In patients with non-valvular atrial fibrillation (NVAF), left atrial appendage closure (LAAC) has demonstrated non-inferior efficacy and safety relative to life-long oral anticoagulation therapy (OAT) in a four-year randomized controlled trial (RCT) (PRAGUE-17). Sub-optimal compliance to OAT in the real-world setting (Simmons 2016) has been associated with increased risk of stroke (Ozaki 2020) and may alter efficacy estimates derived from RCTs in which compliance is generally higher. The study aims to model disease outcomes in NVAF patients treated with LAAC versus lifelong OAT when applying trial versus real world compliance to OAT.

Real-world compliance to OAT in the Australian setting was investigated in a 10 percent Pharmaceutical Benefits Schedule (PBS) sample scripts analysis which measured treatment adherence and persistence to new oral anticoagulants (NOACs) and warfarin. Design of the 10 percent PBS analysis was informed by the compliance to medicine working group report and included the longest follow-up of any OAT compliance study identified in the literature. A Markov cohort model was developed to estimate the expected numbers of strokes and major bleeding events in NVAF patients.

Rates of NOAC discontinuation in PRAGUE-17 was higher at 20 months median follow-up (6.5%) versus compliance in the Australian setting (35.4% and 30.0% according to 3 and 6 month ceasing rules at 20 months follow-up). Applying sub-optimal compliance to lifelong OAT demonstrated in the Australian setting resulted in higher numbers of strokes over a life time modelled time horizon compared with LAAC.

Real world compliance to medicines should be a consideration in economic analysis comparing lifelong medications to one-off surgical interventions.