BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay- per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected]

http://bmjopen.bmj.com/ on September 28, 2021 by guest. Protected copyright. BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open

Atrial Fibrillation Screen, Management And Guideline Recommended Therapy (AF SMART II) in the rural primary care setting): a cross-sectional implementation study protocol

ForJournal: peerBMJ Open review only

Manuscript ID bmjopen-2018-023130

Article Type: Protocol

Date Submitted by the Author: 22-Mar-2018

Complete List of Authors: Orchard, Jessica; University of Sydney, Sydney Medical School and Charles Perkins Centre Neubeck, Lis; Edinburgh Napier University, Health and Social Care Freedman, Saul; Heart Research Institute University of Sydney, Charles Perkins Centre; University of Sydney, Sydney Medical School and Charles Perkins Centre Webster, Ruth; The George Institute for Global Health, University of New South Wales Patel, Anushka; The George Institute for Global Health, University of New

South Wales http://bmjopen.bmj.com/ Gallagher, Robyn; University of Sydney, Sydney Nursing School Li, Jialin; University of Sydney, Sydney Nursing School Hespe, Charlotte; The University of Notre Dame Australia, School of Medicine Ferguson, Caleb; Western Sydney University and Western Sydney Local Health District, Western Sydney Nursing and Midwifery Centre Zwar, Nicholas; University of Wollongong Faculty of Science Medicine and Health

Lowres, Nicole; Heart Research Institute University of Sydney, Charles on September 28, 2021 by guest. Protected copyright. Perkins Centre; University of Sydney

atrial fibrillation, screening, oran anticoagulants, electronic decision Keywords: support

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 1 of 23 BMJ Open

1 2 3 Atrial Fibrillation Screen, Management And Guideline Recommended Therapy (AF SMART II) in the 4 rural primary care setting): a cross-sectional implementation study protocol 5 6

7 1 2 1,3 4 4 8 Authors: Jessica Orchard, * Lis Neubeck, Ben Freedman, Ruth Webster, Anushka Patel, Robyn 9 Gallagher, 5 Jialin Li,5 Charlotte Hespe,6 Caleb Ferguson,7 Nick Zwar8 and Nicole Lowres 1,3 10 11 12 13 Affiliations 14 1. Sydney Medical School and Charles Perkins Centre, University of Sydney, Sydney, NSW, 15 16 Australia For peer review only 17 2. School of Health and Social Care, Edinburgh Napier University, Sighthill, Edinburgh UK 18 3. Heart Research Institute, Sydney, NSW, Australia 19 4. The George Institute for Global Health, University of New South Wales, Sydney, Australia 20 5. Sydney Nursing School, University of Sydney, Sydney, NSW, Australia 21 22 6. General Practice Research, School of Medicine, The University of Notre Dame, Sydney, 23 Australia 24 7. Western Sydney Nursing and Midwifery Centre, Western Sydney University & Western 25 Sydney Local Health District, Sydney Australia 26 8. School of Medicine, Faculty of Science, Medicine and Health, University of Wollongong, 27 28 Wollongong, Australia 29 30 Keywords: atrial fibrillation, screening, stroke, oral anticoagulants, electronic decision support 31 http://bmjopen.bmj.com/

32 33 Word count: 2784 words, excluding abstract 34 35 36 37 *Corresponding author 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 Address for correspondence: 42 43 Jessica Orchard 44 Charles Perkins Centre (D17), Level 2 45 46 University of Sydney NSW 2006 47 48 Ph+ 61 2 8627 1664 49 50 Email: [email protected] 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 2 of 23

1 2 3 ABSTRACT 4 5 Background: Screening for atrial fibrillation (AF) in people ≥65 years is now recommended by 6 guidelines and expert consensus. While AF is often asymptomatic, it is the most common heart 7 arrhythmia, and is associated with increased risk of stroke. Early identification and treatment with 8 oral anticoagulants can substantially reduce stroke risk. The general practice setting is ideal for 9 opportunistic screening and provides a natural pathway for treatment for those identified. 10 11 This study aims to investigate the feasibility of implementing screening for AF in rural general 12 practice using novel electronic tools. It will assess whether screening will fit within an existing 13 14 workflow to quickly and accurately identify AF, and will potentially inform a generalisable, scalable 15 approach. 16 For peer review only 17 Methods and analysis: Screening with a smartphone electrocardiogram will be conducted by 18 general practitioners (GPs) and practice nurses in rural general practices in New South Wales, 19 Australia for 3-4 months during 2018-19. Up to 10 practices will be recruited, and we aim to screen 20 2,000 patients aged ≥65 years. Practices will be given an electronic screening prompt and electronic 21 22 decision support to guide evidence-based treatment for those with AF. De-identified data will be 23 collected using a clinical audit tool and qualitative interviews will be conducted with selected 24 practice staff. A process evaluation and cost-effectiveness analysis will also be undertaken. 25 Outcomes include implementation success (proportion of eligible patients screened, fidelity to 26 protocol), proportion of people screened identified with actionable AF, and rates of treatment with 27 28 anticoagulants and antiplatelets at baseline and completion. Results will be compared against an 29 earlier metropolitan study and a ‘control’ dataset of practices. 30 31 Ethics and dissemination: Ethics approval was received from the University of Sydney Human http://bmjopen.bmj.com/ 32 Research Ethics Committee on 27 February 2018 (Project no.: 2017/1017). Results will be 33 disseminated through various forums, including peer-reviewed publication and conference 34 presentations. 35 36 Clinical trials registration number: ACTRN12618000004268p 37 38 Trial status: not yet recruiting 39 on September 28, 2021 by guest. Protected copyright. 40 Data sharing statement: Data will not be made available as data sharing is not permitted by our 41 ethics approval. 42 43 ARTICLE SUMMARY: STRENGTHS AND LIMITATIONS OF THIS STUDY 44 45 • New technology is used to quickly and efficiently identify and treat AF, within an existing 46 workflow in the general practice setting 47 • General practices are ideally placed to screen for and diagnose AF, as well to provide a 48 pathway to care for those identified during screening 49 50 • Conducting the study in the rural setting will inform future AF screening policies to extend 51 the evidence base beyond metropolitan settings 52 • A potential limitation is the small size and non-randomised design, which would ideally be 53 overcome in a future large-scale study 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 3 of 23 BMJ Open

1 2 3 INTRODUCTION 4 AF is the most common heart arrhythmia, affecting 33.5 million people worldwide.1 A person’s 5 likelihood of developing AF increases with age, reaching 37% for those aged ≥55 years.2 Importantly, 6 3 4 7 AF is associated with a fivefold increase in risk of stroke. However, 1.4%-1.6% of the population 8 aged ≥65 have undiagnosed AF,5 6 which is commonly asymptomatic.7 8 Stroke caused by AF is largely 9 preventable with appropriate oral anticoagulant (OAC) therapy, which can reduce AF-related stroke 10 risk by 64%.9 Although OAC is strongly recommended by clinical practice guidelines for those with 11 additional stroke risk factors, as set out in the CHA₂DS₂-VASc score,10 11 there has historically been a 12 12 13 large gap between evidence and practice with only 60% of eligible patients prescribed OAC. 14 Recently, there are signs of improvement, with a recent Public Health England report noting 77.9% 15 of eligible AF patients in England were treated with OAC.13 Similar increases in OAC prescription, 16 largely due to theFor introduction peer of novel anticoagulants review (NOACs), haveonly been reported in Spain and 17 14 15 18 Denmark. 19 Screening for AF in people ≥65 years by opportunistic pulse palpation or ECG rhythm strip is now 20 11 16 21 recommended by clinical practice guidelines and international expert consensus. Many studies 22 have demonstrated the effectiveness of opportunistic screening for AF in general practice. The SAFE 23 study demonstrated that opportunistic screening of patients aged ≥65 years by general practitioners 24 (GPs), with a flag prompt on the patient file, was an effective screening method.5 AF screening in 25 general practice was found to be promising during flu vaccination in the Netherlands,17 and was 26 18 27 successfully performed by practice nurses in the UK. Our previous research showed AF screening 28 by practice nurses to be feasible, both opportunistically and during the influenza (flu) vaccination 29 period.19 20 30 31 With 84% of Australian general practices employing practice nurses,21 there is scope for AF screening http://bmjopen.bmj.com/ 32 to be performed during existing practice nurse health appointments. Ideally, practice nurses could 33 screen patients for AF during the flu or shingles vaccination and/or coordinated chronic care 34 consultations. These appointments are already funded by the Australian Government: the flu 35 22 36 vaccination is recommended annually and provided free for patients aged ≥65 years; shingles 37 vaccines will be provided free for people aged ≥70 with a 5 year catch-up program for people aged 38 71-79;23 annual health assessments are funded for those aged ≥75;24 and annual coordination on September 28, 2021 by guest. Protected copyright. 39 consultations are funded by Medicare for patients with chronic conditions (eg diabetes, 40 25 41 cardiovascular) on a care plan. These are all occasions when patients in the target age group are 42 likely to be available and receptive to AF screening, in an appropriate setting. 43 44 While numerous studies have been conducted in metropolitan areas, there is a paucity of data 45 regarding AF screening in rural areas. A recent systematic review of AF screening studies conducted 46 in rural settings acknowledged the need for more information on AF burden and risk factors 47 specifically in rural areas.26 In the Australian context, it is known that people living in rural areas have 48 worse cardiovascular outcomes.27 It has been reported that there is a lack of evidence about rural 49 50 cardiovascular disease prevention and “characteristics associated with quality of care”, which 51 highlights the need for more programs in rural primary care, especially for high risk patients.28 This 52 study aims to implement a screening program with a quality improvement focus, to better inform 53 future AF screening policies to extend the evidence base beyond metropolitan settings. 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 4 of 23

Once a practice has been recruited, the pre-intervention phase will involve setting up the electronic http://bmjopen.bmj.com/ 31 32 study tools and relevant training. Each practice will be provided with several Kardia smartphone 33 electrocardiogram (iECG) devices to screen for AF (Figure 1). This device has approval from the 34 Australian Therapeutic Goods Administration as a Medical Device, Class IIa and has a validated, 35 automated algorithm for detecting AF with 95% sensitivity and 99% specificity.29 36 37 [Figure 1 - Kardia device and iECG reading] 38 39 In addition, practices will be provided with two custom-designed apps that were developed and on September 28, 2021 by guest. Protected copyright. 40 tested in the AF-SMART study ([Figure) (ACTRN12616000850471). These apps are the: 41 42 (1) AF app with screening prompt, which extracts data from the electronic patient record and 43 provides a prompt for AF screening when an eligible patient’s file is opened, and allows 44 clinical staff to record the provisional screening results (Figure 2 and Figure 3). 45 46 47 (2) Electronic decision support system (EDS), which guides evidence-based treatment of those 48 diagnosed with AF. The EDS is a clinically validated quality improvement tool, designed to 49 bridge gaps between evidence and practice in AF risk management.30 The EDS extracts data 50 51 from electronic patient records regarding current OAC and antiplatelet prescriptions and 52 calculates individual stroke risk scores (CHA2DS2-VASc) (Figure 4). It then provides clinical 53 staff with evidence-based medication guidelines to support decisions on OAC prescription. 54 55 The apps are located in the TopBar hosting platform that integrates with the general practice 56 electronic patient record management systems. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 5 of 23 BMJ Open

