Development and testing of a decision tree for blunt trauma.

The aim of the present study was to examine the essential problems in a retrospective study of 381 organ injuries in 260 patients, to identify problems, to define criteria, to describe decision rules, and to organize these rules into branch-chain decision trees or clinical algorithms. The basic hypothesis of this study is that criteria organized into a prioritized decision tree can provide objective standards to evaluate the quality of trauma care and to compare alternative approaches. The algorithm was designed to provide prompt therapy for the most life-threatening problems: respiratory and cardiac arrest, shock, head injury, tamponade, lacerations of the great vessels, cardiac contusion, ruptured parenchymal organs, lacerated viscera, and injury to other intraperitoneal organs. Resuscitation from shock, correction of circulatory problems, and monitoring of physiologic variables were prioritized to evaluate the presence of circulatory deficits and the adequacy of specific therapy to correct them. Concomitantly, diagnosis of the underlying problems was approached using peritoneal lavage, abdominal and chest x-rays, iv urograms, cystograms, endoscopy, upper and lower GI barium or hypaque studies, ultrasound, scintograms, and CT scans. In emergency conditions these are limited to a large extent by time factors. The diagnostic accuracy, priorities, and limitations of each of these were evaluated in emergency conditions. The algorithm was used to track management decisions in a prospective series; the mortality of 51 patients with satisfactory compliance was 4% and 44% in nine patients with major deviations from the algorithm.