Quality assurance (QA) systems use audit filters to help identify not only deaths and medical complications, but also cases that reflect deficiencies in the timeliness or appropriateness of care. Through our trauma center QA process, we studied three groups of audit screens for trauma care: two set forth by the Committee on Trauma of the American College of Surgeons, a minimum set of 12 audit filters proposed in 1987 and an expanded version recently proposed in 1990 (studied retrospectively), and one set of filters already in use in our system. A peer review committee determined whether deaths and complication were preventable, and judged the timeliness and appropriateness of care. From July 1, 1989, through June 30, 1990, 844 admitted trauma patients came under QA review. During this period 13 (1.5%) died; 35 (4.1%) suffered at least one medical complication. 140 children (16.6%) violated one of the 12 minimum audit filters suggested by the Committee on Trauma (which includes deaths and medical complications), one of the additional filters used by the Trauma Program QA system, or both. Ninety patients (10.6%) failed one of the 12 minimum audit filters; 104 (12.3%) failed one of the additional filters used by the Trauma Program QA system. Filters that involved medical management issues (late operations, return to operating room, airway, failed reduction, infections, missed injuries, readmissions to intensive care unit, return to the emergency department) frequently involved aspects of inappropriate care (72.5%, 37/51 violations) and were associated with actual deaths or medical complications (52.9%, 27/51).(ABSTRACT TRUNCATED AT 250 WORDS)