1 2 3 4 5 [Figure 2 – Screenshot showing test patient file and TopBar apps] 6 7 [Figure 3 – entering an iECG result in TopBar AF app (test patient shown)] 8 9 [Figure 4 – Screenshot showing EDS recommendation for test patient] 10 11 12 Once the technology setup is complete, practice nurses and GPs will receive training on the use of 13 14 the electronic tools and iECG as well as recent developments in evidence-based management of AF 15 for stroke prevention, highlighting guideline recommendations for OAC, recommendations against 16 using aspirin to treatFor AF.11 Practices peer will be providedreview with posters only for the reception area advertising 17 the study and a laminated participant information statement will be available at reception. 18 19 Patient eligibility 20 All patients aged ≥65 years presenting to the general practice for annual flu vaccination, shingles 21 vaccination, chronic care assessment, or seeing their GP, will be eligible for AF screening if they meet 22 23 the following criteria: 24 25 • age ≥65 years, 26 • no current recorded diagnosis of AF, and 27 • have not already been screened with the iECG in the past 12 months. 28 29 Patients with terminal illness, and patients unable to provide informed consent will be excluded. 30 31 Screening protocol http://bmjopen.bmj.com/ 32 The screening intervention will be conducted for a period of 3-4 months in each practice. The 33 process is shown in Figure 5, and summarised below: 34 35 • The AF screening prompt will notify clinical staff a patient is eligible for iECG screening. 36 • The GP/nurse will explain the screening process and purpose, and obtain informed oral 37 38 consent from patients. 39 • The GP/nurse will record a 30 second iECG. An automated interpretation (‘Possible AF’, on September 28, 2021 by guest. Protected copyright. 40 ‘Normal’, or ‘Unclassified’) is provided immediately after successful recording. 41 • The GP/nurse will enter the iECG screening result into the AF app in TopBar. 42 43 • Depending on the iECG result, follow the appropriate protocol: 44 45 Protocol if a patient receives “Possible AF” diagnosis 46 The GP/nurse will inform the patient of the diagnosis and give a brief explanation. If the nurse 47 48 performed the screening, the patient will be referred to their GP. Further investigation is at the GP’s 49 discretion although a 12-lead ECG is recommended to provide additional confirmation for all new AF 50 diagnoses, and to add extra leads for AF diagnostic work-up. For patients with confirmed AF, GPs will 51 be encouraged to use the EDS to review each patient’s CHA₂DS₂-VASc score and guideline- 52 recommended treatment to reduce stroke risk. 53 54 Protocol if a patient receives “Normal” diagnosis 55 56 The GP/nurse will inform the patient of the diagnosis. No further action is required. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 6 of 23

1 2 3 Protocol if a patient receives “Unclassified” diagnosis 4 5 The GP/nurse will inform the patient of the diagnosis and give a brief explanation based on 6 information provided by researchers. Depending on individual patient’s history and iECG, a 12-lead 7 ECG may be recommended but follow-up is at the GP’s discretion. There are a number of conditions 8 that can lead to this particular diagnosis (e.g. sinus tachycardia/bradycardia, left bundle branch 9 block, right bundle branch block or multiple ectopic beats) which may or may not be clinically 10 11 significant, or previously known. 12 13 Continuing professional development (CDP) and quality improvement (QI) 14 points 15 In order to assist in practice recruitment, education and training will be structured to be eligible for 16 CPD points for nursesFor and GPs. peer In addition, allreview practices will be encouraged only to use the research 17 project to conduct an AF quality improvement program that will attract specific QI professional 18 development points. 19 20 [Figure 5 - Process flow of AF screening] 21 22 Study outcomes 23

24 25 • Implementation success through process measures including proportion of eligible patients 26 seen in the practice during the period who are screened, fidelity to the protocol, and time 27 taken to complete the intervention. 28 • Proportion of people screened identified with actionable AF. 29 Proportion of patients where EDS page accessed. 30 • 31 • Proportion of eligible patients prescribed OAC as a result of the intervention. http://bmjopen.bmj.com/ 32 • Proportion of patients with diagnosed AF prescribed antiplatelet therapy during the study 33 period. 34 • Prevalence of AF at baseline (compared to metropolitan and control group, as defined 35 36 below). 37 • Incidence of new AF at completion of the intervention (compared to metropolitan and 38 control group). 39 on September 28, 2021 by guest. Protected copyright. • Rates of treatment with anticoagulants and antiplatelets at baseline and at completion of 40 41 the intervention (compared to metropolitan and control group). 42 • Acceptability, competing demands, barriers, and enablers according to staff involved in the 43 intervention. 44 • Cost-effectiveness analysis (i.e. incremental cost-effectiveness ratio (ICER) of screening per 45 46 quality adjusted life year (QALY) gained, and per stroke avoided), using data from Cadillac et 31 47 al to estimate the present value of QALYs gained for each ischaemic stroke prevented. 48 Data collection and sample size 49 The study aims to screen approximately 2,000 patients in total from up to 10 general practices. 50 Assuming an incidence of 1.4% of people aged ≥65 with unknown AF,32 we estimate 28 new cases of 51 52 AF would be identified. A sample size of 2,000, assuming 1.4% incidence, would provide a 95% 53 confidence interval of 0.93–2.02% (18-40 cases). 54 55 PenCAT will be configured to collect de-identified data from electronic patient records. These data 56 include demographic, medication, and diagnostic information. Data extracts will be taken at 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 7 of 23 BMJ Open

1 2 3 baseline, end of month 1, end of month 2 and end of month 3. The screening period will be 4 extended if required so that practices get all 4 data extracts. Following each data extract, feedback 5 will be provided to practices including a summary of results (number screened by each staff member 6 in total and for that month). 7 8 Following completion of the iECG screening period, selected staff from each practice will be 9 interviewed to gather their feedback as part of a comprehensive process evaluation. Semi-structured 10 11 interviews with selected GPs, practice nurses, and practice managers will be carried out by 12 researchers at the end of the screening intervention. The interviews will be audio-recorded and 13 transcribed. 14 15 Statistical analysis 16 Quantitative analysisFor peer review only 17 Descriptive analyses will be carried out both at the individual practice level and all practices pooled 18 together. Data collected in this study will be compared with data collected from the earlier study in 19 20 metropolitan Sydney, and a database of ‘control’ practices with the same data-parameters collected 21 through concurrent studies run by The George Institute. Analyses of these data will be conducted 22 using SAS, with χ2 tests used to test associations between AF incidence and method of detection 23 (screen-detected or clinically-detected), and two-tail p values <0.05 will be considered significant. 24 25 Cost-effectiveness analysis 26 A basic economic model for AF screening has previously been developed and utilised in the SEARCH- 27 AF33 study. This model will be adapted and extended to evaluate screening for AF by the practice 28 29 nurse in the general practice. A modelled cost-effectiveness analysis from an Australian health 30 funder perspective will be performed comparing the cost of iECG population-based screening for AF, 31 to diagnosed AF in an unscreened population of Australian men and women aged 65 to 84 years. http://bmjopen.bmj.com/ 32 Future costs will be discounted at a rate of 5%. The results will be expressed as an ICER per stroke 33 34 avoided and per QALY gained. 35 Sensitivity analyses will be performed, including varying the base assumptions of a guideline- 36 37 adherence rate for OAC prescription from 40% to 90% and increasing the cost of treatment. We will 34 38 obtain 95% confidence intervals (CIs) for all ICERs using a multifactorial Monte Carlo simulation as 39 described in SEARCH-AF.33 on September 28, 2021 by guest. Protected copyright. 40 41 Qualitative analysis 42 A detailed process evaluation using mixed methods will be undertaken to evaluate the iECG 43 screening process. Realist evaluation35 will be used to analyse the interaction of context, 44 mechanisms for change and the outcomes that are produced in both this study and the metropolitan 45 46 component. In this way we will piece together a detailed understanding of how the intervention 47 influenced the capacity, opportunity and motivation of providers to improve detection and 48 evidence-based treatment of AF. 49 50 Patient involvement 51 Development and refinement of the protocol was an iterative process. Feedback obtained through 52 interviews patients, nurses, GPs and practice managers about their experience and preferences 53 during previous AF screening studies, 19 29 contributed to the development of the protocol. 54 Following the study, participating practices will be provided with a summary of the findings of the 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 8 of 23

1 2 3 study, together with a full copy of any publications, which will be made available to patients in the 4 practice reception area. 5 6 ETHICS AND DISSEMINATION 7 The study will comply with the National Health and Medical Research Council (NHMRC) ethical 8 9 guidelines. Approval for the study was received from the University of Sydney Human Research 10 Ethics Committee on 27 February 2018 (Project no.: 2017/017). Participating practices will provide 11 written, informed consent and patients being screened will provide oral consent. We used the 12 Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist when writing 13 our protocol report.36 The study will be administered by the Charles Perkins Centre, University of 14 15 Sydney. Results of the study will be disseminated through various forums, including peer-reviewed 16 publications, and Forpresentation peer at national andreview international conferences. only 17 18 CONCLUSIONS 19 Screening for AF in people ≥65 years is now recommended by guidelines and international expert 20 11 16 21 consensus. Importantly, a comprehensive screening program should include a system to support 22 evidence-based treatment for patients ultimately diagnosed with the condition.16 General practices 23 are well-placed to opportunistically screen older patients, and are ideal in terms of the pathway for 24 treatment for those identified. This study will provide electronic prompts for screening and test a 25 26 patient-level EDS integrated with the practice’s clinical software, to support evidence-based 27 treatment. 28 29 In previous studies, a range of evidence-based interventions have been developed to increase 30 effective prescribing of guideline-recommended OAC in primary care settings. These include use of 31 electronic decision support (EDS) tools,37 targeted GP-education programs,38 consultant-led primary http://bmjopen.bmj.com/ 32 care anticoagulation assessment clinics,39 and patient focused education interventions.40 Overall, 33 these interventions increase quality prescribing, but the results are varied in terms of effect size and 34 35 duration. One study with a large effect size used a consultant-led anticoagulation assessment within 36 the general practice to identify and review AF patients with a CHA₂DS₂-VASc score ≥1 who were 37 “suboptimally anticoagulated”.39 This study reported a substantial and significant increase in 38 appropriate OAC treatment from 77% to 95% (p<0.0001), suggesting that an intensive, systematic 39 on September 28, 2021 by guest. Protected copyright. approach, together with numerous reminders to improve patient attendance at in-person reviews, 40 39 41 can be very successful in increasing OAC uptake. In contrast, a cluster-randomised study providing 42 doctors with patient level recommendations via an EDS did not show significant improvements in 43 evidence-based OAC treatment.37 44 45 This study aims to investigate the feasibility of implementing screening for AF in rural general 46 practice. It follows on from the first phase of the study conducted in metropolitan general practice, 47 allowing comparison against a metropolitan setting and a control group of standard practices. The 48 49 project will assess if screening will efficiently fit within an existing workflow to quickly and accurately 50 identify AF. The study will also explore the effect of GP and practice nurse education, electronic 51 screening prompts and EDS to facilitate evidence-based treatment. 52 53 The study will potentially have important implications for the implementation of a large scale 54 opportunistic AF screening program for those aged ≥65 years. It is anticipated that the study will 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 9 of 23 BMJ Open

1 2 3 demonstrate that the general practice setting, provided with additional electronic tools, can meet 4 the challenges of large-scale opportunistic screening with a cheap, convenient, scalable solution. 5 6 7 8 Author contributions: LN conceived the original concept of the study, and all authors were involved 9 in the design of the study. LN provided qualitative analysis expertise and NL provided cost- 10 effectiveness analysis expertise. JO, LN, BF and NL were involved in drafting the article. All authors 11 contributed to the critical revision of the drafting and final approval of the manuscript. 12 13 Funding statement: This work was supported by a National Heart Foundation of Australia/ New 14 South Wales Health Cardiovascular Research Network Research Development Project Grant 15 (101133) with top-up funding to complete the study in rural NSW through an investigator-initiated 16 For peer review only 17 grant from Pfizer. The funders have no influence on study design, data collection or the decision to 18 submit reports for publication. AliveCor have provided free Kardia devices for study purposes. JO is 19 supported by an Australian Government Research Training Program (RTP) Scholarship. RW is funded 20 by a NHMRC Early Career Fellowship (APP1125044). NL is funded by a NSW Health Early Career 21 22 Fellowship (H16/ 52168). 23 Competing interests statement: JO, JL, CH, RG and NZ have no disclosures. NL and SBF report grants 24 25 from BMS/Pfizer during the conduct of the study. RW and AP report other financial interests from 26 George Health, outside the submitted work. CF reports personal fees from Pfizer outside the 27 submitted work. LN reports grants from Pfizer/BMS and Boehringer Ingelheim outside the submitted 28 work. BF reports prior fees and advisory board honoraria from Bayer Pharma AG, Boehringer 29 Ingelheim, and BMS/Pfizer but for the past year has removed himself from pharmaceutical advisory 30 31 boards and receives speaker fees only for accredited educational meetings. http://bmjopen.bmj.com/ 32

