Evaluating the use of a mobile device for detection of atrial fibrillation in primary care

Br J Cardiol 2021;28(1)doi:10.5837/bjc.2021.005 Leave a comment
Click any image to enlarge

Atrial fibrillation (AF) increases cardio-embolic stroke risk, yet AF diagnosis and subsequent prophylactic anticoagulant prescription rates are suboptimal globally. This project aimed to increase AF diagnosis and subsequent anticoagulation prescription rates in East Midlands Clinical Commissioning Groups (CCGs).

This service improvement evaluation of the East Midlands AF Advance programme investigated the implementation of mobile AF detection devices (Kardia, AliveCor) into primary-care practices within East Midlands CCGs, along with audit tools and clinician upskilling workshops designed to increase AF diagnosis and anticoagulation prescription rates. AF prevalence and prescription data were collected quarterly from July to September (Q3) 2017/18 to April to June/July to September (Q2/3) 2018/19.

AF prevalence increased from 1.9% (22,975 diagnoses) in Q3 2017/18 to 2.4% (24,246 diagnoses) in Q2 2018/19 (p=0.026), while the percentage of high-risk AF patients receiving anticoagulants increased from 80.5% in Q3 2017/18 to 86.9% in Q3 2018/19 (p=0.57), surpassing the Public Health England 2019 target of 85%.

The East Midlands AF Advance programme increased AF diagnosis and anticoagulation rates, which is expected to be of significant clinical benefit. The mobile AF detection devices provide a more practical alternative to traditional 12-lead electrocardiograms (ECGs) and should be incorporated into routine clinical practice for opportunistic AF detection, in combination with medication reviews to increase anticoagulant prescription.


For UK healthcare professionals only


Atrial fibrillation (AF) presents as an abnormal cardiac rhythm characterised by an irregular or abnormally fast (>100 bpm) resting heart rate (HR). AF risk factors include increasing age, diabetes, hypertension and coronary artery disease.1

AF increases stroke risk by roughly fivefold, greater than the risk elicited by hypertension, coronary artery disease or previous heart failure.2 AF-related stroke patients experience greater mortality rates, disability, hospitalisation time and healthcare costs relative to non-AF stroke patients.3

The East Midlands primary healthcare services comprise 19 Clinical Commissioning Groups (CCGs), covering a population of 4.6 million. Public Health England (PHE) estimated that, as of February 2017, 36,183 East Midlands AF patients were undiagnosed, including paroxysmal AF,4 possibly caused by the difficulty of capturing intermittent AF events using 12-lead electrocardiogram (ECG) assessments in clinic visits. Currently, it is estimated that only 70% of AF patients in the East Midlands are diagnosed (CCGs range from 56% to 77%). East Midlands AF rates are greater than the national average (1.79% vs. 1.71%), though this may be due to the greater proportion of individuals over 50 years of age. While the necessity of anticoagulation therapy depends on clinical indication (e.g. rate vs. rhythm control)5 and investigation is required to determine optimum treatment plans, typically AF patients at high stroke risk (CHA2DS2VASc score ≥2) require anticoagulation therapy for stroke prevention. However, prescription rates remain suboptimal globally. PHE has recommended that, by 2019, 85% of high-risk patients should be receiving anticoagulation treatment. Currently, only one East Midlands CCG has achieved this target (86.3%), with a range from 73.1% to 86.3%. The six worst-performing CCGs make up 54% of the gap between the East Midlands average and the 85% target, suggesting these are easy targets for improvement. While current limitations related to AF treatment include the unknown degree of AF burden associated with thromboembolic risk and the direction of causality (i.e. AF may either precede or follow a thromboembolic event),6 the East Midlands Clinical Network has estimated that optimising AF diagnosis and management could prevent an additional 948 strokes and 316 deaths per year, avoiding £11 million yearly hospital admissions costs. Strategies proposed include targeted education of clinical staff and the implementation of mobile AF diagnosis technologies to guide future treatment. Previous efforts to incorporate mobile health technology into AF patient education have proved successful in increasing patient disease knowledge, however, less is known concerning anticoagulation prescription and adherence.7

In response, the East Midlands Clinical Network implemented measures designed to increase AF diagnosis and management in order to reach national targets. This service evaluation aimed to assess the efficacy of the East Midlands AF Advance programme, which was designed to improve patient care, reduce stroke risk and, consequentially, reduce hospital admissions in AF patients.


