Interpretation of the Ws statistic: application to an integrated trauma system

Comparison of trauma survival rates between institutions and reference data bases is hampered by different injury severity mixes. To overcome this, a standardized comparison using a stratified W statistic (Ws) has been proposed. Ws enables comparisons but does not represent actual survival rates at an institution.

This discussion paper uses, as an example, data from 632 patients with blunt injuries treated by an integrated trauma system comprising a helicopter emergency medical service and the Royal London Hospital. Using the United Kingdom Major Trauma Outcome Study as a reference data set, Ws is calculated and its usefulness discussed. The problem of overprediction of survival is discussed. A modification of the Ws method is used to allow for this and to estimate the actual survival rate.

Ws is useful for comparison between centers. Trauma and Injury Severity Score methodology overpredicts survival in certain probability of survival intervals. Adjustment can be made to estimate the actual survival rate. When applied to an integrated trauma system, such adjustments demonstrate overall improved survival, most noticeable in a more severely injured subgroup.