33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 10 of 23

1 2 3 REFERENCES 4 5 1. Chugh SS, Havmoeller R, Narayanan K, et al. Worldwide epidemiology of atrial fibrillation: a Global 6 Burden of Disease 2010 Study. Circulation 2014;129(8):837-47. doi: 7 10.1161/CIRCULATIONAHA.113.005119 [published Online First: 2013/12/19] 8 2. Weng LC, Preis SR, Hulme OL, et al. Genetic Predisposition, Clinical Risk Factor Burden, and 9 Lifetime Risk of Atrial Fibrillation. Circulation 2018;137(10):1027-38. doi: 10 10.1161/circulationaha.117.031431 [published Online First: 2017/11/14] 11 3. Hughes M, Lip GY, Guideline Development Group NCGfMoAFiP, et al. Stroke and 12 thromboembolism in atrial fibrillation: a systematic review of stroke risk factors, risk 13 stratification schema and cost effectiveness data. Thromb Haemost 2008;99(2):295-304. doi: 14 10.1160/TH07-08-0508 [published Online First: 2008/02/19] 15 4. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the 16 FraminghamFor Study. Strokepeer 1991;22(8):983-88. review only 17 5. Fitzmaurice DA, Hobbs FD, Jowett S, et al. Screening versus routine practice in detection of atrial 18 fibrillation in patients aged 65 or over: cluster randomised controlled trial. BMJ 19 2007;335(7616):383. doi: 10.1136/bmj.39280.660567.55 [published Online First: 20 2007/08/04] 21 6. Lowres N, Neubeck L, Redfern J, et al. Screening to identify unknown atrial fibrillation. A 22 systematic review. Thromb Haemost 2013;110(2):213-22. doi: 10.1160/TH13-02-0165 23 24 [published Online First: 2013/04/19] 25 7. Lee J, Reyes BA, McManus DD, et al. Atrial fibrillation detection using an iPhone 4S. IEEE Trans 26 Biomed Eng 2013;60(1):203-6. doi: 10.1109/TBME.2012.2208112 [published Online First: 27 2012/08/08] 28 8. Savelieva I, Camm AJ. Clinical relevance of silent atrial fibrillation: prevalence, prognosis, quality of 29 life, and management. Journal of Interventional Cardiac Electrophysiology 2000;4(2):369-82. 30 9. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: antithrombotic therapy to prevent stroke in 31 patients who have nonvalvular atrial fibrillation. Annals of internal medicine http://bmjopen.bmj.com/ 32 2007;146(12):857-67. 33 10. January CT, Wann LS, Alpert JS, et al. 2014 AHA/ACC/HRS guideline for the management of 34 patients with atrial fibrillation: a report of the American College of Cardiology/American 35 Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll 36 Cardiol 2014;64(21):e1-76. doi: 10.1016/j.jacc.2014.03.022 [published Online First: 37 2014/04/02] 38 11. Kirchhof P, Benussi S, Kotecha D, et al. 2016 ESC Guidelines for the management of atrial on September 28, 2021 by guest. Protected copyright. 39 fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37(38):2893-962. doi: 40 10.1093/eurheartj/ehw210 [published Online First: 2016/08/28] 41 12. Ogilvie IM, Newton N, Welner SA, et al. Underuse of oral anticoagulants in atrial fibrillation: a 42 systematic review. Am J Med 2010;123(7):638-45 e4. doi: 10.1016/j.amjmed.2009.11.025 43 44 [published Online First: 2010/07/09] 45 13. Public Health England. CVD: Primary Care Intelligence Packs, NHS South Norfolk CCG. 2017. 46 available from: 47 https://wwwgovuk/government/uploads/system/uploads/attachment_data/file/622956/NH 48 S_East_Riding_of_Yorkshire_CCG_CVD_intelligence_packpdf accessed 24 October 2017 49 14. Gadsbøll K, Staerk L, Fosbøl EL, et al. Increased use of oral anticoagulants in patients with atrial 50 fibrillation: temporal trends from 2005 to 2015 in Denmark. Eur Heart J 2017;38(12):899- 51 906. doi: 10.1093/eurheartj/ehw658 52 15. Rodríguez-Bernal CL, Hurtado I, García-Sempere A, et al. Oral Anticoagulants Initiation in Patients 53 with Atrial Fibrillation: Real-World Data from a Population-Based Cohort. Front Pharmacol 54 2017;8:63. doi: 10.3389/fphar.2017.00063 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 11 of 23 BMJ Open

1 2 3 16. Freedman B, Camm J, Calkins H, et al. Screening for Atrial Fibrillation: A Report of the AF-SCREEN 4 International Collaboration. Circulation 2017;135(19):1851-67. doi: 5 10.1161/CIRCULATIONAHA.116.026693 [published Online First: 2017/05/10] 6 17. Tieleman RG, Plantinga Y, Rinkes D, et al. Validation and clinical use of a novel diagnostic device 7 for screening of atrial fibrillation. Europace 2014;16(9):1291-5. doi: 8 10.1093/europace/euu057 [published Online First: 2014/05/16] 9 18. Morgan S, Mant D. Randomised trial of two approaches to screening for atrial fibrillation in UK 10 general practice. The British Journal of General Practice 2002;52(478):373-80. 11 19. Orchard J, Freedman SB, Lowres N, et al. iPhone ECG screening by practice nurses and 12 receptionists for atrial fibrillation in general practice: the GP-SEARCH qualitative pilot study. 13 Aust Fam Physician 2014;43(5):315-9. [published Online First: 2014/05/06] 14 20. Orchard J, Lowres N, Ben Freedman S, et al. Screening for atrial fibrillation during influenza 15 vaccinations by primary care nurses using a smartphone electrocardiograph (iECG): A 16 feasibilityFor study. European peer Journal ofreview Preventive Cardiology only 2016;23(Supp 2):13-20. doi: 17 10.1177/2047487316670255 18 19 21. Britt H, Miller GC, Henderson J, et al. General Practice Activity in Australia 2015-16: BEACH: 20 Bettering the Evaluation and Care of Health. Sydney: Sydney University Press 2016. 21 22. Media Release, Prime Minister and Minister for Health. Ground-breaking flu vaccines to protect 22 millions of Aussies 19 February 2018. 23 http://www.health.gov.au/internet/ministers/publishing.nsf/Content/health-mediarel- 24 yr2018-hunt020.htm, accessed 9 March 2018 25 23. NSW Government. Shingles Vaccination Program 2018. 26 http://wwwhealthnswgovau/immunisation/Pages/Shingles-programaspx, accessed 9 March 27 2018 28 24. Australian Government Department of Health. Health assessment for people aged 75 years and 29 older (2014). 30 http://wwwhealthgovau/internet/main/publishingnsf/Content/mbsprimarycare_mbsitem_7 31 5andolder accessed 19 March 2018 http://bmjopen.bmj.com/ 32 25. Australian Government Department of Health. Chronic Disease Management - Provider 33 Information 2016. 34 http://wwwhealthgovau/internet/main/publishingnsf/content/mbsprimarycare-factsheet- 35 chronicdiseasehtm, accessed 9 March 2018 36 26. Gavino AI, McLachlan CS. Review of screening studies for atrial fibrillation in rural populations of 37 11 countries. Proc (Bayl Univ Med Cent) 2017;30(3):280-85. 38 39 27. Australian Institute of Health and Welfare (AIHW). Cardiovascular Medicines and Primary Health on September 28, 2021 by guest. Protected copyright. 40 Care: A Regional Analysis. Canberra, 2010. 41 28. Allenby A, Kinsman L, Tham R, et al. The quality of cardiovascular disease prevention in rural 42 primary care. Aust J Rural Health 2016;24(2):92-8. doi: 10.1111/ajr.12224 [published Online 43 First: 2015/08/11] 44 29. Orchard J, Lowres N, Freedman SB, et al. Screening for atrial fibrillation during influenza 45 vaccinations by primary care nurses using a smartphone electrocardiograph (iECG): A 46 feasibility study. Eur J Prev Cardiol 2016;23(2 suppl):13-20. doi: 10.1177/2047487316670255 47 [published Online First: 2016/11/29] 48 30. Peiris D, Usherwood T, Panaretto K, et al. The Treatment of cardiovascular Risk in Primary care 49 using Electronic Decision supOrt (TORPEDO) study-intervention development and protocol 50 for a cluster randomised, controlled trial of an electronic decision support and quality 51 improvement intervention in Australian primary healthcare. BMJ Open 2012;2(6):e002177. 52 doi: 10.1136/bmjopen-2012-002177 [published Online First: 2012/11/21] 53 31. Cadilhac DA, Dewey HM, Vos T, et al. The health loss from ischemic stroke and intracerebral 54 hemorrhage: evidence from the North East Melbourne Stroke Incidence Study (NEMESIS). 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 12 of 23

1 2 3 Health Qual Life Outcomes 2010;8:49. doi: 10.1186/1477-7525-8-49 [published Online First: 4 2010/05/18] 5 32. Lowres N, Neubeck L, Redfern J, et al. Community screening programs to identify unknown atrial 6 fibrillation: a systematic review. Eur Heart J 2012;33:61-61. doi: 7 http://dx.doi.org/10.1016/j.hlc.2012.05.018 8 33. Lowres N, Neubeck L, Salkeld G, et al. Feasibility and cost-effectiveness of stroke prevention 9 through community screening for atrial fibrillation using iPhone ECG in pharmacies. The 10 SEARCH-AF study. Thromb Haemost 2014;111(6):1167-76. doi: 10.1160/TH14-03-0231 11 [published Online First: 2014/04/02] 12 34. Briggs AH. Statistical approaches to handling uncertainty in health economic evaluation. Eur J 13 Gastroenterol Hepatol 2004;16(6):551-61. 14 35. Pawson R, Tilley N. Realistic Evaluation. London, UK: Sage Publications 1997. 15 36. Chan AW, Tetzlaff JM, Altman DG, et al. SPIRIT 2013 statement: defining standard protocol items 16 for clinicalFor trials. Ann peer Intern Med 2013;158(3):200-7. review doi: only10.7326/0003-4819-158-3- 17 201302050-00583 [published Online First: 2013/01/09] 18 19 37. Eckman MH, Lip GY, Wise RE, et al. Impact of an Atrial Fibrillation Decision Support Tool on 20 thromboprophylaxis for atrial fibrillation. Am Heart J 2016;176:17-27. doi: 21 10.1016/j.ahj.2016.02.009 [published Online First: 2016/06/07] 22 38. Jackson SL, Peterson GM, Vial JH. A community-based educational intervention to improve 23 antithrombotic drug use in atrial fibrillation. Ann Pharmacother 2004;38(11):1794-9. doi: 24 10.1345/aph.1E152 [published Online First: 2004/09/30] 25 39. Das M, Panter L, Wynn GJ, et al. Primary Care Atrial Fibrillation Service: outcomes from 26 consultant-led anticoagulation assessment clinics in the primary care setting in the UK. BMJ 27 Open 2015;5(12):e009267. doi: 10.1136/bmjopen-2015-009267 [published Online First: 28 2015/12/15] 29 40. McAlister FA, Man-Son-Hing M, Straus SE, et al. Impact of a patient decision aid on care among 30 patients with nonvalvular atrial fibrillation: a cluster randomized trial. CMAJ 31 2005;173(5):496-501. doi: 10.1503/cmaj.050091 [published Online First: 2005/09/01] http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 13 of 23 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 Figure 1 - Kardia device and iECG recording 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 14 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

16 For peer review only 17 Figure 2 - Screenshot showing test patient file and TopBar apps 18 19 2055x640mm (96 x 96 DPI) 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 15 of 23 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 Figure 3 - entering an iECG result in TopBar AF app (test patient shown) 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 16 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Figure 4 - Screenshot showing EDS recommendation for test patient 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 17 of 23 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 Figure 5 - Process flow of AF screening 46 47 280x450mm (300 x 300 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 18 of 23