East Midlands AF Advance programme

The aim of the AF Advance programme was to increase AF detection and prescription rates by focusing on the following three areas:

  1. Practice audit and care improvement. Each practice was provided with training and access to the GRASP-AF (Guidance on Risk Assessment and Stroke Prevention for Atrial Fibrillation) audit tool to support the review of their current AF diagnosis and management in comparison with National Institute for Health and Care Excellence (NICE) guidelines and clinical best practice. Audit was completed within each CCG (quarterly where possible) to determine the impact of any changes made.
  2. Clinician upskilling. AF ‘ambassadors’ (general practitioners and practice nurses) from participating practices were delivered an interactive three-hour training session from a general practitioner with expertise in AF physiology, detection and treatment. Follow-up support was provided to develop improvement plans within their practice, which were reviewed by the CCGs every three months.
  3. AF diagnosis. Mobile AF detection devices were implemented into each participating practice (Kardia, AliveCor, USA) to allow the opportunistic detection of AF in routine primary-care appointments. Training in the use of these devices was provided to the relevant healthcare professionals. Devices were provided in line with practice populations (average of three devices provided per practice). Each participating practice registered their devices in order to allow the collection of usage data.

Data collection

Quantitative data regarding AF diagnosis, prevalence and anticoagulation treatment were collected manually from each participating practice.

Data were collected in Q3 and Q4 of 2017/18, and Q1 and Q2 (and Q3 where possible) of 2018/19. As GP practices could join the programme after data collection commenced and GP practices were subject to closure, merger or withdrawal, the number of participating GP practices varied between quarters.

Statistical analysis

As the data incorporated in this service evaluation are basic descriptive data, only basic statistical analysis was completed. Following normality testing, AF prevalence and anticoagulation rates were examined using repeated measures analysis of variance (IBM SPSS version 25). A p value <0.05 was accepted as statistical significance, and effect sizes are presented as partial eta squared (η2).


This service quality improvement initiative and evaluation collected no patient-level or identifiable data. As such, no prior ethical approval or patient consent was necessary.


All analysed data were normally distributed and, thus, not transformed prior to analysis. Data are presented from 14 participating CCGs with varying numbers of participating practices.

AF detection

The relative percentage of diagnosed AF patients (after accounting for practice dropouts) displayed an increase from 1.9 ± 0.4% in Q3 2017/18 to 2.4 ± 0.4% in Q2 2018/19 (p=0.026, η2=0.39) (data available upon request). GPs reported using the devices with ~40% of patients.

AF anticoagulation rates

Overall, anticoagulation rates were above the target of 85% by the end of the programme, increasing from 80.5 ± 15.3% in Q3 2017/18 to 86.9 ± 2.7% in Q3 2018/19 (range 82.1% to 92.9%), though this effect did not display statistical significance (p=0.57, η2=0.11). However, three CCGs remained below the 85% target in Q3 2018/19 (data available on request).


Implementation of the mobile AF detection device into GP practices, in combination with the practice audit tool and clinician training sessions, resulted in increased rates of AF detection and prevalence, and increased anticoagulation prescription rates, though the increase in anticoagulation prescription rate was not statistically significant.

The increase in AF diagnosis rates using mobile AF detection devices has been observed previously in other settings and contexts. High sensitivity and specificity to AF episodes has been previously demonstrated by the AliveCor device when compared to ECG monitoring in healthy young participants, elite athletes, cardiology patients (94% and 94%, respectively),8 patients at risk of AF (98% and 97%),9 and AF patients following ablation (100% and 97%).10 While previous research suggests that the AliveCor system is a cost-effective stroke avoidance strategy,9 this was not evaluated here.

The final mean anticoagulation rate for the included CCGs was 86.9%, above the PHE target of 85% for 2019, and, as such, this element of the programme can certainly be considered a success. This effect is likely due to the impact of the clinician upskilling sessions, as the increased AF prevalence was accompanied by an appropriate increase in anticoagulation prescription rates, which can only be actioned by the prescribing clinician. This effect has significant implications for the prevention of strokes and the improvement of associated health outcomes. Future research should investigate long-term outcomes (e.g. stroke rates, survival) and the cost-effectiveness of projects incorporating device-enabled AF detection and clinician upskilling. Additionally, treatment persistence and adherence remain important targets for adequate anticoagulation and effective stroke prevention,4 and should be considered in future research. Lastly, future work should address ways to more appropriately integrate the devices into clinical workflow and promote their usage on more patients, provided that long-term clinical efficacy is demonstrated.

A strength of this evaluation is the inclusion of a large number of practices across the East Midlands, giving the programme a large geographical reach and increasing the transferability of the results to the general population. Furthermore, the comprehensive nature (i.e. clinician education, audit support and device provision) represents a novel approach to AF diagnosis and treatment improvements. Limitations of this service evaluation include the dynamic nature of the practice involvement and the lack of control practices. This makes it hard to state definitively that increases in AF diagnosis and treatment were down to the programme, as other factors such as increased general awareness of AF may have played a role. Furthermore, as no patient-level data were available, it was not possible to account for certain potential confounding variables, such as the presence of comorbidities. However, these limitations are fundamental in service improvement initiatives as this was not a strictly controlled trial.

