Trauma centre patient volume and inpatient mortality risk reconsidered

Several studies have examined the relationship between injury volumes and trauma centre outcomes, with varying results attributable to differences in the measurement of volume's effect on mortality and differences in how characteristics are addressed as potential confounders.

This analysis includes all trauma cases reported to the NTDB 2012. The effect of trauma centre volume on patient mortality risk was measured in three different contexts: as a linear function of trauma centre volume, as a dichotomous function comparing patients in trauma centres with and without 1200 or more cases, and as a non-linear function of trauma centre volume. Multivariable weighted Hierarchical Generalized Linear Models were used to account for the combined effects of facility level and patient level covariates. Patient level mortality risk was assessed using the ACS Trauma Quality Improvement Programme methodology.

Trauma centre volume was not a statistically significant predictor (at the α=0.01 level) of patient mortality risk, in any of the three models. Comprehensive adjustments for patient level risk were obtained, with excellent discrimination between survivor and decedent cases. The addition of trauma volume to baseline patient mortality risk yielded no improvement in the accuracy of any model. These results were not sensitive to the inclusion of Level II trauma centres. Equivalent results were obtained by repeating the analysis for the Level I subpopulation only.

Case volume may be a reasonable standard for determining whether adequate numbers of injured patients are available to support training needs and experience requirements of a Level I trauma centre. However, case volume is not a useful predictor of patient mortality in individual facilities. Trauma centre volume has no independent effect, after accounting for the patient level characteristics that predominantly influence mortality.