1 2 3 Reporting checklist for protocol of a clinical trial. 4 5 6 Based on the SPIRIT guidelines. 7 8 Instructions to authors 9 10 11 Complete this checklist by entering the page numbers from your manuscript where readers will find 12 each of the items listed below. 13 14 15 Your article may not currently address all the items on the checklist. Please modify your text to 16 include the missing information.For Ifpeer you are certain review that an item onlydoes not apply, please write "n/a" and 17 18 provide a short explanation. 19 20 Upload your completed checklist as an extra file when you submit to a journal. 21 22 In your methods section, say that you used the SPIRIT reporting guidelines, and cite them as: 23 24 Chan AW, Tetzlaff JM, Altman DG, Laupacis A, Gøtzsche PC, KrležaJerić K, Hróbjartsson A, Mann 25 26 H, Dickersin K, Berlin J, Doré C, Parulekar W, Summerskill W, Groves T, Schulz K, Sox H, Rockhold 27 FW, Rennie D, Moher D. SPIRIT 2013 Statement: Defining standard protocol items for clinical trials. 28 29 Ann Intern Med. 2013;158(3):200207 30 31 Page http://bmjopen.bmj.com/ 32 33 Reporting Item Number 34 35 Title #1 Descriptive title identifying the study design, population, 1 36 37 interventions, and, if applicable, trial acronym 38 39 Trial registration #2a Trial identifier and registry name. If not yet registered, name 2 on September 28, 2021 by guest. Protected copyright. 40 of intended registry 41 42 43 Trial registration: #2b All items from the World Health Organization Trial n/a 44 data set Registration Data Set 45 46 47 Protocol version #3 Date and version identifier n/a 48 49 Funding #4 Sources and types of financial, material, and other support 11 50 51 Roles and #5a Names, affiliations, and roles of protocol contributors 11 52 53 responsibilities: 54 contributorship 55 56 57 Roles and #5b Name and contact information for the trial sponsor 1 58 responsibilities: 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 19 of 23 BMJ Open

sponsor contact 1 2 information 3 4 Roles and #5c Role of study sponsor and funders, if any, in study design; 11 5 6 responsibilities: collection, management, analysis, and interpretation of 7 sponsor and funder data; writing of the report; and the decision to submit the 8 9 report for publication, including whether they will have 10 ultimate authority over any of these activities 11 12 Roles and #5d Composition, roles, and responsibilities of the coordinating n/a 13 14 responsibilities: centre, steering committee, endpoint adjudication 15 committees committee, data management team, and other individuals or 16 For peer review only 17 groups overseeing the trial, if applicable (see Item 21a for 18 data monitoring committee) 19 20 21 Background and #6a Description of research question and justification for 3 22 rationale undertaking the trial, including summary of relevant studies 23 24 (published and unpublished) examining benefits and harms 25 for each intervention 26 27 28 Background and #6b Explanation for choice of comparators 3 29 rationale: choice of 30 31 comparators http://bmjopen.bmj.com/ 32 33 Objectives #7 Specific objectives or hypotheses 4 34 35 Trial design #8 Description of trial design including type of trial (eg, parallel 4 36 37 group, crossover, factorial, single group), allocation ratio, 38 and framework (eg, superiority, equivalence, noninferiority, 39 on September 28, 2021 by guest. Protected copyright. 40 exploratory) 41 42 Study setting #9 Description of study settings (eg, community clinic, 4 43 academic hospital) and list of countries where data will be 44 45 collected. Reference to where list of study sites can be 46 obtained 47 48 49 Eligibility criteria #10 Inclusion and exclusion criteria for participants. If applicable, 6 50 eligibility criteria for study centres and individuals who will 51 52 perform the interventions (eg, surgeons, psychotherapists) 53 54 Interventions: #11a Interventions for each group with sufficient detail to allow 6 55 56 description replication, including how and when they will be 57 administered 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 20 of 23

1 Interventions: #11b Criteria for discontinuing or modifying allocated 7 2 3 modifications interventions for a given trial participant (eg, drug dose 4 change in response to harms, participant request, or 5 6 improving / worsening disease) 7 8 Interventions: #11c Strategies to improve adherence to intervention protocols, 7 9 adherance and any procedures for monitoring adherence (eg, drug 10 11 tablet return; laboratory tests) 12 13 Interventions: #11d Relevant concomitant care and interventions that are n/a 14 15 concomitant care permitted or prohibited during the trial 16 For peer review only 17 Outcomes #12 Primary, secondary, and other outcomes, including the 9 18 19 specific measurement variable (eg, systolic blood pressure), 20 analysis metric (eg, change from baseline, final value, time 21 22 to event), method of aggregation (eg, median, proportion), 23 and time point for each outcome. Explanation of the clinical 24 25 relevance of chosen efficacy and harm outcomes is strongly 26 recommended 27 28 Participant timeline #13 Time schedule of enrolment, interventions (including any 5 29 30 runins and washouts), assessments, and visits for http://bmjopen.bmj.com/ 31 participants. A schematic diagram is highly recommended 32 33 (see Figure) 34 35 Sample size #14 Estimated number of participants needed to achieve study 9 36 37 objectives and how it was determined, including clinical and 38 statistical assumptions supporting any sample size 39 on September 28, 2021 by guest. Protected copyright. 40 calculations 41 42 Recruitment #15 Strategies for achieving adequate participant enrolment to 7 43 reach target sample size 44 45 46 Allocation: sequence #16a Method of generating the allocation sequence (eg, n/a 47 generation computergenerated random numbers), and list of any 48 49 factors for stratification. To reduce predictability of a random 50 sequence, details of any planned restriction (eg, blocking) 51 52 should be provided in a separate document that is 53 unavailable to those who enrol participants or assign 54 55 interventions 56 57 Allocation #16b Mechanism of implementing the allocation sequence (eg, n/a 58 59 concealment central telephone; sequentially numbered, opaque, sealed 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 21 of 23 BMJ Open

mechanism envelopes), describing any steps to conceal the sequence 1 2 until interventions are assigned 3 4 Allocation: #16c Who will generate the allocation sequence, who will enrol 6 5 6 implementation participants, and who will assign participants to 7 interventions 8 9 Blinding (masking) #17a Who will be blinded after assignment to interventions (eg, n/a 10 11 trial participants, care providers, outcome assessors, data 12 analysts), and how 13 14 15 Blinding (masking): #17b If blinded, circumstances under which unblinding is n/a 16 emergency Forpermissible, peer and review procedure for revealingonly a participant’s 17 18 unblinding allocated intervention during the trial 19 20 Data collection plan #18a Plans for assessment and collection of outcome, baseline, 9 21 22 and other trial data, including any related processes to 23 promote data quality (eg, duplicate measurements, training 24 25 of assessors) and a description of study instruments (eg, 26 questionnaires, laboratory tests) along with their reliability 27 28 and validity, if known. Reference to where data collection 29 forms can be found, if not in the protocol 30 http://bmjopen.bmj.com/ 31 Data collection plan: #18b Plans to promote participant retention and complete follow 9 32 33 retention up, including list of any outcome data to be collected for 34 participants who discontinue or deviate from intervention 35 36 protocols 37 38 Data management #19 Plans for data entry, coding, security, and storage, including 9 39 on September 28, 2021 by guest. Protected copyright. 40 any related processes to promote data quality (eg, double 41 data entry; range checks for data values). Reference to 42 43 where details of data management procedures can be 44 found, if not in the protocol 45 46 Statistics: outcomes #20a Statistical methods for analysing primary and secondary 9 47 48 outcomes. Reference to where other details of the statistical 49 analysis plan can be found, if not in the protocol 50 51 52 Statistics: additional #20b Methods for any additional analyses (eg, subgroup and 10 53 analyses adjusted analyses) 54 55 56 Statistics: analysis #20c Definition of analysis population relating to protocol non n/a 57 population and adherence (eg, as randomised analysis), and any statistical 58 59 missing data methods to handle missing data (eg, multiple imputation) 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 22 of 23

1 Data monitoring: #21a Composition of data monitoring committee (DMC); summary n/a 2 3 formal committee of its role and reporting structure; statement of whether it is 4 independent from the sponsor and competing interests; and 5 6 reference to where further details about its charter can be 7 found, if not in the protocol. Alternatively, an explanation of 8 9 why a DMC is not needed 10 11 Data monitoring: #21b Description of any interim analyses and stopping guidelines, n/a 12 interim analysis including who will have access to these interim results and 13 14 make the final decision to terminate the trial 15 16 Harms #22For Plans peer for collecting, review assessing, reporting,only and managing n/a 17 18 solicited and spontaneously reported adverse events and 19 other unintended effects of trial interventions or trial conduct 20 21 22 Auditing #23 Frequency and procedures for auditing trial conduct, if any, n/a 23 and whether the process will be independent from 24 25 investigators and the sponsor 26 27 Research ethics #24 Plans for seeking research ethics committee / institutional 2 28 approval review board (REC / IRB) approval 29 30 31 Protocol #25 Plans for communicating important protocol modifications 2 http://bmjopen.bmj.com/ 32 amendments (eg, changes to eligibility criteria, outcomes, analyses) to 33 34 relevant parties (eg, investigators, REC / IRBs, trial 35 participants, trial registries, journals, regulators) 36 37 38 Consent or assent #26a Who will obtain informed consent or assent from potential 7 39 trial participants or authorised surrogates, and how (see on September 28, 2021 by guest. Protected copyright. 40 41 Item 32) 42 43 Consent or assent: #26b Additional consent provisions for collection and use of n/a 44 ancillary studies participant data and biological specimens in ancillary 45 46 studies, if applicable 47 48 Confidentiality #27 How personal information about potential and enrolled 9 49 50 participants will be collected, shared, and maintained in 51 order to protect confidentiality before, during, and after the 52 53 trial 54 55 Declaration of #28 Financial and other competing interests for principal 12 56 interests investigators for the overall trial and each study site 57 58 59 Data access #29 Statement of who will have access to the final trial dataset, n/a 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 23 of 23 BMJ Open

and disclosure of contractual agreements that limit such 1 2 access for investigators 3 4 Ancillary and post #30 Provisions, if any, for ancillary and posttrial care, and for n/a 5 6 trial care compensation to those who suffer harm from trial 7 participation 8 9 Dissemination policy: #31a Plans for investigators and sponsor to communicate trial 10 10 11 trial results results to participants, healthcare professionals, the public, 12 and other relevant groups (eg, via publication, reporting in 13 14 results databases, or other data sharing arrangements), 15 including any publication restrictions 16 For peer review only 17 18 Dissemination policy: #31b Authorship eligibility guidelines and any intended use of 11 19 authorship professional writers 20 21 22 Dissemination policy: #31c Plans, if any, for granting public access to the full protocol, 2 23 reproducible participantlevel dataset, and statistical code 24 25 research 26 27 Informed consent #32 Model consent form and other related documentation given n/a 28 materials to participants and authorised surrogates 29 30 31 Biological specimens #33 Plans for collection, laboratory evaluation, and storage of n/a http://bmjopen.bmj.com/ 32 biological specimens for genetic or molecular analysis in the 33 34 current trial and for future use in ancillary studies, if 35 applicable 36 37 38 The SPIRIT checklist is distributed under the terms of the Creative Commons Attribution License CC 39 BYND 3.0. This checklist was completed on 16. March 2018 using http://www.goodreports.org/, a on September 28, 2021 by guest. Protected copyright. 40 41 tool made by the EQUATOR Network in collaboration with Penelope.ai 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open

Atrial Fibrillation Screen, Management And Guideline Recommended Therapy (AF SMART II) in the rural primary care setting): an implementation study protocol

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2018-023130.R1

Article Type: Protocol

Date Submitted by the Author: 04-Jun-2018

Gallagher, Robyn; University of Sydney, Sydney Nursing School http://bmjopen.bmj.com/ Li, Jialin; University of Sydney, Sydney Nursing School Hespe, Charlotte; The University of Notre Dame Australia, School of Medicine Ferguson, Caleb; Western Sydney University and Western Sydney Local Health District, Western Sydney Nursing and Midwifery Centre Zwar, Nicholas; University of Wollongong Faculty of Science Medicine and Health Lowres, Nicole; Heart Research Institute University of Sydney, Charles Perkins Centre; University of Sydney on September 28, 2021 by guest. Protected copyright. Primary Subject Cardiovascular medicine Heading:

