OBJECTIVES: To review health economic models of population screening for abdominal aortic aneurysm (AAA) among elderly males and assess their credibility for informing decision-making. METHODS: A literature review identified health economic models of ultrasound screening for AAA. For each model focussing on population screening in elderly males, model structure and input parameter values were critically appraised using published good practice guidelines for decision analytic models. RESULTS: Twelve models published between 1989 and 2003 were identified. Converting costs to a common currency and base year, substantial variability in cost-effectiveness results were revealed. Appraisals carried out for the nine models focusing on population screening showed differences in their complexity, with the simpler models generating results most favourable to screening. Eight of the nine models incorporated two or more simplifying structural assumptions favouring screening; uncertainty surrounding these assumptions was not investigated by any model. Quality assessments on a small number of parameters revealed input values varied between models, methods used to identify and incorporate input data were often not described, and few sensitivity analyses were reported. CONCLUSIONS: Large variation exists in the cost-effectiveness results generated by AAA screening models. The substantial number of factors potentially contributing to such disparities means that reconciliation of model results is impossible. In addition, poor reporting of methods makes it difficult to identify the most plausible and thus most useful model of those developed.
OBJECTIVES: To review health economic models of population screening for abdominal aortic aneurysm (AAA) among elderly males and assess their credibility for informing decision-making. METHODS: A literature review identified health economic models of ultrasound screening for AAA. For each model focussing on population screening in elderly males, model structure and input parameter values were critically appraised using published good practice guidelines for decision analytic models. RESULTS: Twelve models published between 1989 and 2003 were identified. Converting costs to a common currency and base year, substantial variability in cost-effectiveness results were revealed. Appraisals carried out for the nine models focusing on population screening showed differences in their complexity, with the simpler models generating results most favourable to screening. Eight of the nine models incorporated two or more simplifying structural assumptions favouring screening; uncertainty surrounding these assumptions was not investigated by any model. Quality assessments on a small number of parameters revealed input values varied between models, methods used to identify and incorporate input data were often not described, and few sensitivity analyses were reported. CONCLUSIONS: Large variation exists in the cost-effectiveness results generated by AAA screening models. The substantial number of factors potentially contributing to such disparities means that reconciliation of model results is impossible. In addition, poor reporting of methods makes it difficult to identify the most plausible and thus most useful model of those developed.
Authors: Jeremy D Goldhaber-Fiebert; Natasha K Stout; Jesse Ortendahl; Karen M Kuntz; Sue J Goldie; Joshua A Salomon Journal: Popul Health Metr Date: 2007-10-29