Summary of the discussion: trauma registry data and TRISS evidence.

I appreciate the opportunity to participate in this symposium and the invitation to serve on a panel presentation with Jerry Jurkovich, Charlie Mock, and Howard Champion. My comments will focus on several issues that we have begun to address today and several topics that need to be incorporated into our deliberations. First, the difficult methodologic issues already raised in our symposium on preventable trauma death studies and trauma registries are not unique to studies of trauma care effectiveness. An overriding concern throughout health care is how to better establish an evidence base for health services provided and how to establish this evidence within existing budgets. Enormous gaps persist in our capacity to study health care processes and outcomes. Mark Pauly, a health economist at the Wharton School, has summarized the general problem: "Current techniques for measuring appropriateness and effectiveness of care are deficient in the sense of not accurately measuring initial severity, not capturing the full spectrum ... of outcomes, and not being based on good statistical methodologies."1 That we lack definitive measures of trauma system effectiveness is not surprising. What would be surprising is if such measures were readily available.