Secondary Subject Heading: Public health, General practice / Family practice, Cardiovascular medicine

atrial fibrillation, screening, oran anticoagulants, electronic decision Keywords: support

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 1 of 23 BMJ Open

1 2 3 Atrial Fibrillation Screen, Management And Guideline Recommended Therapy (AF SMART II) in the 4 rural primary care setting): an implementation study protocol 5 6

7 1 2 1,3 4 8 Authors: Jessica Orchard, * Lis Neubeck, Saul Benedict Freedman, Ruth Webster, Anushka Patel, 9 4 Robyn Gallagher, 5 Jialin Li,5 Charlotte Mary Hespe,6 Caleb Ferguson,7 Nick Zwar8 and Nicole Lowres 10 1,3 11 12 13 14 Affiliations 15 16 1. Sydney MedicalFor School peer and Charles reviewPerkins Centre, University only of Sydney, Sydney, NSW, 17 Australia 18 2. School of Health and Social Care, Edinburgh Napier University, Sighthill, Edinburgh UK 19 3. Heart Research Institute, Sydney, NSW, Australia 20 4. The George Institute for Global Health, University of New South Wales, Sydney, Australia 21 22 5. Sydney Nursing School, University of Sydney, Sydney, NSW, Australia 23 6. General Practice Research, School of Medicine, The University of Notre Dame, Sydney, 24 Australia 25 7. Western Sydney Nursing and Midwifery Centre, Western Sydney University & Western 26 Sydney Local Health District, Sydney Australia 27 28 8. School of Medicine, Faculty of Science, Medicine and Health, University of Wollongong, 29 Wollongong, Australia 30 31 Keywords: atrial fibrillation, screening, stroke, oral anticoagulants, electronic decision support http://bmjopen.bmj.com/ 32 33 34 35 Word count: 2882 words, excluding abstract 36 37 38 *Corresponding author 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 Address for correspondence: 43 44 Jessica Orchard 45 46 Charles Perkins Centre (D17), Level 2 47 University of Sydney NSW 2006 48 49 Ph+ 61 2 8627 1664 50 51 Email: [email protected] 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 2 of 23

1 2 3 ABSTRACT 4 5 Introduction: Screening for atrial fibrillation (AF) in people ≥65 years is now recommended by 6 guidelines and expert consensus. While AF is often asymptomatic, it is the most common heart 7 arrhythmia, and is associated with increased risk of stroke. Early identification and treatment with 8 oral anticoagulants can substantially reduce stroke risk. The general practice setting is ideal for 9 opportunistic screening and provides a natural pathway for treatment for those identified. 10 11 This study aims to investigate the feasibility of implementing screening for AF in rural general 12 practice using novel electronic tools. It will assess whether screening will fit within an existing 13 14 workflow to quickly and accurately identify AF, and will potentially inform a generalisable, scalable 15 approach. 16 For peer review only 17 Methods and analysis: Screening with a smartphone electrocardiogram will be conducted by 18 general practitioners (GPs) and practice nurses in rural general practices in New South Wales, 19 Australia for 3-4 months during 2018-19. Up to 10 practices will be recruited, and we aim to screen 20 2,000 patients aged ≥65 years. Practices will be given an electronic screening prompt and electronic 21 22 decision support to guide evidence-based treatment for those with AF. De-identified data will be 23 collected using a clinical audit tool and qualitative interviews will be conducted with selected 24 practice staff. A process evaluation and cost-effectiveness analysis will also be undertaken. 25 Outcomes include implementation success (proportion of eligible patients screened, fidelity to 26 protocol), proportion of people screened identified with new AF, and rates of treatment with 27 28 anticoagulants and antiplatelets at baseline and completion. Results will be compared against an 29 earlier metropolitan study and a ‘control’ dataset of practices. 30 31 Ethics and dissemination: Ethics approval was received from the University of Sydney Human http://bmjopen.bmj.com/ 32 Research Ethics Committee on 27 February 2018 (Project no.: 2017/1017). Results will be 33 disseminated through various forums, including peer-reviewed publication and conference 34 presentations. 35 36 Trial registration number: ACTRN12618000004268 37 38 ARTICLE SUMMARY: STRENGTHS AND LIMITATIONS OF THIS STUDY 39 on September 28, 2021 by guest. Protected copyright. 40 • New technology is used to quickly and efficiently identify and treat AF, within an existing 41 workflow in the general practice setting 42 • General practices are ideally placed to screen for and diagnose AF, as well to provide a 43 44 pathway to care for those identified during screening 45 • Conducting the study in the rural setting will inform future AF screening policies to extend 46 the evidence base beyond metropolitan settings 47 • A potential limitation is the small size and non-randomised design, which would ideally be 48 overcome in a future large-scale study 49 50

51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 3 of 23 BMJ Open

1 2 3 INTRODUCTION 4 AF is the most common heart arrhythmia, affecting 33.5 million people worldwide.1 A person’s 5 likelihood of developing AF increases with age, reaching 37% for those aged ≥55 years.2 Importantly, 6 3 4 7 AF is associated with a fivefold increase in risk of stroke. However, 1.4%-1.6% of the population 8 aged ≥65 have undiagnosed AF,5 6 which is commonly asymptomatic.7 8 Stroke caused by AF is largely 9 preventable with appropriate oral anticoagulant (OAC) therapy, which can reduce AF-related stroke 10 risk by 64%.9 Although OAC is strongly recommended by clinical practice guidelines for those with 11 additional stroke risk factors, as set out in the CHA₂DS₂-VASc score,10 11 there has historically been a 12 12 13 large gap between evidence and practice with only 60% of eligible patients prescribed OAC. 14 Recently, there are signs of improvement, with a recent Public Health England report noting 77.9% 15 of eligible AF patients in England were treated with OAC.13 Similar increases in OAC prescription, 16 largely due to theFor introduction peer of novel anticoagulants review (NOACs), haveonly been reported in Spain and 17 14 15 18 Denmark. 19 Screening for AF in people ≥65 years by opportunistic pulse palpation or ECG rhythm strip is now 20 11 16 17 21 recommended by clinical practice guidelines and international expert consensus. Many studies 22 have demonstrated the effectiveness of opportunistic screening for AF in general practice. The SAFE 23 study demonstrated that opportunistic screening of patients aged ≥65 years by general practitioners 24 (GPs), with a flag prompt on the patient file, was an effective screening method.5 AF screening in 25 general practice was found to be promising during flu vaccination in the Netherlands,18 and was 26 19 27 successfully performed by practice nurses in the UK. Our previous research showed AF screening 28 by practice nurses to be feasible, both opportunistically and during the influenza (flu) vaccination 29 period.20 21 30 31 With 84% of Australian general practices employing practice nurses,22 there is scope for AF screening http://bmjopen.bmj.com/ 32 to be performed during existing practice nurse health appointments. Ideally, practice nurses could 33 screen patients for AF during the flu or shingles vaccination and/or coordinated chronic care 34 consultations. These appointments are already funded by the Australian Government: the flu 35 23 36 vaccination is recommended annually and provided free for patients aged ≥65 years; shingles 37 vaccines will be provided free for people aged ≥70 with a 5 year catch-up program for people aged 38 71-79;24 annual health assessments are funded for those aged ≥75;25 and annual coordination on September 28, 2021 by guest. Protected copyright. 39 consultations are funded by Medicare for patients with chronic conditions (eg diabetes, 40 26 41 cardiovascular) on a care plan. These are all occasions when patients in the target age group are 42 likely to be available and receptive to AF screening, in an appropriate setting. 43 44 While numerous studies have been conducted in metropolitan areas, there is a paucity of data 45 regarding AF screening in rural areas. A recent systematic review of AF screening studies conducted 46 in rural settings acknowledged the need for more information on AF burden and risk factors 47 specifically in rural areas.27 In the Australian context, it is known that people living in rural areas have 48 worse cardiovascular outcomes.28 It has been reported that there is a lack of evidence about rural 49 50 cardiovascular disease prevention and “characteristics associated with quality of care”, which 51 highlights the need for more programs in rural primary care, especially for high risk patients.29 This 52 study aims to implement a screening program with a quality improvement focus, to better inform 53 future AF screening policies to extend the evidence base beyond metropolitan settings. 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 4 of 23

1 2 3 METHODS AND ANALYSIS 4 5 Design 6 This study, AF-SMART II, will investigate the feasibility of implementation of an AF screening 7 program and use of an AF electronic decision support tool (EDS), in rural general practices. It is a 8 cross-sectional, observational implementation study of AF screening using a smartphone 9 electrocardiogram (ECG) conducted by GPs and practice nurses in up to 10 rural general practices in 10 New South Wales (NSW), Australia. It is very similar to the study protocol used in the AF-SMART 11 12 study (ACTRN12616000850471) which was conducted in 8 metropolitan general practices in Sydney, 13 Australia from 2016-2018 . AF-SMART II is a rural extension of AF-SMART with some important 14 modifications based on feedback obtained in the metropolitan study. These modifications include: 15 (1) an improved electronic screening prompt to improve the proportion of eligible patients screened; 16 (2) the introductionFor GP Continuing peer Professional review Development (CPD) only / Quality Improvement (QI) 17 18 points to provide additional incentives for GPs; (3) a more structured approach to nurse screening; 19 and (4) a shorter screening period. 20 Practice recruitment 21 A convenience sample of up to 10 rural general practices, recruited through several NSW Primary 22 23 Health Networks, will be invited to participate in the study. Practices will be required to: 24 • have at least one practice nurse; 25 26 • use either ‘Best Practice’ or ‘Medical Director’ as the practice electronic patient record 27 management system; 28 • be willing to have ‘TopBar’ (PenCS third party software) and the clinical audit tool ‘PenCAT’ 29 for data collection installed (licences are provided by the Primary Health Networks, New 30 31 South Wales); and http://bmjopen.bmj.com/ 32 • have WiFi. 33 34 Pre-intervention: training and set up electronic study tools 35 Once a practice has been recruited, the pre-intervention phase will involve setting up the electronic 36 study tools and relevant training. Each practice will be provided with several Kardia smartphone 37 electrocardiogram (iECG) devices to screen for AF (Figure 1). This device has approval from the 38 39 Australian Therapeutic Goods Administration as a Medical Device, Class IIa and has a validated, on September 28, 2021 by guest. Protected copyright. 40 automated algorithm for detecting AF with 95% sensitivity and 99% specificity.30 41 42 In addition, practices will be provided with two custom-designed apps that were developed and 43 tested in the AF-SMART study (Error! Reference source not found.) (ACTRN12616000850471). These 44 apps are the: 45 46 (1) AF app with screening prompt, which extracts data from the electronic patient record and 47 provides a prompt for AF screening when an eligible patient’s file is opened, and allows 48 clinical staff to record the provisional screening results (Figure 2 and Figure 3). 49 50 51 (2) Electronic decision support system (EDS), which guides evidence-based treatment of those 52 diagnosed with AF. The EDS is a clinically validated quality improvement tool, designed to 53 bridge gaps between evidence and practice in AF risk management.31 The EDS extracts data 54 from electronic patient records regarding current OAC and antiplatelet prescriptions and 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 5 of 23 BMJ Open

