Previously, we developed a protocol for shock resuscitation of severe trauma patients to reverse shock and regain hemodynamic stability during the first 24 intensive care unit (ICU) hours. Key hemodynamic measurements of cardiac output and preload were obtained using a pulmonary artery catheter (PAC). As an alternative, we developed a protocol that used central venous pressure (CVP) to guide decision making for interventions to regain hemodynamic stability [mean arterial pressure (MAP) >or= 65 mmHg and heart rate (HR) or= 6 mEq/L or systolic blood pressure < 90 mmHg, 3) transfusion of >or= 1 unit packed red blood cells (PRBC), or >or= age 65 years with two of three criteria. Patients with brain injury were excluded. Data were recorded prospectively. In 24 months ending July 31, 2006, of 193 patients, 114 (59%) were assigned CVP- directed resuscitation, and 79 (41%) were assigned PAC-directed resuscitation. A subgroup of 11 (10%) initially assigned CVP was reassigned PAC-directed resuscitation (7 +/- 2 h after start) due to hemodynamic instability. Crystalloid fluid and PRBC resuscitation volumes for PAC (8 +/- 1 L lactated Ringer's [LR], 5 +/- 0.4 units PRBC) were > CVP (5 +/- 0.4 L LR, 3 +/- 0.3 units PRBC) and similar to CVP - PAC protocol subgroup patients (9 +/- 2 L LR, 5 +/- 1 units PRBC). Intensive care unit (ICU) stay and survival rate for PAC (18 +/- 2 days, 75%) were similar to CVP - PAC (17 +/- 4 days, 73%) and worse than CVP protocol subgroup patients (9 +/- 1 days, 98%). Traumatic shock resuscitation is feasible using CVP as a primary hemodynamic monitor as part of a protocol that includes explicit definition of hemodynamic instability and where PAC monitoring is readily available. Computerized decision support provides a technique to implement complex protocol care processes and analyze patient response.