This JAMA Guide to Statistics and Methods summarizes the data compilation, structure, and methods of the National Trauma Data Bank (NTDB) for use in surgical research.
Trauma remains a leading cause of death and disability and accounts for a substantial portion of health care expenditures.1 Therefore, research to improve trauma care is a leading public health priority.
Spearheading this effort, the American College of Surgeons Committee on Trauma coordinated a landmark multi-institutional endeavor, the Major Trauma Outcomes Study.2 On its completion in 1989, the American College of Surgeons Committee on Trauma recognized the importance of national trauma data aggregation to inform quality improvement. In 1997, it formed a subcommittee to develop the National Trauma Data Bank (NTDB), a standardized collection of national trauma data (Box 11). Today, to our knowledge, the NTDB is the world’s largest trauma data repository, with more than 7.5 million electronic records from more than 900 trauma centers.
BOX 11 Attributes of the National Trauma Data Bank
The National Trauma Data Bank (NTDB) is the world’s largest trauma data repository with more than 7.5 million electronic records.
The NTDB is considered a convenience sample because of voluntary trauma data submission.
The large sample size of NTDB facilitates hypothesis-generating research and the study of rare injuries, procedures, and outcomes.
The NTDB contains prehospital and in-hospital injury data, including anatomic and physiologic severity.
The NTDB does not include data on costs, laboratory values, readmissions, or long-term outcomes.
Careful study design, sample selection, and analytics can help mitigate NTDB limitations resulting from missing data and selection/information bias.
DATA COMPILATION AND STRUCTURE
Annually, between February and May, the NTDB collects voluntarily submitted data from individual hospitals, concordant with the National Trauma Data Standard, a set of standardized data definitions.3 Inclusion/exclusion criteria are based on International Statistical Classification of Diseases, Tenth Revision, Clinical Modification diagnoses codes and certain admission characteristics.4 The NTDB is an incident-based record; each incident is recorded independently of repeated injuries. Validation rules mitigate the submission of missing or nonsensical data. A fully deidentified data set is reported annually in compliance with the Health Insurance Portability and Accountability Act, such that no patient can be identified.
Data quality in the NTDB has improved significantly since the adoption of the National Trauma Data Standard in 2007 and the implementation of American College of Surgeons Trauma Quality Improvement Program (ACS TQIP) in 2010, which is based on NTDB data. Over time, ACS TQIP has added several fields that are only required for participating centers. Further, ACS TQIP has focused significantly on registrar education to achieve greater data standardization. The quality of data in the NTDB has improved significantly as the number of centers participating in ACS TQIP has increased. Data from ACS TQIP centers represent a large subset of NTDB data, which differ from the non-TQIP centers ...