TRISS is the result of the Major Trauma Outcome Study [9]. It combines the following three most important and independent predictive factors: (1) anatomic injury severity, quantified as ISS; (2) the physiological response to these injuries, quantified as RTS; and (3) the age of the patient. However, age is only considered as < / ≥55 years. Different formulas for patients with blunt and penetrating injuries provide a probability of survival as the final TRISS score. Many trauma registries use the TRISS or updated and modified versions of this score.

RISC was developed in a large trauma registry (Trauma Register of the German Trauma Society) to improve outcome prediction [10]. In addition to age, physiology, and injury severity, it also takes into consideration the first laboratory values upon admission, such as base deficit or coagulation marker (partial thromboplastin time, international normalized ration, hemoglobin) and prehospital cardiac arrest.

Other trauma registries such as the British Trauma Audit and Research Network (TARN) also use predictive models to estimate the prognosis. The TARN model includes fewer variables but is updated every year; the actual version can be retrieved from their web site.

Finally, scores can not only be used to predict mortality or survival, but also to predict other conditions. As an example, the TASH score determines the probability of a patient needing a mass transfusion (defined as 10 or more units of blood) [11]. This easy-to-calculate score could increase the preparedness for blood trans-fusion when a patient with severe bleeding is admitted, or it can easily be included in treatment algorithms. The score uses blood pressure, heart rate, hemoglobin, base deficit, initial ultrasound results, femur/pelvic fracture, and male gender as predictive factors.

What is a ‘good’ score? A good trauma score should be able to discriminate between patients with minor and severe injuries. Severity, however, is usually associated with mortality and other unfavorable outcomes (Fig. 3.1). Thus, predicted and observed mortality rates are frequently used to evaluate trauma scores. The following four aspects should be analyzed when the quality of a score is measured:

Discrimination describes the ability of a score to separate survivors from non-survivors. Therefore, mean score values (or predictions) should be as different as possible in the two groups. The most frequently used measure for discrimination is the area under the receiver operating characteristic curve (AUC of ROC), a summary measure where all possible score values are used for prediction of survival (or death). Each score value is used as a cut-off value for prediction of death, and sensitivity and specificity (or true and false positive rate) are determined and plotted in the diagram. AUC of ROC values range from 0.5 to 1.0, and the higher the value, the better the outcome. AUC of ROC values of different scoring systems derived from the same dataset are appropriate tools for comparing their discriminating power. However, if results come from different publications or datasets, comparability is limited because the AUC strongly depends on the portion of easy-to-predict cases. In populations where 95 % of cases survived, AUC ROC values were much higher than those obtained in more severely injured patients.