In summary, implementation of the AliveCor mobile AF detection device into GP practices within East Midlands CCGs successfully increased AF diagnosis rates. When supplemented by a practice audit tool and clinician upskilling, an increase in anticoagulation rates to 86.9% was observed, over the PHE target of 85% for 2019. Future work should focus on investigating long-term outcomes and cost-effectiveness of similar multi-faceted interventions to optimise stroke detection and anticoagulation rates and targeting currently under-performing CCGs.

Key messages

  • This article presents the results of the East Midlands Atrial Fibrillation Advance programme service evaluation, which aimed to increase atrial fibrillation (AF) diagnosis by implementing simple hand-held detection devices into GP practices, and increase anticoagulation prescription rates by providing clinician education sessions
  • By combining opportunistic screening with clinician education and audit support, AF prevalence and anticoagulation prescription rates increased to above national targets within the East Midlands
  • This is predicted to reduce future stroke events and subsequent hospitalisations, and the results of this programme will inform future initiatives aiming to increase AF detection and improve anticoagulation prescription

Conflicts of interest

None declared.


NIHR Collaboration for Leadership in Applied Health Research and Care East Midlands (CLAHRC-EM – now recommissioned as NIHR Applied Research Collaboration East Midlands), Health Education England and the East Midlands Academic Health Science Network.


The authors would like to thank: Martin Cassidy (East Midlands Clinical Network); Ben Andersen (Public Health England); Nigel Scarborough (Health Education England); Trudi Lobban (Heart Rhythm Alliance); Kiran Loi (RightCare); Deb Morris (The Stroke Organisation). The authors acknowledge support from the National Institute for Health Research Collaboration for Leadership in Applied Health Research and Care East Midlands (NIHR CLAHRC-EM) and the NIHR Leicester Diet, Lifestyle and Physical Activity Biomedical Research Centre. The views expressed in this publication are those of the authors and not necessarily those of the NHS or the National Institute for Health Research.

Study approval

This service quality improvement initiative and evaluation collected no patient-level or identifiable data. As such, no ethical approval or patient consent was necessary.


1. Benjamin EJ, Levy D, Vaziri SM, D’agostino RB, Belanger AJ, Wolf PA. Independent risk factors for atrial fibrillation in a population-based cohort: the Framingham Heart Study. JAMA 1994;271:840–4. https://doi.org/10.1001/jama.1994.03510350050036

2. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991;22:983–8. https://doi.org/10.1161/01.STR.22.8.983

3. Jørgensen HS, Nakayama H, Reith J, Raaschou HO, Olsen TS. Acute stroke with atrial fibrillation: the Copenhagen Stroke Study. Stroke 1996;27:1765–9. https://doi.org/10.1161/01.STR.27.10.1765

4. Public Health England. Atrial fibrillation prevalence estimates in England: application of recent population estimates of AF. London: Public Health England, 2017. Available from: https://www.gov.uk/government/publications/atrial-fibrillation-prevalence-estimates-for-local-populations

5. Okcun B, Yigit Z, Yildiz A et al. What should be the primary treatment in atrial fibrillation: ventricular rate control or sinus rhythm control with long-term anticoagulation? J Int Med Res 2009;37:464–71. https://doi.org/10.1177/147323000903700222

6. Sposato LA, Cipriano LE, Saposnik G, Vargas ER, Riccio PM, Hachinski V. Diagnosis of atrial fibrillation after stroke and transient ischaemic attack: a systematic review and meta-analysis. Lancet Neurol 2015;14:377–87. https://doi.org/10.1016/S1474-4422(15)70027-X

7. Stephan LS, Almeida ED, Guimarães RB et al. Oral anticoagulation in atrial fibrillation: development and evaluation of a mobile health application to support shared decision-making. Arq Bras Cardiol 2018;110:7–15. https://doi.org/10.5935/abc.20170181

8. Haberman ZC, Jahn RT, Bose R et al. Wireless smartphone ECG enables large-scale screening in diverse populations. J Cardiovasc Electrophysiol 2015;26:520–6. https://doi.org/10.1111/jce.12634

9. Halcox JP, Wareham K, Cardew A et al. Assessment of remote heart rhythm sampling using the AliveCor heart monitor to screen for atrial fibrillation: the REHEARSE-AF study. Circulation 2017;136:1784–94. https://doi.org/10.1161/CIRCULATIONAHA.117.030583

10. Tarakji KG, Wazni OM, Callahan T et al. Using a novel wireless system for monitoring patients after the atrial fibrillation ablation procedure: the iTransmit study. Heart Rhythm 2015;12:554–9. https://doi.org/10.1016/j.hrthm.2014.11.015