Ultrasonographic screening for abdominal aortic aneurysm: Analysis of surgical decisions for cost‐effectiveness

Abstract

A mathematic model is created to determine the economic cost per year of anticipated prolongation of life that would result from a program of abdominal ultrasonographic (US) screening for abdominal aortic aneurysm. The protocol involves US screening at age 60, 67, and 74 years with additional annual follow-up US and examination if an aneurysm of less than 40 mm is detected. Larger aneurysms are assumed to be sent for early elective resection. The benefits and risks for a subset of men with symptoms of intermittent claudication (IC) as an additional risk factor of atherosclerosis is calculated for comparison. Many of the factors on which these calculated costs and benefits are based are approximations and inferences. These include operative mortality for elective and emergent cases, charges for each such condition, cost of US, and anticipated survival following successful aneurysmectomy both with and without concomitant IC. Sensitivity analysis is performed to show how variations in the major parameters alter the outcome of the calculated cost per year of anticipated extension of life.

The credibility of health economic models for health policy decision-making: the case of population screening for abdominal aortic aneurysm

“…(Table 3 on web) Around half of the analysts used a base-case cost that they had estimated themselves [16][17][18][19] and in one paper both a locally calculated cost and a published cost were used.…”

“…In six papers, [13][14][15][16][17][18] a policy of a single (or prevalence) population screen was modelled and in three papers, policies involving multiple population screens were presented. 5,6,19 If estimates of the annual incidence of new AAA of 0.1% per annum are to be believed, 15,19 then the prevalence of AAA at subsequent screens will be small. In terms of costeffectiveness, and when comparing against a policy of no screening, one might expect a policy involving re-screens of the same cohort to appear less favourable than a single screen strategy.…”

“…16 Four of the nine models seem to be structured for 100% attendance at screening. 6,15,16,19 Pilot studies of AAA screening have suggested attendance of around 80%. An assumption that all invited patients attend will tend to make screening appear more favourable than it would be in routine practice.…”

“…Table 3 shows that one study assumes that screening will completely eliminate rupture and emergency surgery. 19 In addition to 100% attendance at screening and 100% sensitivity and specificity of ultrasound, two further assumptions would need to hold in order for this to be true: firstly no ruptures in patients with screendetected AAA below the threshold for surgery, and secondly all patients with screen-detected AAA exceeding this threshold undergo elective repair without delay. As the threshold for surgery used in this study is small (40 mm) and rupture of an AAA of this size is unlikely, the first of these two assumptions could be considered valid.…”

“…5,16,18 Of the six studies that did cite sources, three referenced just one publication [13][14][15] (although one of these contained results of a Medline review), and one each referenced two papers, 17 three papers, 6 and four papers. 19 Of those studies citing more than one source, none provided information on how estimates from these studies were synthesized to generate the base-case value incorporated within the model.Four of the nine studies reported sensitivity analyses for this parameter. Of the three performing one-way analyses, two demonstrated that small changes in mortality following elective AAA repair would have a large impact upon cost-effectiveness.…”