1 2 3 calculates individual stroke risk scores (CHA2DS2-VASc) (Figure 4). It then provides clinical 4 staff with evidence-based medication guidelines to support decisions on OAC prescription. 5 6 The apps are located in the TopBar hosting platform that integrates with the general practice 7 electronic patient record management systems. 8 9 Once the technology setup is complete, practice nurses and GPs will receive training on the use of 10 the electronic tools and iECG as well as recent developments in evidence-based management of AF 11 for stroke prevention, highlighting guideline recommendations for OAC, recommendations against 12 using aspirin to treat AF.11 Practices will be provided with posters for the reception area advertising 13 14 the study and a laminated participant information statement will be available at reception. 15 16 Patient eligibilityFor peer review only 17 All patients aged ≥65 years presenting to the general practice for annual flu vaccination, shingles 18 vaccination, chronic care assessment, or seeing their GP, will be eligible for AF screening if they meet 19 the following criteria: 20 21 • age ≥65 years, 22 • no current recorded diagnosis of AF, and 23 24 • have not already been screened with the iECG in the past 12 months. 25 26 Patients with terminal illness, and patients unable to provide informed consent will be excluded. 27 28 Screening protocol 29 The screening intervention will be conducted for a period of 3-4 months in each practice. The 30 process is shown in Figure 5, and summarised below: 31 http://bmjopen.bmj.com/ 32 • The AF screening prompt will notify clinical staff a patient is eligible for iECG screening. 33 • The GP/nurse will explain the screening process and purpose, and obtain informed oral 34 consent from patients. 35 • The GP/nurse will record a 30 second iECG. An automated interpretation (‘Possible AF’, 36 37 ‘Normal’, or ‘Unclassified’) is provided immediately after successful recording. 38 • The GP/nurse will enter the iECG screening result into the AF app in TopBar. 39 • Depending on the iECG result, follow the appropriate protocol: on September 28, 2021 by guest. Protected copyright. 40 41 Protocol if a patient receives “Possible AF” diagnosis 42 43 The GP/nurse will inform the patient of the diagnosis and give a brief explanation. If the nurse 44 performed the screening, the patient will be referred to their GP. Further investigation and 45 management is at the GP’s discretion. A 12-lead ECG is recommended to provide additional 46 confirmation for all new AF diagnoses, and to add extra leads for AF diagnostic work-up. Following a 47 48 12-lead ECG, for patients with confirmed AF, GPs will be encouraged to use the EDS to review each 49 patient’s CHA₂DS₂-VASc score and commence guideline-recommended treatment to reduce stroke 50 risk. 51 52 Protocol if a patient receives “Normal” diagnosis 53 54 The GP/nurse will inform the patient of the diagnosis. No further action is required. 55 Protocol if a patient receives “Unclassified” diagnosis 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 6 of 23

1 2 3 The GP/nurse will inform the patient of the diagnosis and give a brief explanation based on 4 information provided by researchers. Depending on individual patient’s history and iECG, a 12-lead 5 ECG may be recommended by the GP but follow-up is at the GP’s discretion. There are a number of 6 conditions that can lead to this particular diagnosis (e.g. sinus tachycardia/bradycardia, left bundle 7 8 branch block, right bundle branch block or multiple ectopic beats) which may or may not be clinically 9 significant, or previously known. 10 11 CPD and QI points 12 In order to assist in practice recruitment, education and training will be structured to be eligible for 13 CPD points for nurses and GPs. In addition, all practices will be encouraged to use the research 14 project to conduct an AF quality improvement program that will attract specific QI professional 15 development points. 16 For peer review only 17 Study outcomes 18 Implementation success through process measures including proportion of eligible patients 19 • 20 seen in the practice during the period who are screened (calculated from information 21 available in the deidentified data extracts), fidelity to the protocol, and time taken to 22 complete the intervention. 23 • Proportion of people screened identified with new AF. 24 25 • Proportion of patients where EDS page accessed. 26 • Proportion of eligible patients prescribed OAC as a result of the intervention. 27 • Proportion of patients with diagnosed AF prescribed antiplatelet therapy during the study 28 period. 29 30 • Prevalence of AF at baseline (compared to metropolitan and control group, as defined 31 below). http://bmjopen.bmj.com/ 32 • Incidence of new AF at completion of the intervention (compared to metropolitan and 33 control group). 34 35 • Rates of treatment with anticoagulants and antiplatelets at baseline and at completion of 36 the intervention (compared to metropolitan and control group). 37 • Acceptability, competing demands, barriers, and enablers according to staff involved in the 38 intervention. 39 on September 28, 2021 by guest. Protected copyright. Cost-effectiveness analysis (i.e. incremental cost-effectiveness ratio (ICER) of screening per 40 • 41 quality adjusted life year (QALY) gained, and per stroke avoided), using data from Cadillac et 42 al32 to estimate the present value of QALYs gained for each ischaemic stroke prevented. 43 44 45 Data collection and sample size 46 The study aims to screen approximately 2,000 patients in total from up to 10 general practices. 47 Assuming an incidence of 1.4% of people aged ≥65 with unknown AF,33 we estimate 28 new cases of 48 AF would be identified. A sample size of 2,000, assuming 1.4% incidence, would provide a 95% 49 confidence interval of 0.93–2.02% (18-40 cases). 50 51 PenCAT will be configured to collect de-identified data from electronic patient records. These data 52 include demographic, medication, and diagnostic information. Data extracts will be taken at 53 54 baseline, end of month 1, end of month 2 and end of month 3. The screening period will be 55 extended if required so that practices get all 4 data extracts. Following each data extract, feedback 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 7 of 23 BMJ Open

1 2 3 will be provided to practices including a summary of results (number screened by each staff member 4 in total and for that month). 5 6 Following completion of the iECG screening period, selected staff from each practice will be 7 interviewed to gather their feedback as part of a comprehensive process evaluation. Semi-structured 8 interviews with selected GPs, practice nurses, and practice managers will be carried out by 9 researchers at the end of the screening intervention. The interviews will be audio-recorded and 10 11 transcribed. 12 13 Statistical analysis 14 Quantitative analysis 15 Descriptive analyses will be carried out both at the individual practice level and all practices pooled 16 together. Data collectedFor in thispeer study will be review compared with data onlycollected from the earlier study in 17 metropolitan Sydney, and a database of ‘control’ practices with the same data-parameters collected 18 through concurrent studies run by The George Institute. Analyses of these data will be conducted 19 2 20 using SAS, with χ tests used to test associations between AF incidence and method of detection 21 (screen-detected or clinically-detected), and two-tail p values <0.05 will be considered significant. 22 23 Cost-effectiveness analysis 24 A basic economic model for AF screening has previously been developed and utilised in the SEARCH- 25 AF34 study. This model will be adapted and extended to evaluate screening for AF by the practice 26 nurse in the general practice. A modelled cost-effectiveness analysis from an Australian health 27 funder perspective will be performed comparing the cost of iECG population-based screening for AF, 28 29 to diagnosed AF in an unscreened population of Australian men and women aged 65 to 84 years. 30 Future costs will be discounted at a rate of 5%. The results will be expressed as an ICER per stroke 31 avoided and per QALY gained. http://bmjopen.bmj.com/ 32 33 Sensitivity analyses will be performed, including varying the base assumptions of a guideline- 34 adherence rate for OAC prescription from 40% to 90% and increasing the cost of treatment. We will 35 obtain 95% confidence intervals (CIs) for all ICERs using a multifactorial Monte Carlo simulation35 as 36 34 37 described in SEARCH-AF. 38

Qualitative analysis on September 28, 2021 by guest. Protected copyright. 39 A detailed process evaluation using mixed methods will be undertaken to evaluate the iECG 40 36 41 screening process. Realist evaluation will be used to analyse the interaction of context, 42 mechanisms for change and the outcomes that are produced in both this study and the metropolitan 43 component. In this way we will piece together a detailed understanding of how the intervention 44 influenced the capacity, opportunity and motivation of providers to improve detection and 45 46 evidence-based treatment of AF. 47 48 Patient involvement 49 Development and refinement of the protocol was an iterative process. Feedback obtained through 50 interviews patients, nurses, GPs and practice managers about their experience and preferences 20 30 51 during previous AF screening studies, contributed to the development of the protocol. 52 Following the study, participating practices will be provided with a summary of the findings of the 53 study, together with a full copy of any publications, which will be made available to patients in the 54 practice reception area. 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 8 of 23

1 2 3 ETHICS AND DISSEMINATION 4 The study will comply with the National Health and Medical Research Council (NHMRC) ethical 5 guidelines. Approval for the study was received from the University of Sydney Human Research 6 7 Ethics Committee on 27 February 2018 (Project no.: 2017/017). Participating practices will provide 8 written, informed consent and patients being screened will provide oral consent. We used the 9 Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist when writing 10 our protocol report.37 The study will be administered by the Charles Perkins Centre, University of 11 Sydney. Results of the study will be disseminated through various forums, including peer-reviewed 12 13 publications, and presentation at national and international conferences. 14 15 CONCLUSIONS 16 Screening for AF inFor people ≥65 peer years is now reviewrecommended by guidelines only and international expert 17 consensus.11 16 Importantly, a comprehensive screening program should include a system to support 18 16 19 evidence-based treatment for patients ultimately diagnosed with the condition. General practices 20 are well-placed to opportunistically screen older patients, and are ideal in terms of the pathway for 21 treatment for those identified. This study will provide electronic prompts for screening and test a 22 patient-level EDS integrated with the practice’s clinical software, to support evidence-based 23 24 treatment. 25 In previous studies, a range of evidence-based interventions have been developed to increase 26 27 effective prescribing of guideline-recommended OAC in primary care settings. These include use of 38 39 28 electronic decision support (EDS) tools, targeted GP-education programs, consultant-led primary 29 care anticoagulation assessment clinics,40 and patient focused education interventions.41 Overall, 30 these interventions increase quality prescribing, but the results are varied in terms of effect size and 31 http://bmjopen.bmj.com/ duration. One study with a large effect size used a consultant-led anticoagulation assessment within 32 33 the general practice to identify and review AF patients with a CHA₂DS₂-VASc score ≥1 who were 40 34 “suboptimally anticoagulated”. This study reported a substantial and significant increase in 35 appropriate OAC treatment from 77% to 95% (p<0.0001), suggesting that an intensive, systematic 36 approach, together with numerous reminders to improve patient attendance at in-person reviews, 37 can be very successful in increasing OAC uptake.40 In contrast, a cluster-randomised study providing 38 39 doctors with patient level recommendations via an EDS did not show significant improvements in on September 28, 2021 by guest. Protected copyright. 40 evidence-based OAC treatment.38 41 42 This study aims to investigate the feasibility of implementing screening for AF in rural general 43 practice. It follows on from the first phase of the study conducted in metropolitan general practice, 44 allowing comparison against a metropolitan setting and a control group of standard practices. The 45 project will assess if screening will efficiently fit within an existing workflow to quickly and accurately 46 47 identify AF. The study will also explore the effect of GP and practice nurse education, electronic 48 screening prompts and EDS to facilitate evidence-based treatment. 49 50 In Australia, the Heart Foundation’s Guideline for the Diagnosis and Management of Atrial 51 Fibrillation will be released in August 2018, the draft guideline recommends opportunistic screening 52 for AF in general practice for people aged ≥65 years. This study will potentially have important 53 implications for the implementation of a large scale opportunistic AF screening program for those 54 aged ≥65 years. It could also inform the design of a future randomised trial of AF screening. It is 55 56 anticipated that the study will demonstrate that the general practice setting, provided with 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 9 of 23 BMJ Open

1 2 3 additional electronic tools, can meet the challenges of large-scale opportunistic screening with a 4 cheap, convenient, scalable solution. 5 6 7 8 Figure legends 9 Figure 1 - Kardia device and iECG reading 10 11 Figure 2 – Screenshot showing test patient file and TopBar apps 12 13 Figure 3 – entering an iECG result in TopBar AF app (test patient shown) 14 15 Figure 4 – Screenshot showing EDS recommendation for test patient 16 For peer review only 17 Figure 5 - Process flow of AF screening 18 19

20 21 Author contributions: LN, SBF, NL and JO conceived the original concept of the study. JO, LN, NL, 22 SBF, CMH CF, RG and NZ were involved in the design of the study. LN provided qualitative analysis 23 24 expertise and NL provided cost-effectiveness analysis expertise. RW and AP were involved in the 25 design and testing of the electronic study apps and provided statistical expertise. JO, LN, SBF, JL and 26 NL were involved in drafting the article. CMH, and NZ provided input on general practice aspects of 27 study design, including in relation to educational materials for practices. JO, LN, SBF, RW, AP, RG, JL, 28 29 CF, NZ and NL contributed to the critical revision of the drafting and final approval of the manuscript. 30

