Breadth of use of The Abbreviated Injury Scale in The National Trauma Data Bank bank

The Abbreviated Injury Scale (AIS) is an anatomic-based injury coding system that strives to provide sufficient detail to differentiate unique injuries for the purposes of research and quality assurance, while limiting the total number of codes to facilitate efficient use. It has been shown that a substantial portion of codes are unused in automotive-trauma specific databases. The goal of this study was to determine the percentage of codes utilized in a nationwide trauma registry that includes multiple mechanisms of injury. Secondary objectives were to examine unused codes and determine the number of codes that were most frequently utilized.

Data were obtained from the National Trauma Data Bank (NTDB) years 2016 and 2017. All injury data were recorded using AIS version 2005 update 2008 (AIS08), which contains 1,999 distinct injury codes. The percentage of the total number of AIS08 codes used in NTDB were determined for each year individually and the combination of both years. The unused codes were then examined manually to identify common characteristics. Finally, the number of codes that provided 95% coverage of all recorded injuries was calculated.

There were 6,661,110 injuries recorded for 1,953,775 patients in NTDB over the two-year period. A small percentage of codes had an incorrect severity level (0.07%) or an incorrect injury code (0.0002%). There were 1,987 (99.4% of the entire AIS dictionary) unique AIS08 codes utilized in each year, with the unused codes varying between years. The unused codes tended to involve specific nerves, dural sinuses, or severe, bilateral injuries. During the combined two-year period, 1,996 (99.8% of the entire dictionary) unique AIS08 codes were used. Although almost every code was used at least once, 95% of the injuries in NTDB used only the 631 (31.6%) most frequent AIS08 codes.

In contrast to automotive specific databases, nearly all the AIS08 codes are used each year in the NTDB. Over a two-year period, only three AIS08 injuries were unused. However, less than a third of AIS08 codes encompass 95% of the injuries. Further research is necessary to determine if common codes should be separated into multiple distinct codes to enhance discriminatory ability of AIS.