Regional variation in anticoagulation and clinical outcomes: scope for improvement

This editorial refers to ‘Inequality in oral anticoagulation use and clinical outcomes in atrial fibrillation: a Danish nationwide perspective’, by A.M.S. Christesen et al., on page 189. The prevalence of atrial fibrillation (AF) in Europe is now over 5 million,1,2 is increasing rapidly,3 and has been described as ‘a modern-day epidemic’.4 Stroke is the most feared complication, and prevention of thrombo-embolic complications is a key component of clinical guidelines.5 Simple, validated tools have been available to clinicians for patient-level thrombo-embolic risk-stratification to guide stroke prophylaxis for over 20 years6,7 and have undergone modification and refinement over time.8 Scoring systems for bleeding risk are also available.9 Smartphone apps and web-based calculators have made these tools easily accessible to busy clinicians, yet adherence to guidelines in practice remains sub-optimal.10 In an elegant analysis using linked registry data in Denmark, Christesen et al. have demonstrated substantial regional variation in oral anticoagulation (OAC) initiation for incident AF.1 Moreover, ischaemic stroke rates assessed in the first year after presentation were observed to be lowest in the region with the highest OAC initiation. The usual limitations relating to observational studies apply and a causal relationship between OAC coverage and clinical outcomes cannot be inferred.11 Nonetheless, this finding is consistent with our understanding of the role of OAC in prevention of cardioembolic stroke.12 The authors calculated, on the assumption that the regional differences in ischaemic stroke represent a causal effect, that 1 in 5 ischaemic strokes within 1 year of baseline could potentially have been prevented if all regions had performed at the level of the best performing region. Reassuringly, in regions with higher OAC uptake there was no increase in haemorrhagic stroke and a small decrease in major bleeding. Whilst the reasons for the observed regional differences in OAC uptake are unknown, the authors propose that differences in health care delivery and in ‘attitude and attention’ towards AF patients between regions may be factors. In particular, they express concern that in some areas responsibility for the initiation of OAC may fall in a ‘no man’s land’ between primary and secondary care which may result in some eligible patients not being treated. In general, regions with the highest levels of initiation of OAC also had the highest persistence on therapy, implying that the factors that contribute to variation in initiation have a continuing influence on the likelihood that a patient will continue on therapy. That increased initiation of therapy was not matched by a rise in rates of discontinuation is reassuring and indicates that facilitating anticoagulation initiation is a worthwhile use of resource. Oral anticoagulation initiation and persistence increased over the course of the 8-year study period, indicating some improvement in practice. The authors identify the publication of European Society of Cardiology (ESC) guidelines in 2010, which updated advice for OAC use in AF, as one possible reason.13 Another is the advent and growth in the use of direct oral anticoagulants (DOAC), the first of which came onto the market in Denmark in 2011, mid-way through the study. Work using primary care data in the UK has also shown temporal improvements in OAC usage, with DOAC used in 40.5% of those newly diagnosed with AF in 2016, compared with 0.3% in 2012 and 9.8% in 2014.14 It seems probable that increasing utilization of DOAC may contribute to closing the gap between guideline-indicated OAC coverage and real-world prescribing through providing an attractive alternative to vitamin-K antagonists for those patients not previously prescribed or reluctant to take OAC. However, there was no apparent association between DOAC use and anticoagulation uptake overall in this study—the lowest performing region in Denmark for overall OAC initiation in fact had the highest proportion of people prescribed a DOAC. Disappointingly, despite the overall rise in OAC initiation with time, regional differences did not appear to decrease over the course of the study. It is possible that some of the observed variation in OAC prescription is clinically appropriate and explained by variables not available within this data set. Whilst the authors identify no significant baseline differences in HAS-BLED score, there are many other explanations for under-utilization of OAC compared to guideline-indicated care that cannot be fully explored in an observational study. Clinical practice is rarely as clear-cut as in guidelines, with many competing considerations. At a patient level, weighing up risk and benefit is a complex and often nuanced decision. Patients’ opinions on OAC therapy for stroke prevention are divergent—in one study 12% of patients reported that they would not take OAC, even if it was 100% effective in preventing stroke.15 There is also evidence of myths and misunderstandings influencing patient’s decisions regarding accepting OAC.16 Clinicians too have anxieties about OAC prescribing, with many physicians reporting themselves as under-skilled in managing complex patients with AF,17 and bleeding a particular (and often over-estimated) concern.18 Decision-making with regard to OAC can be particularly challenging in an ageing and increasingly frail population.19 Christesen et al. give examples of nurse-led, structured AF services which have improved guideline adherence and clinical outcomes20,21 and have hence contributed to overcoming some of the barriers to OAC uptake. The authors suggest that these structured AF services may have been unevenly distributed in Denmark. A further possible barrier is clinician understanding. The ESC has recently completed an educational needs assessment to guide targeted education and advocacy with the aim of increasing guideline implementation in AF care.17 The authors also highlight the recent IMPACT-AF study.22 This was a cluster-randomized trial of a multifaceted educational intervention to improve OAC use in patients with AF, and showed a 12% increase in OAC use in the intervention group compared with 3% in the control arm. These findings highlight the importance of educating patients, their families and their doctors in order to optimize OAC therapy for stroke prevention. They provide evidence that assisting physicians with access to the required knowledge, removing practical barriers to guideline implementation and ongoing support all have a role in reducing unwarranted variation in care. That such variation in stroke prophylaxis still exists despite many years of robust evidence of benefit and internationally agreed guidelines is concerning5,13 and should act as a stimulus for improvement. Cardioembolic stroke disease has a huge associated societal and human cost23 and is largely preventable.24 The authors describe how initiatives have recently been launched in Denmark aimed at improving guideline compliance and standardization of care through audit of performance measures including the initiation and persistence of OAC therapy. International emulation of the Danish example is warranted with a particular focus on OAC optimization and addressing unwarranted variation in practice. Conflict of interest: Dr Cowan chaired the NICE Guideline on the management of atrial fibrillation published in 2014. Dr Wilkinson has nothing to disclose.