Hospitals vary greatly in the quality of their trauma care
Research Activities, February 2011, No. 366
Patients treated in the worst-performing hospital trauma centers have a 50 percent higher chance of dying than those treated at the average-performing trauma centers, even after adjusting for severity of patient injury. Patients treated in the best-performing hospitals had a comparable reduction in risk of death. These findings from a new study suggest that hospitals vary greatly in the quality of their trauma care.
Interventions to improve patient outcomes in trauma care need to be developed and tested, suggest Laurent G. Glance, M.D., of the University of Rochester Medical Center, and colleagues. Their study was based on 157,045 trauma patients admitted to 1 of 125 hospitals that contribute patient information to the American College of Surgeons' National Trauma Databank (NTDB) and that treat at least 250 trauma patients each year. The NTDB also includes information on hospital characteristics (trauma certification level, number of beds, teaching status, geographic region, and whether nonprofit or for-profit) and patient characteristics (including mechanism of injury). Seventy percent of the hospitals were either Level I or Level II trauma centers, and nearly two-fifths of the hospitals had more than 400 beds. The researchers adjusted mortality outcomes for patient injury severity, age, gender, mechanism of injury, physiologic information, and whether the patient was admitted from another hospital.
The researchers calculated an adjusted odds ratio (OR) for each hospital (OR <1 if the patient was less likely to die than if treated at an average hospital; OR >1 if the patient was more likely to die than at an average hospital). The study was funded by the Agency for Healthcare Research and Quality (HS16737).
More details are in "The Survival Measurement and Reporting Trial for Trauma (SMARTT): Background and study design," by Dr. Glance, Turner M. Osler, M.D., Andrew W. Dick, Ph.D., and others in the June 2010 Journal of Trauma 68(6), pp. 1491-1497.