The major aim was to develop a branch-chain decision tree for penetrating truncal injury and to subject this to a prospective trial of its feasibility to track management decisions. In contrast to the conventional trauma study which focuses on highly selected, well-defined surgical problems, this algorithmic approach was designed to look at the whole gamut of problems of unselected patients with penetrating injury as they enter the ED. The branch-chain algorithm primarily focuses on priorities, order of procedures, and the immediate therapeutic options. The algorithm was used to track clinical management of 280 consecutive patients with penetrating truncal injuries; 31 were dead on arrival, 12 were alive on admission but died during their hospitalization. Of 209 patients whose management was in satisfactory compliance, four (2%) died; eight (20%) of 40 patients who had major deviations from the algorithm died (p less than .01).