Funding statement: This work was supported by a National Heart Foundation of Australia/ New http://bmjopen.bmj.com/ 31 32 South Wales Health Cardiovascular Research Network Research Development Project Grant 33 (101133) with top-up funding to complete the study in rural NSW through an investigator-initiated 34 grant from Pfizer. The funders have no influence on study design, data collection or the decision to 35 submit reports for publication. AliveCor have provided free Kardia devices for study purposes. JO is 36 supported by an Australian Government Research Training Program (RTP) Scholarship. RW is funded 37 38 by a NHMRC Early Career Fellowship (APP1125044). NL is funded by a NSW Health Early Career 39 Fellowship (H16/ 52168). on September 28, 2021 by guest. Protected copyright. 40 41 Competing interests statement: JO, JL, CH, RG and NZ have no disclosures. NL and SBF report grants 42 from BMS/Pfizer during the conduct of the study. RW and AP report other financial interests from 43 George Health, outside the submitted work. CF reports personal fees from Pfizer outside the 44 submitted work. LN reports grants from Pfizer/BMS and Boehringer Ingelheim outside the submitted 45 46 work. BF reports prior fees and advisory board honoraria from Bayer Pharma AG, Boehringer 47 Ingelheim, and BMS/Pfizer but for the past year has removed himself from pharmaceutical advisory 48 boards and receives speaker fees only for accredited educational meetings. 49 50 Data sharing statement: Data will not be made available as data sharing is not permitted by our 51 ethics approval. 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 10 of 23

1 2 3 16. Freedman B, Camm J, Calkins H, et al. Screening for Atrial Fibrillation: A Report of the AF-SCREEN 4 International Collaboration. Circulation 2017;135(19):1851-67. doi: 5 10.1161/CIRCULATIONAHA.116.026693 [published Online First: 2017/05/10] 6 17. Mairesse GH, Moran P, Van Gelder IC, et al. Screening for atrial fibrillation: a European Heart 7 Rhythm Association (EHRA) consensus document endorsed by the Heart Rhythm Society 8 (HRS), Asia Pacific Heart Rhythm Society (APHRS), and Sociedad Latinoamericana de 9 Estimulacion Cardiaca y Electrofisiologia (SOLAECE). Europace 2017;19(10):1589-623. doi: 10 10.1093/europace/eux177 [published Online First: 2017/10/20] 11 18. Tieleman RG, Plantinga Y, Rinkes D, et al. Validation and clinical use of a novel diagnostic device 12 for screening of atrial fibrillation. Europace 2014;16(9):1291-5. doi: 13 10.1093/europace/euu057 [published Online First: 2014/05/16] 14 19. Morgan S, Mant D. Randomised trial of two approaches to screening for atrial fibrillation in UK 15 general practice. The British Journal of General Practice 2002;52(478):373-80. 16 20. Orchard J, FreedmanFor SB, peerLowres N, et al. review iPhone ECG screening only by practice nurses and 17 receptionists for atrial fibrillation in general practice: the GP-SEARCH qualitative pilot study. 18 19 Aust Fam Physician 2014;43(5):315-9. [published Online First: 2014/05/06] 20 21. Orchard J, Lowres N, Ben Freedman S, et al. Screening for atrial fibrillation during influenza 21 vaccinations by primary care nurses using a smartphone electrocardiograph (iECG): A 22 feasibility study. European Journal of Preventive Cardiology 2016;23(Supp 2):13-20. doi: 23 10.1177/2047487316670255 24 22. Britt H, Miller GC, Henderson J, et al. General Practice Activity in Australia 2015-16: BEACH: 25 Bettering the Evaluation and Care of Health. Sydney: Sydney University Press 2016. 26 23. Media Release, Prime Minister and Minister for Health. Ground-breaking flu vaccines to protect 27 millions of Aussies 19 February 2018. 28 http://www.health.gov.au/internet/ministers/publishing.nsf/Content/health-mediarel- 29 yr2018-hunt020.htm, accessed 9 March 2018 30 24. NSW Government. Shingles Vaccination Program 2018. 31 http://wwwhealthnswgovau/immunisation/Pages/Shingles-programaspx, accessed 9 March http://bmjopen.bmj.com/ 32 2018 33 25. Australian Government Department of Health. Health assessment for people aged 75 years and 34 older (2014). 35 http://wwwhealthgovau/internet/main/publishingnsf/Content/mbsprimarycare_mbsitem_7 36 5andolder accessed 19 March 2018 37 26. Australian Government Department of Health. Chronic Disease Management - Provider 38 39 Information 2016. on September 28, 2021 by guest. Protected copyright. 40 http://wwwhealthgovau/internet/main/publishingnsf/content/mbsprimarycare-factsheet- 41 chronicdiseasehtm, accessed 9 March 2018 42 27. Gavino AI, McLachlan CS. Review of screening studies for atrial fibrillation in rural populations of 43 11 countries. Proc (Bayl Univ Med Cent) 2017;30(3):280-85. 44 28. Australian Institute of Health and Welfare (AIHW). Cardiovascular Medicines and Primary Health 45 Care: A Regional Analysis. Canberra, 2010. 46 29. Allenby A, Kinsman L, Tham R, et al. The quality of cardiovascular disease prevention in rural 47 primary care. Aust J Rural Health 2016;24(2):92-8. doi: 10.1111/ajr.12224 [published Online 48 First: 2015/08/11] 49 30. Orchard J, Lowres N, Freedman SB, et al. Screening for atrial fibrillation during influenza 50 vaccinations by primary care nurses using a smartphone electrocardiograph (iECG): A 51 feasibility study. Eur J Prev Cardiol 2016;23(2 suppl):13-20. doi: 10.1177/2047487316670255 52 [published Online First: 2016/11/29] 53 31. Peiris D, Usherwood T, Panaretto K, et al. The Treatment of cardiovascular Risk in Primary care 54 using Electronic Decision supOrt (TORPEDO) study-intervention development and protocol 55 for a cluster randomised, controlled trial of an electronic decision support and quality 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 12 of 23

1 2 3 improvement intervention in Australian primary healthcare. BMJ Open 2012;2(6):e002177. 4 doi: 10.1136/bmjopen-2012-002177 [published Online First: 2012/11/21] 5 32. Cadilhac DA, Dewey HM, Vos T, et al. The health loss from ischemic stroke and intracerebral 6 hemorrhage: evidence from the North East Melbourne Stroke Incidence Study (NEMESIS). 7 Health Qual Life Outcomes 2010;8:49. doi: 10.1186/1477-7525-8-49 [published Online First: 8 2010/05/18] 9 33. Lowres N, Neubeck L, Redfern J, et al. Community screening programs to identify unknown atrial 10 fibrillation: a systematic review. Eur Heart J 2012;33:61-61. doi: 11 http://dx.doi.org/10.1016/j.hlc.2012.05.018 12 34. Lowres N, Neubeck L, Salkeld G, et al. Feasibility and cost-effectiveness of stroke prevention 13 through community screening for atrial fibrillation using iPhone ECG in pharmacies. The 14 SEARCH-AF study. Thromb Haemost 2014;111(6):1167-76. doi: 10.1160/TH14-03-0231 15 [published Online First: 2014/04/02] 16 35. Briggs AH. StatisticalFor approaches peer to handling review uncertainty in health only economic evaluation. Eur J 17 Gastroenterol Hepatol 2004;16(6):551-61. 18 19 36. Pawson R, Tilley N. Realistic Evaluation. London, UK: Sage Publications 1997. 20 37. Chan AW, Tetzlaff JM, Altman DG, et al. SPIRIT 2013 statement: defining standard protocol items 21 for clinical trials. Ann Intern Med 2013;158(3):200-7. doi: 10.7326/0003-4819-158-3- 22 201302050-00583 [published Online First: 2013/01/09] 23 38. Eckman MH, Lip GY, Wise RE, et al. Impact of an Atrial Fibrillation Decision Support Tool on 24 thromboprophylaxis for atrial fibrillation. Am Heart J 2016;176:17-27. doi: 25 10.1016/j.ahj.2016.02.009 [published Online First: 2016/06/07] 26 39. Jackson SL, Peterson GM, Vial JH. A community-based educational intervention to improve 27 antithrombotic drug use in atrial fibrillation. Ann Pharmacother 2004;38(11):1794-9. doi: 28 10.1345/aph.1E152 [published Online First: 2004/09/30] 29 40. Das M, Panter L, Wynn GJ, et al. Primary Care Atrial Fibrillation Service: outcomes from 30 consultant-led anticoagulation assessment clinics in the primary care setting in the UK. BMJ 31 Open 2015;5(12):e009267. doi: 10.1136/bmjopen-2015-009267 [published Online First: http://bmjopen.bmj.com/ 32 2015/12/15] 33 41. McAlister FA, Man-Son-Hing M, Straus SE, et al. Impact of a patient decision aid on care among 34 patients with nonvalvular atrial fibrillation: a cluster randomized trial. CMAJ 35 2005;173(5):496-501. doi: 10.1503/cmaj.050091 [published Online First: 2005/09/01] 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 13 of 23 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 Figure 1 - Kardia device and iECG reading 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 14 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 Figure 2 - Screenshot showing test patient file and TopBar apps 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 15 of 23 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 Figure 3 – Entering an iECG result in TopBar AF app (test patient shown) 20 30x14mm (300 x 300 DPI) 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 16 of 23

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 Figure 4 – Screenshot showing EDS recommendation for test patient 28 29 20x13mm (600 x 600 DPI) 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 17 of 23 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 Figure 5 - Process flow of AF screening 46 47 46x68mm (600 x 600 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 18 of 23

Atrial Fibrillation Screen, Management And Guideline Recommended Therapy (AF SMART II) in the rural primary care setting): an implementation study protocol

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2018-023130.R2

Article Type: Protocol

Date Submitted by the Author: 25-Jul-2018

Secondary Subject Heading: Public health, General practice / Family practice, Cardiovascular medicine

atrial fibrillation, screening, oran anticoagulants, electronic decision Keywords: support

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 1 of 24 BMJ Open

1 2 3 Atrial Fibrillation Screen, Management And Guideline Recommended Therapy (AF SMART II) in the 4 rural primary care setting): an implementation study protocol 5 6

1 2 3 INTRODUCTION 4 AF is the most common heart arrhythmia, affecting 33.5 million people worldwide.1 A person’s 5 likelihood of developing AF increases with age, reaching 37% for those aged ≥55 years.2 Importantly, 6 3 4 7 AF is associated with a fivefold increase in risk of stroke. However, 1.4%-1.6% of the population 8 aged ≥65 have undiagnosed AF,5 6 which is commonly asymptomatic.7 8 Stroke caused by AF is largely 9 preventable with appropriate oral anticoagulant (OAC) therapy, which can reduce AF-related stroke 10 risk by 64%.9 Although OAC is strongly recommended by clinical practice guidelines for those with 11 additional stroke risk factors, as set out in the CHA₂DS₂-VASc score,10 11 there has historically been a 12 12 13 large gap between evidence and practice with only 60% of eligible patients prescribed OAC. 14 Recently, there are signs of improvement, with a recent Public Health England report noting 77.9% 15 of eligible AF patients in England were treated with OAC.13 Similar increases in OAC prescription, 16 largely due to theFor introduction peer of novel anticoagulants review (NOACs), haveonly been reported in Spain and 17 14 15 18 Denmark. 19 Screening for AF in people ≥65 years by opportunistic pulse palpation or ECG rhythm strip is now 20 11 16 17 21 recommended by clinical practice guidelines and international expert consensus. Many studies 22 have demonstrated the effectiveness of opportunistic screening for AF in general practice. The SAFE 23 study demonstrated that opportunistic screening of patients aged ≥65 years by general practitioners 24 (GPs), with a flag prompt on the patient file, was an effective screening method.5 AF screening in 25 general practice was found to be promising during flu vaccination in the Netherlands,18 and was 26 19 27 successfully performed by practice nurses in the UK. Our previous research showed AF screening 28 by practice nurses to be feasible, both opportunistically and during the influenza (flu) vaccination 29 period.20 21 30 31 With 84% of Australian general practices employing practice nurses,22 there is scope for AF screening http://bmjopen.bmj.com/ 32 to be performed during existing practice nurse health appointments. Ideally, practice nurses could 33 screen patients for AF during the flu or shingles vaccination and/or coordinated chronic care 34 consultations. These appointments are already funded by the Australian Government: the flu 35 23 36 vaccination is recommended annually and provided free for patients aged ≥65 years; shingles 37 vaccines will be provided free for people aged ≥70 with a 5 year catch-up program for people aged 38 71-79;24 annual health assessments are funded for those aged ≥75;25 and annual coordination on September 28, 2021 by guest. Protected copyright. 39 consultations are funded by Medicare for patients with chronic conditions (eg diabetes, 40 26 41 cardiovascular) on a care plan. These are all occasions when patients in the target age group are 42 likely to be available and receptive to AF screening, in an appropriate setting. 43 44 While numerous studies have been conducted in metropolitan areas, there is a paucity of data 45 regarding AF screening in rural areas. A recent systematic review of AF screening studies conducted 46 in rural settings acknowledged the need for more information on AF burden and risk factors 47 specifically in rural areas.27 In the Australian context, it is known that people living in rural areas have 48 worse cardiovascular outcomes.28 It has been reported that there is a lack of evidence about rural 49 50 cardiovascular disease prevention and “characteristics associated with quality of care”, which 51 highlights the need for more programs in rural primary care, especially for high risk patients.29 This 52 study aims to implement a screening program with a quality improvement focus, to better inform 53 future AF screening policies to extend the evidence base beyond metropolitan settings. 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 4 of 24

