Diagnostic evaluation of patients with blunt abdominal trauma: a decision analysis

Using decision analysis, to compare the expected utility (EU) of diagnostic peritoneal lavage (DPL), computed tomography (CT), and ultrasonography (US) to determine the optimal modality for the evaluation of blunt abdominal trauma (BAT) in hemodynamically stable adults.

Data points for the decision analysis were obtained from three sources: 1) prevalence of BAT and the sensitivity and specificity of each diagnostic modality were determined through a criteria-based review of the literature; 2) rate of BAT necessitating immediate intervention, perioperative complication rate, and operative mortality rate were calculated using data from the authors' institution's trauma registry; and 3) outcome utilities were determined by telephone survey of adults in a random sample of households in the region. The decision tree was constructed and evaluated in standard fashion. For each diagnostic modality, the authors calculated the EU using the minimum, mean, and maximum sensitivity and specificity across a range of prevalence. Mean outcome utilities were used for each branch of the tree when calculating the EU.

The EU of CT was consistently lower than the EUs of DPL and US at all levels of prevalence. However, the rank order of the EUs of US and DPL varied with the prevalence of BAT. When the prevalence was <30%, the EU of US was higher than that for DPL. When the prevalence was 30-40%, the EUs were similar. When the prevalence was >40%, the EU of US was less than that of DPL.

Among institutions operating under constraints similar to those used in this model, the optimal diagnostic modality for the evaluation of BAT can be determined based on the sensitivity and specificity of the modality at their institution and the prevalence of BAT in their patient population.