Background
Abdominal aortic aneurysm (AAA) screening programmes have been established for men in the UK to reduce deaths from AAA rupture. Whether or not screening should be extended to women is uncertain.
Objective
To evaluate the cost-effectiveness of population screening for AAAs in women and compare a range of screening options.
Design
A discrete event simulation (DES) model was developed to provide a clinically realistic model of screening, surveillance, and elective and emergency AAA repair operations. Input parameters specifically for women were employed. The model was run for 10 million women, with parameter uncertainty addressed by probabilistic and deterministic sensitivity analyses.
Setting
Population screening in the UK.
Participants
Women aged ≥ 65 years, followed up to the age of 95 years.
Interventions
Invitation to ultrasound screening, followed by surveillance for small AAAs and elective surgical repair for large AAAs.
Main outcome measures
Number of operations undertaken, AAA-related mortality, quality-adjusted life-years (QALYs), NHS costs and cost-effectiveness with annual discounting.
Data sources
AAA surveillance data, National Vascular Registry, Hospital Episode Statistics, trials of elective and emergency AAA surgery, and the NHS Abdominal Aortic Aneurysm Screening Programme (NAAASP).
Review methods
Systematic reviews of AAA prevalence and, for elective operations, suitability for endovascular aneurysm repair, non-intervention rates, operative mortality and literature reviews for other parameters.
Results
The prevalence of AAAs (aortic diameter of ≥ 3.0 cm) was estimated as 0.43% in women aged 65 years and 1.15% at age 75 years. The corresponding attendance rates following invitation to screening were estimated as 73% and 62%, respectively. The base-case model adopted the same age at screening (65 years), definition of an AAA (diameter of ≥ 3.0 cm), surveillance intervals (1 year for AAAs with diameter of 3.0–4.4 cm, 3 months for AAAs with diameter of 4.5–5.4 cm) and AAA diameter for consideration of surgery (5.5 cm) as in NAAASP for men. Per woman invited to screening, the estimated gain in QALYs was 0.00110, and the incremental cost was £33.99. This gave an incremental cost-effectiveness ratio (ICER) of £31,000 per QALY gained. The corresponding incremental net monetary benefit at a threshold of £20,000 per QALY gained was –£12.03 (95% uncertainty interval –£27.88 to £22.12). Almost no sensitivity analyses brought the ICER below £20,000 per QALY gained; an exception was doubling the AAA prevalence to 0.86%, which resulted in an ICER of £13,000. Alternative screening options (increasing the screening age to 70 years, lowering the threshold for considering surgery to diameters of 5.0 cm or 4.5 cm, lowering the diameter defining an AAA in women to 2.5 cm and lengthening the surveillance intervals for the smallest AAAs) did not bring the ICER below £20,000 per QALY gained when considered either singly or in combination.
Limitations
The model for women was not directly validated against empirical data. Some parameters were poorly estimated, potentially lacking relevance or unavailable for women.
Conclusion
The accepted criteria for a population-based AAA screening programme in women are not currently met.
Future work
A large-scale study is needed of the exact aortic size distribution for women screened at relevant ages. The DES model can be adapted to evaluate screening options in men.
Study registration
This study is registered as PROSPERO CRD42015020444 and CRD42016043227.
Funding
The National Institute for Health Research Health Technology Assessment programme.
Discrete Event Simulation for Decision Modeling in Health Care: Lessons from Abdominal Aortic Aneurysm Screening
