, the American College of Surgery National Surgical Quality Improvement Program (ACS NSQIP), the National Cancer Database (NCDB), the Healthcare Cost and Utilization Project National Inpatient Sample (HCUP NIS), and the Military Health System Data Repository (MDR).
The number of databases that contained variables within each category are as follows: patient identification (N=5), patient demographics (N=8), patient geographics (N=6), socioeconomic status (N=3), social history (N=4), height/weight/BMI (N=4), comorbidities (N=8), facility characteristics (N=4), diagnosis codes (N=6), preoperative medications (N=3), preoperative laboratory results (N=2), payer source (N=7), procedure codes (N=7), intraoperative details (N=4), surgeon identification (N=5), surgeon details (N=5), surgical outcomes (N=8), quality of life (N=1), and cost (N=4).
Conclusion:
The use of big data in healthcare research continues to expand. Large national databases offer increased power and external validity when utilized for surgical outcomes research. This study analyzed the contents of the most comprehensive healthcare databases and will serve as a reference tool for future studies in MIS/bariatrics outcomes research.
