The American College of Surgeon’s National Trauma Data Bank (NTDB) is the world’s largest central aggregation of injury data, submitted by individual hospitals. The type of data ranges from EMS through the continuum of care to patient discharge information. Hospitals reporting data to the NTDB are expected to make efforts to collect accurate and complete data elements. Data consistency in the NTDB is of vital importance but often overlooked and under-analyzed in research efforts. Despite its all-encompassing nature, a review of the medical literature raises concern regarding a lack of relevant clinical physiological data in the NTDB. The missing data could be a limitation because these data are essential components in generating risk adjustment strategies and are needed to further analyze the outcomes from trauma. The lack of critical physiological data may make it challenging to draw valid conclusions, address the comparison between cases, or adopt methods for future improvement. Moreover, an analysis of a select group of registrars who input data in the NTDB found that there is great variability in their reporting, which may present bias.1 This variability influences the consistency of the data in the NTDB. Another study evaluated the effects of actually excluding patients with missing data in the NTDB. The study found that excluding patients with missing data leads to researcherbiased estimates.2 This difficult condition demonstrates that the patients with missing data are of importance to any research findings, but the missing data itself create complications with the validity of findings. Moore et al. looked at multiple imputation methods to measure missing physiologic data, but this leads to unnecessary inaccuracies compared with simply recording the actual correct information.3, 4 The objective of this review was to evaluate the consistency of the simple key data elements of initial hospital and EMS vital signs reported to the NTDB. The consistency of these data in the NTDB is important and of concern because medical researchers are increasingly using the NTDB to improve trauma prevention and clinical care efforts both in the field and on initial arrival to the hospital. To contend with this, the American College of Surgeons has made diligent efforts to increase the validity of the data, including the implementation of business rules to decrease the amount of reporting of null values (i.e., not applicable (NA), not known (NK), and not recorded (NR)). In this review, the consistency of the simple key data elements of initial hospital and EMS vital signs reported to the NTDB was evaluated. The NTDB’s National Sample Program (NSP) and Research Data Set (RDS) were reviewed for initial hospital and EMS vital signs and data consistency for 2013 and 2016. The amount of NA and NK/NR data for initial hospital and EMS vital signs was collected and compared between 2013 and 2016. The data elements vital sign type and frequency tables were used to review simple uncollected data based on reporting of NA or NK/NR. The NSP and RDS held patient data for 172,387 and 968,665 patients, respectively. Initial hospital vital signs (blood pressure, pulse, respiratory rate, oxygen saturation, supplemental oxygen, temperature, and Glasgow Coma Scale) were reviewed for data consistency. The NSP missing data ranged from 2.2 to 12.9 per cent (reported as NA or NK/NR) for initial hospital vital signs. The RDS missing data ranged from 2.4 to 10.6 per cent (reported as NA or NK/NR). For the EMS vital signs, missing data ranged from 34.7 to 49.3 per cent in the NSP and 24.1 to 33.9 per cent in the RDS. In comparison with the 2013 EMS data, the rate of NK/NR in the NTDB for EMS in 2016 has decreased. This could be attributed to the implementation of the business rules suggested by the American College of Surgeons. The decrease in the amount of null reporting also suggests that the EMS system is improving and that hospitals are being more conscientious and meticulous with data documentation submitted to the NTDB. Address correspondence and reprint requests to Adel Elkbuli, M.D., M.P.H., Department of Surgery, Kendall Regional Medical Center, 11750 Bird Road, Miami, FL 33175. E-mail: Adel. Elkbuli@hcahealthcare.com.
[1]
Kevin A. Captain,et al.
Variability in interhospital trauma data coding and scoring: A challenge to the accuracy of aggregated trauma registries
,
2015,
The journal of trauma and acute care surgery.
[2]
Lynne Moore,et al.
A multiple imputation model for imputing missing physiologic data in the national trauma data bank.
,
2009,
Journal of the American College of Surgeons.
[3]
Lynne Moore,et al.
Evaluating the validity of multiple imputation for missing physiological data in the national trauma data bank
,
2009,
Journal of emergencies, trauma, and shock.
[4]
R Caetano,et al.
Clustered and missing data in the US National Trauma Data Bank: implications for analysis
,
2008,
Injury Prevention.