Background: Performance improvement is an essential component of the trauma center. TRISS methodology has been applied as a national standard against which trauma centers can compare their outcomes. Earlier reviews of TRISS unexpected survivors sustained the classification of unexpected survivor in the vast majority of cases. Our hypothesis was that the level of care that is currently expected has made the TRISS unexpected survivors a statistical phenomenon only.
Methods: Two hundred seventy TRISS unexpected survivors at a Level I trauma center from 1991 to 1995 were reviewed. Each case was reviewed as a blinded abstract by six reviewers (three of whom are directors at other facilities) and classified as clinically unexpected survivor (confirmed TRISS classification) or clinically expected survivor (did not sustain TRISS classification as unexpected survivor). Data are expressed as mean +/- SD. Statistical significance was achieved at p < 0.05.
Results: Among the 270 patients categorized by TRISS as unexpected survivors, only 10.7% were corroborated as clinically unexpected survivors by this peer review process and 89.3% were reclassified as clinically expected survivors. Confirmed clinically unexpected survivors were more likely to go directly from the emergency department to the operating room (82 vs. 46%; p < 0.05). Age (32 +/- 12 years vs. 40 +/- 19 years; p < 0.05), Injury Severity Score (46 +/- 20 vs. 32 +/- 14; p < 0.05), Revised Trauma Score (2.46 +/- 1.89 vs. 3.11 +/- 1.21; p < 0.05), probability of survival (0.13 +/- 0.13 vs. 0.24 +/- 0.15; p < 0.05), systolic blood pressure in the emergency department (60 +/- 51 mm Hg vs. 109 +/- 33 mm Hg; p < 0.05), hospital length of stay (39.6 +/- 30.3 days vs. 24.0 +/- 23.0 days; p < 0.05), and intensive care unit length of stay (19.5 +/- 20.6 days vs. 9.6 +/- 10.1 days; p < 0.05) were significantly different comparing confirmed versus unsustained classification as unexpected survivors.
Conclusion: Only 10.7% of survivors classified as unexpected by TRISS were corroborated as unexpected by a blinded, peer-review process. TRISS needs to be updated for meaningful interpretation; modifications need to be made and coefficients need to be revised.