C. Yeung1, E. Chang1, K. Bressler1, H. Roberts1, J. Feliciano2, S. DiRusso1,2 1New York College Of Osteopathic Medicine, Old Westbury, NY, USA 2St. Barnabas Hospital Health System, Bronx, NY, USA
Introduction:
The National Trauma Data Bank (NTDB), created and maintained by the American College of Surgeons (ACS) is the largest aggregation of trauma patient data, sourced from participating member hospitals. However, the structure of the data is not easily accessible for analysis in standard statistical programs. A protocol was developed to prepare and organize the variables and tables of the NTDB for accurate and practical use in SPSS.
Methods:
Data: years 2021-2022. The major Trauma Data file (Participant_Use_Files: PUF_Trauma) contained all patient data besides Abbreviated Injury Scale (AIS) and International Classification of Diseases, Tenth Revision (ICD10) diagnosis and procedure codes, hospital events, and pre-existing conditions. Files containing AIS predots and severities, and Injury Severity Score regions were used to calculate New Injury Severity Scores (NISS) which was not in the original files. AIS predots and severities, and NISS were added to the major Trauma Data file. ICD10 diagnosis and procedure codes with start times were also added. Binary variables were created to indicate each hospital event and pre-existing condition description. Analysis of Hospital Discharge variables showed a large percentage of missing values and showed the majority of these were due to ED discharges (non admissions). These were eliminated. Dead on Arrival (DOAs) were defined and eliminated. A new binary variable was created to indicate mortality based on hospital discharge disposition and died in ED. Accuracy of variable reorganization and creation were checked using descriptive statistics in SPSS. Combination of data created a 3.44 GB master file. Python code was written to extract manageable data subsets of interest for use in SPSS.
Results:
2442053 trauma patients were obtained from NTDB, 1209097 from 2021 and 1232956 from 2022. 174799 (7.2%) patients were removed for missing sex and/or age as these were considered essential variables. 387607 (17.1%) patients were removed as they were not admitted to hospital as indicated by ED discharge disposition of home, other, left against medical advice, or transferred to another hospital. 12840 (0.7%) patients were removed for DOA defined as systolic blood pressure, pulse rate, respiratory rate, temperature, pulse oximetry = 0, total Glasgow Coma Scale = 3, and ED discharge disposition of deceased/expired. 1866807 (76.4%) trauma patients remained.
Conclusion:
A protocol optimizing NTDB for analytical use in SPSS was developed. Access to quality national databases is limited due to human input error/discrepancies and hardware restrictions and the complex structure of the database. However, large datasets such as NTDB are necessary for performing comprehensive research on diverse patient populations. Thus, data refinement protocols may give researchers the ability to access and analyze national datasets with more accurate and relevant results.