Decision analytical modelling of strategies for investigating suspected acute aortic syndrome

Abstract

Background Acute aortic syndrome (AAS) requires urgent diagnosis with computed tomographic angiography (CTA). Diagnostic strategies need to weigh the benefits of detecting AAS against the costs of using CTA with a low yield of AAS when the prevalence of AAS is low. We aimed to estimate the cost-effectiveness of diagnostic strategies using clinical probability scoring and D-dimer to select patients with potential symptoms of AAS for CTA. Methods We developed a decision analytical model to simulate the management of patients attending hospital with possible AAS. We modelled diagnostic strategies that used the Aortic Dissection Detection Risk Score (ADD-RS) and D-dimer to select patients for CTA. We used estimates from our meta-analysis, existing literature and clinical experts to model the consequences of diagnostic strategies on survival, health utility, and health and social care costs. We estimated the incremental cost per quality-adjusted life-years gained by each strategy compared with the next most effective alternative on the efficiency frontier. Results A strategy based on the Canadian guideline (CTA if ADD-RS>1 or ADD-RS=1 with D-dimer >500 ng/mL) is cost-effective but would result in high rates of CTA if applied to an unselected population (AAS prevalence 0.26%). The strategy is also cost-effective and would result in lower rates of CTA if applied to a more selected population, such as those with a non-zero clinical suspicion of AAS (prevalence 0.61%). For patients currently receiving CTA, using ADD-RS>1 or D-dimer >500 ng/mL to select patients for CTA is cost-effective. Conclusions A strategy using ADD-RS>1 or ADD-RS=1 with D-dimer >500 ng/mL to select patients for CTA appears cost-effective but primary research is required to evaluate this strategy in practice and determine how suspicion of AAS is identified.