The AIS was described in 1971 after being developed jointly by the American Medical Association (AMA), the Association for the Advancement of Automotive Medicine (AAAM) and the Society of Automotive Engineers [8]. The development was consensus-driven and was initially devised to classify injuries sustained in motor vehicle crashes with the aim of improving vehicular safety, but has now expanded to include other mechanisms of injury. The score provides a simple numerical method that is anatomically based (head, face, neck, thorax, abdomen, spine, upper and lower extremities, and external). It has no physiological contribution. The head AIS ranks injuries from 0 to 6 (Table 6.2) using neuroradiologic and operative findings. In addition, it has been expanded to include intracranial injuries based upon location, number, and size. Hence the scoring is often performed after the fact, commonly at discharge from hospital.

The ISS [9] is a composite measure derived from the AIS score that rates the three most severely injured body regions out of six (head or neck, face, chest, abdomen or pelvis, extremities or pelvis, and external). ISS can range from 0 to 75 and is calculated by summing the squares of the highest three component values. A patient with an AIS score of 6 in any anatomic region is automatically assigned an ISS of 75. Major trauma or polytrauma is defined by a total ISS greater than 15 [10]. The ISS score correlates linearly with several markers and surrogates for severity, such as mortality, morbidity, and hospital stay.

There have been numerous attempts to combine anatomic (and thereby radiographic) scores (AIS, ISS, Marshall CT grading) with physiological or clinical data (GCS, Revised Trauma Score) to predict outcome after trauma. The New ISS, the Acute Physiology and Chronic Health Evaluation (APACHE), and the Trauma and Injury Severity Score (TrISS) have also been utilized for predicting trauma mortality. However, the use of these scores has been demonstrated to be limited due to frequent misclassification and false positives or negatives [11, 12]. On the other hand, studies have demonstrated that utilizing combined physiological data (GCS or GCS-M) and anatomic scores (AIS or ISS) can improve outcome predictions , although this has been inconsistent. Furthermore, the fact that the determination of AIS and ISS are determined over time, as radiographic and other injury progression information becomes available, makes their utility in the acute phases of resuscitation limited.

Assessment of the literature must take into account the durability of prognostic value as well as the accuracy. Furthermore, care must be taken in interpretation of predictive scores for groups in the early clinical stages of treatment for an individual patient. Scores predicting finite outcomes measures (such as mortality) are limited in their utility in the resuscitation phases of clinical decision-making, because in retrospective clinical studies, decisions to treat or not to treat are not random, and in randomized trials, intention to treat may impact the outcomes of select patient groups. Therefore, the self-fulfilling prophecy of early mortality in those with worse “predictive scores” will select out patients who, had they been treated aggressively and survived, may have had a broad spectrum of outcomes but who would likely have fared poorer.