is necessary to better characterize the physiologic response to injury at the extremes of age.
Statistical modeling techniques have been used to combine information on the anatomic descriptors of tissue damage, the physiologic response of the body to the injury, age, preexisting chronic conditions, and mechanism of injury in order to produce estimates of the probability of survival. The two most widely used models are the TRISS and ASCOT (A Severity Characterization of Trauma), both of which are based on the AIS, elements of the RTS, age, and mechanism of the injury (Champion et al., 1994, 1996). Refinements of TRISS and ASCOT have been proposed (Cayten et al., 1991). In addition, models using ICD-based versus AIS-based measures of severity have been introduced (Rutledge et al., 1997). Research is needed to validate existing models and to develop new models where necessary. Common databases and standard measures of scale performance should be employed.
Finally, and perhaps most importantly, research is needed to extend existing approaches or to develop new approaches for modeling outcomes other than death, including resource utilization and functional outcomes. Existing measures of severity and case mix have been developed principally for use in studies of mortality and are inadequate for assessing severity in terms of hospital length of stay, treatment costs, or disability (Bull, 1985; MacKenzie et al., 1986; Rutledge et al., 1998). With increasing attention to the determinants and consequences of nonfatal injuries, several efforts have been undertaken to develop a companion to the AIS that maps AIS injury descriptors into scores that better reflect probable degree of impairment or disability when the patient survives the injury (AAAM, 1994; MacKenzie et al., 1996; McClure and Douglas, 1996). Additional work is under way to develop approaches for predicting hospital length of stay and charges (Rutledge et al., 1998). High priority should be given to the evaluation of these approaches across the wide range of injury types and severities.
Evaluation of the effectiveness of trauma care has traditionally focused on survival. However, as more lives are saved, attention also is shifting to nonfatal outcomes. However, in measuring nonfatal outcomes following trauma, it is important to move beyond the use of narrowly defined measures of morbidity, impairment, and performance in basic activities of daily living to include more global measures of health status and health-related quality of life (HRQOL). Ultimately, the goal of good trauma care is the restoration of function that will allow the patient to resume his or her normal everyday activities. Although numerous measures of health status and HRQOL have been proposed in the literature, few have been applied to the study of trauma care and rehabilitation. There is an urgent need for broader use of these measures and for the development of