Abbreviated Injury Scale

BackgroundA threshold Injury Severity Score (ISS) ≥ 16 is common in classifying major trauma (MT), although the Abbreviated Injury Scale (AIS) has been extensively revised over time. The aim of this study was to determine effects of different AIS revisions (1998, 2008 and 2015) on clinical outcome measures.MethodsA retrospective observational cohort study including all primary admitted trauma patients was performed (in 2013-2014 AIS98 was used, in 2015-2016 AIS08, AIS08 mapped to AIS15). Different ISS thresholds for MT and their corresponding observed mortality and intensive care (ICU) admission rates were compared between AIS98, AIS08, and AIS15 with Chi-square tests and logistic regression models.ResultsThirty-nine thousand three hundred seventeen patients were included. Thresholds ISS08 ≥ 11 and ISS15 ≥ 12 were similar to a threshold ISS98 ≥ 16 for in-hospital mortality (12.9, 12.9, 13.1% respectively) and ICU admission (46.7, 46.2, 46.8% respectively). AIS98 and AIS08 differed significantly for in-hospital mortality in ISS 4-8 (χ2 = 9.926, p = 0.007), ISS 9-11 (χ2 = 13.541, p = 0.001), ISS 25-40 (χ2 = 13.905, p = 0.001) and ISS 41-75 (χ2 = 7.217, p = 0.027). Mortality risks did not differ significantly between AIS08 and AIS15.ConclusionISS08 ≥ 11 and ISS15 ≥ 12 perform similarly to a threshold ISS98 ≥ 16 for in-hospital mortality and ICU admission. This confirms studies evaluating mapped datasets, and is the first to present an evaluation of implementation of AIS15 on registry datasets. Defining MT using appropriate ISS thresholds is important for quality indicators, comparing datasets and adjusting for injury severity.Level of evidencePrognostic and epidemiological, level III.

BackgroundThe purpose of Abbreviated Injury Scale (AIS) is to code various types of Traumatic Brain Injuries (TBI) based on their anatomical location and severity. The Marshall CT Classification is used to identify those subgroups of brain injured patients at higher risk of deterioration or mortality. The purpose of this study is to determine whether and how AIS coding can be translated to the Marshall ClassificationMethodsInitially, a Marshall Class was allocated to each AIS code through cross-tabulation. This was agreed upon through several discussion meetings with experts from both fields (clinicians and AIS coders). Furthermore, in order to make this translation possible, some necessary assumptions with regards to coding and classification of mass lesions and brain swelling were essential which were all approved and made explicit.ResultsThe proposed method involves two stages: firstly to determine all possible Marshall Classes which a given patient can attract based on allocated AIS codes; via cross-tabulation and secondly to assign one Marshall Class to each patient through an algorithm.ConclusionThis method can be easily programmed in computer softwares and it would enable future important TBI research programs using trauma registry data.

BackgroundTrauma systems should consistently monitor a given trauma population over a period of time. The Abbreviated Injury Scale (AIS) and derived scores such as the Injury Severity Score (ISS) are commonly used to quantify injury severities in trauma registries. To reflect contemporary trauma management and treatment, the most recent version of the AIS (AIS08) contains many codes which differ in severity from their equivalents in the earlier 1998 version (AIS98). Consequently, the adoption of AIS08 may impede comparisons between data coded using different AIS versions. It may also affect the number of patients classified as major trauma.MethodsThe entire AIS98-coded injury dataset of a large population based trauma registry was retrieved and mapped to AIS08 using the currently available AIS98-AIS08 dictionary map. The percentage of codes which had increased or decreased in severity, or could not be mapped, was examined in conjunction with the effect of these changes to the calculated ISS. The potential for free text information accompanying AIS coding to improve the quality of AIS mapping was explored.ResultsA total of 128280 AIS98-coded injuries were evaluated in 32134 patients, 15471 patients of whom were classified as major trauma. Although only 4.5% of dictionary codes decreased in severity from AIS98 to AIS08, this represented almost 13% of injuries in the registry. In 4.9% of patients, no injuries could be mapped. ISS was potentially unreliable in one-third of patients, as they had at least one AIS98 code which could not be mapped. Using AIS08, the number of patients classified as major trauma decreased by between 17.3% and 30.3%. Evaluation of free text descriptions for some injuries demonstrated the potential to improve mapping between AIS versions.ConclusionsConverting AIS98-coded data to AIS08 results in a significant decrease in the number of patients classified as major trauma. Many AIS98 codes are missing from the existing AIS map, and across a trauma population the AIS08 dataset estimates which it produces are of insufficient quality to be used in practice. However, it may be possible to improve AIS98 to AIS08 mapping to the point where it is useful to established registries.

BackgroundThe Abbreviated Injury Scale (AIS) and Injury Severity Score (ISS) grade the severity of injuries and are useful for trauma audit and benchmarking. However, AIS coding is complex and requires specifically trained staff. A simple yet reliable scoring system is needed. The aim of this study was two-fold. First, to develop and validate a simplified AIS (sAIS) chart centred on the most frequent injuries for use by non-trained healthcare professionals. Second, to evaluate the diagnostic accuracy of the sAIS (index test) to calculate the simplified ISS (sISS) to identify major trauma, compared with the reference AIS (rAIS) to calculate the reference ISS (rISS).MethodsThis retrospective study used data (2013-2014) from the Northern French Alps Trauma Registry to develop and internally validate the sAIS. External validation was performed with data from the Trauma Registry of Acute Care of Lausanne University Hospital, Switzerland (2019-2021). Both datasets comprised a random sample of 100 injured patients. Following the Standards for Reporting of Diagnostic Accuracy Studies 2015 guidelines, all patients completed the rAIS and the sAIS. The sISS and the rISS were calculated using the sAIS and the rAIS, respectively. Accuracy was evaluated with the mean difference between the sISS and the rISS and the Pearson correlation coefficient. A clinically relevant equivalence limit was set at ± 4 ISS points. Precision was analyzed using Bland-Altmann plots with 95% limits of agreement.ResultsAccuracy was good. The mean ISS difference of 0.97 (95% CI, -0.03 to 1.97) in the internal validation dataset and - 1.77 (95% CI, - 3.04 to 0.50) in the external validation dataset remained within the equivalence limit. The Pearson correlation coefficient was 0.93 in the internal validation dataset (95% CI, 0.90-0.95) and 0.82 in the external validation dataset (95% CI, 0.75-0.88). The limits of agreement were wider than the predetermined relevant range.ConclusionsThe sAIS is accurate, but slightly imprecise in calculating the ISS. The development of this scale increases the possibilities to use a scoring system for severely injured patients in settings with a reduced availability of the AIS.Trial registrationRetrospectively registered.