Epidemiologic studies, based on retrospective data from heterogeneous populations with poor control of confounders, led early investigators to conclude that infection was the overriding risk factor for multiple organ failure (MOF). More recent studies have convincingly shown that MOF frequently occurs in the absence of infection. Consequently, we have shifted our research focus away from the traditional infectious models of MOF to the newer "one-hit" and "two-hit" inflammatory models. Clinically, we have chosen to study trauma patients because they are a relatively homogeneous group with a low incidence of common confounders. Trauma also permits a clear distinction between the first insult and the outcome, both temporally and with respect to the definition criteria. In this review we discuss the background, rationale, and our initial attempts to use indicators of the first insult (i.e., tissue injury quantification and clinical signs of shock) and indicators of the host response (i.e., systemic inflammatory response syndrome) to predict MOF early after injury.