Economic and clinical impact of diagnostic sensitivity for the use of insertable cardiac monitors to detect atrial fibrillation in cryptogenic stroke patients in the United Kingdom

Insertable cardiac monitors (ICMs) are recommended by international clinical guidelines for prolonged cardiac monitoring in patients with cryptogenic stroke (CS). However, only LINQ ICM has been recognised as cost-effective by the National Institute for Health and Care Excellence (NICE) for this indication, as other ICMs may vary in diagnostic yield. It is unclear how this might impact patient outcomes and cost-effectiveness. This study aims to evaluate the clinical and economic impacts of different diagnostic sensitivity levels of ICMs for detecting atrial fibrillation (AF) in CS patients from the perspective of the UK National Health Service (NHS). We built a Markov model with a three-month cycle and a lifetime horizon to assess the cost-effectiveness of ICMs. Patient characteristics and detection probabilities were taken from the CRYSTAL-AF trial, which compared ICM use with the standard of care (SoC), with costs sourced from NHS. Upon detection of AF, patients transition to direct oral anticoagulants to prevent secondary ischaemic stroke (IS). Outcomes were measured in quality-adjusted life years (QALYs) and incremental cost-effectiveness ratios (ICERs). Scenario analyses were performed to examine various levels of AF detection sensitivity levels. LINQ ICM versus SoC resulted in an ICER of £8,670 per QALY, below the NICE willingness-to-pay threshold. Compared with a hypothetical ICM with 30% points lower sensitivity, AF was 37.5% more likely to be detected during long-term follow-up, leading to a 9.5% reduction in the odds of recurrent IS. The lower sensitivity ICM is less cost-effective as a result, with an ICER of £13,479 per QALY versus SoC. These results remain robust under sensitivity analyses. Diagnostic sensitivity substantially impacts the clinical outcomes of ICMs for prolonged cardiac monitoring as well as the cost-effectiveness. ICMs with modestly lower sensitivity than LINQ are likely to be less effective at detecting AF, leading to increased stroke risks and costs, and thereby be less cost-effective.