A commonly used indicator of performance of trauma care systems is the preventable death rate, i.e. the proportion of all deaths which could have been avoided, had the care been optimal. While the preventable death rate may be useful for assessing the quality of care, it cannot be used either to demonstrate the need for, or to evaluate the impact of, regional trauma programs.

Basic flaws are described from the review of 29 preventable deaths studies:​ 1) Time trends in trauma severity unrelated to care are not taken into account. 2) Selection bias occurs when there is a restriction of cases by region, hospital, type or cause of injury, time of death, or other variables. 3) Other problems arise if the clinical and post mortem data are drawn inconsistently from varying sources. 4) Also, the choice of criteria for consensus about preventability could by itself explain observed differences in preventable death rates.

A more serious issue is that the preventable death rate denominator does not take into account those patients who would have died but actually survived because of adequate care. The preventable death rate is indeed a function of the proportion of all trauma victims seen in the system who should survive but are at increased risk of dying under conditions of sub-optimal care. An analysis of joint variation of the preventable death rate, the efficacy, and the proportion of salvageable victims illustrates the possibility of erroneous conclusions from the use of preventable death rates, and from the comparison of preventable death rates.

Guidelines for the choice of adequate indicators for evaluation of trauma care conclude the paper.