1 2 3 calculates individual stroke risk scores (CHA2DS2-VASc) (Figure 4). It then provides clinical 4 staff with evidence-based medication guidelines to support decisions on OAC prescription. 5 6 The apps are located in the TopBar hosting platform that integrates with the general practice 7 electronic patient record management systems. 8 9 Once the technology setup is complete, practice nurses and GPs will receive training on the use of 10 the electronic tools and iECG as well as recent developments in evidence-based management of AF 11 for stroke prevention, highlighting guideline recommendations for OAC, recommendations against 12 using aspirin to treat AF.11 Practices will be provided with posters for the reception area advertising 13 14 the study and a laminated participant information statement will be available at reception. 15 16 Patient eligibilityFor peer review only 17 All patients aged ≥65 years presenting to the general practice for annual flu vaccination, shingles 18 vaccination, chronic care assessment, or seeing their GP, will be eligible for AF screening if they meet 19 the following criteria: 20 21 • age ≥65 years, 22 • no current recorded diagnosis of AF, and 23 24 • have not already been screened with the iECG in the past 12 months. 25 26 Patients with terminal illness, and patients unable to provide informed consent will be excluded. 27 28 Screening protocol 29 The screening intervention will be conducted for a period of 3-4 months in each practice. The 30 process is shown in Figure 5, and summarised below: 31 http://bmjopen.bmj.com/ 32 • The AF screening prompt will notify clinical staff a patient is eligible for iECG screening. 33 • The GP/nurse will explain the screening process and purpose, and obtain informed oral 34 consent from patients. 35 • The GP/nurse will record a 30 second iECG. The iECG trace will be visible in real-time on the 36 37 phone, and a pdf will also be available to download from the secure Kardia website. An 38 automated interpretation (‘Possible AF’, ‘Normal’, or ‘Unclassified’) is provided immediately 39 after successful recording. on September 28, 2021 by guest. Protected copyright. 40 • The GP/nurse will enter the iECG screening result into the AF app in TopBar. 41 • Depending on the iECG result, follow the appropriate protocol: 42 43 Protocol if a patient receives “Possible AF” diagnosis 44 45 The GP/nurse will inform the patient of the diagnosis and give a brief explanation. If the nurse 46 performed the screening, the patient will be referred to their GP. Further investigation and 47 48 management is at the GP’s discretion, which could include referral to a specialist. A 12-lead ECG is 49 recommended to provide additional confirmation for all new AF diagnoses, and to add extra leads 50 for AF diagnostic work-up. In some cases of paroxysmal AF, the 12-lead ECG may show sinus rhythm 51 while the iECG showed AF. This will be obvious from a comparison of the p-waves on the lead I iECG 52 and lead I of the 12-lead ECG, as well as the regularity of the rhythm. Following a 12-lead ECG, for 53 54 patients with confirmed AF, GPs will be encouraged to use the EDS to review each patient’s 55 CHA₂DS₂-VASc score and commence guideline-recommended treatment to reduce stroke risk. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 6 of 24

1 2 3 Protocol if a patient receives “Normal” diagnosis 4 5 The GP/nurse will inform the patient of the diagnosis. No further action is required. 6 Protocol if a patient receives “Unclassified” diagnosis 7 8 The GP/nurse will inform the patient of the diagnosis and give a brief explanation based on 9 information provided by researchers. Depending on individual patient’s history and iECG, a 12-lead 10 11 ECG may be recommended by the GP but follow-up is at the GP’s discretion. There are a number of 12 conditions that can lead to this particular diagnosis (e.g. sinus tachycardia/bradycardia, left bundle 13 branch block, right bundle branch block or multiple ectopic beats) which may or may not be clinically 14 significant, or previously known. 15 16 CPD and QI pointsFor peer review only 17 In order to assist in practice recruitment, education and training will be structured to be eligible for 18 CPD points for nurses and GPs. In addition, all practices will be encouraged to use the research 19 20 project to conduct an AF quality improvement program that will attract specific QI professional 21 development points. 22 23 Study outcomes 24 • Implementation success through process measures including proportion of eligible patients 25 seen in the practice during the period who are screened (calculated from information 26 available in the deidentified data extracts), fidelity to the protocol, and time taken to 27 28 complete the intervention. 29 • Proportion of people screened identified with new AF. 30 • Proportion of patients where EDS page accessed. http://bmjopen.bmj.com/ 31 • Proportion of eligible patients prescribed OAC as a result of the intervention. 32 33 • Proportion of patients with diagnosed AF prescribed antiplatelet therapy during the study 34 period. 35 • Prevalence of AF at baseline (compared to metropolitan and control group, as defined 36 below). 37 38 • Incidence of new AF at completion of the intervention (compared to metropolitan and 39 control group). on September 28, 2021 by guest. Protected copyright. 40 • Rates of treatment with anticoagulants and antiplatelets at baseline and at completion of 41 the intervention (compared to metropolitan and control group). 42 • Acceptability, competing demands, barriers, and enablers according to staff involved in the 43 44 intervention. 45 • Cost-effectiveness analysis (i.e. incremental cost-effectiveness ratio (ICER) of screening per 46 quality adjusted life year (QALY) gained, and per stroke avoided), using data from Cadillac et 47 al32 to estimate the present value of QALYs gained for each ischaemic stroke prevented. 48 49 50 Data collection and sample size 51 The study aims to screen approximately 2,000 patients in total from up to 10 general practices. 52 Assuming an incidence of 1.4% of people aged ≥65 with unknown AF,33 we estimate 28 new cases of 53 AF would be identified. A sample size of 2,000, assuming 1.4% incidence, would provide a 95% 54 55 confidence interval of 0.93–2.02% (18-40 cases). 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 7 of 24 BMJ Open

1 2 3 PenCAT will be configured to collect de-identified data from electronic patient records. These data 4 include demographic, medication, and diagnostic information. Data extracts will be taken at 5 baseline, end of month 1, end of month 2 and end of month 3. The screening period will be 6 extended if required so that practices get all 4 data extracts. Following each data extract, feedback 7 8 will be provided to practices including a summary of results (number screened by each staff member 9 in total and for that month). 10 11 Following completion of the iECG screening period, selected staff from each practice will be 12 interviewed to gather their feedback as part of a comprehensive process evaluation. Semi-structured 13 interviews with selected GPs, practice nurses, and practice managers will be carried out by 14 researchers at the end of the screening intervention. The interviews will be audio-recorded and 15 transcribed. 16 For peer review only 17 Statistical analysis 18 19 Quantitative analysis 20 Descriptive analyses will be carried out both at the individual practice level and all practices pooled 21 together. Data collected in this study will be compared with data collected from the earlier study in 22 metropolitan Sydney, and a database of ‘control’ practices with the same data-parameters collected 23 through concurrent studies run by The George Institute. Analyses of these data will be conducted 24 2 25 using SAS, with χ tests used to test associations between AF incidence and method of detection 26 (screen-detected or clinically-detected), and two-tail p values <0.05 will be considered significant. 27 28 Cost-effectiveness analysis 29 A basic economic model for AF screening has previously been developed and utilised in the SEARCH- 30 AF34 study. This model will be adapted and extended to evaluate screening for AF by the practice 31 nurse in the general practice. A modelled cost-effectiveness analysis from an Australian health http://bmjopen.bmj.com/ 32 funder perspective will be performed comparing the cost of iECG population-based screening for AF, 33 34 to diagnosed AF in an unscreened population of Australian men and women aged 65 to 84 years. 35 Future costs will be discounted at a rate of 5%. The results will be expressed as an ICER per stroke 36 avoided and per QALY gained. 37 38 Sensitivity analyses will be performed, including varying the base assumptions of a guideline- 39 adherence rate for OAC prescription from 40% to 90% and increasing the cost of treatment. We will on September 28, 2021 by guest. Protected copyright. 40 obtain 95% confidence intervals (CIs) for all ICERs using a multifactorial Monte Carlo simulation35 as 41 described in SEARCH-AF.34 42 43 Qualitative analysis 44 A detailed process evaluation using mixed methods will be undertaken to evaluate the iECG 45 36 46 screening process. Realist evaluation will be used to analyse the interaction of context, 47 mechanisms for change and the outcomes that are produced in both this study and the metropolitan 48 component. In this way we will piece together a detailed understanding of how the intervention 49 influenced the capacity, opportunity and motivation of providers to improve detection and 50 evidence-based treatment of AF. 51 52 Patient involvement 53 Development and refinement of the protocol was an iterative process. Feedback obtained through 54 interviews patients, nurses, GPs and practice managers about their experience and preferences 55 20 30 56 during previous AF screening studies, contributed to the development of the protocol. 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 8 of 24

1 2 3 Following the study, participating practices will be provided with a summary of the findings of the 4 study, together with a full copy of any publications, which will be made available to patients in the 5 practice reception area. 6 7 8 ETHICS AND DISSEMINATION 9 The study will comply with the National Health and Medical Research Council (NHMRC) ethical 10 guidelines. Approval for the study was received from the University of Sydney Human Research 11 Ethics Committee on 27 February 2018 (Project no.: 2017/017). Participating practices will provide 12 written, informed consent and patients being screened will provide oral consent. We used the 13 Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) checklist when writing 14 37 15 our protocol report. The study will be administered by the Charles Perkins Centre, University of 16 Sydney. Results ofFor the study peerwill be disseminated review through various only forums, including peer-reviewed 17 publications, and presentation at national and international conferences. 18 19 20 CONCLUSIONS 21 Screening for AF in people ≥65 years is now recommended by guidelines and international expert 11 16 22 consensus. Importantly, a comprehensive screening program should include a system to support 23 evidence-based treatment for patients ultimately diagnosed with the condition.16 General practices 24 are well-placed to opportunistically screen older patients, and are ideal in terms of the pathway for 25 treatment for those identified. This study will provide electronic prompts for screening and test a 26 27 patient-level EDS integrated with the practice’s clinical software, to support evidence-based 28 treatment. 29 30 In previous studies, a range of evidence-based interventions have been developed to increase 31 effective prescribing of guideline-recommended OAC in primary care settings. These include use of http://bmjopen.bmj.com/ 32 electronic decision support (EDS) tools,38 targeted GP-education programs,39 consultant-led primary 33 care anticoagulation assessment clinics,40 and patient focused education interventions.41 Overall, 34 35 these interventions increase quality prescribing, but the results are varied in terms of effect size and 36 duration. One study with a large effect size used a consultant-led anticoagulation assessment within 37 the general practice to identify and review AF patients with a CHA₂DS₂-VASc score ≥1 who were 38 “suboptimally anticoagulated”.40 This study reported a substantial and significant increase in on September 28, 2021 by guest. Protected copyright. 39 appropriate OAC treatment from 77% to 95% (p<0.0001), suggesting that an intensive, systematic 40 41 approach, together with numerous reminders to improve patient attendance at in-person reviews, 40 42 can be very successful in increasing OAC uptake. In contrast, a cluster-randomised study providing 43 doctors with patient level recommendations via an EDS did not show significant improvements in 44 evidence-based OAC treatment.38 45 46 This study aims to investigate the feasibility of implementing screening for AF in rural general 47 practice. It follows on from the first phase of the study conducted in metropolitan general practice, 48 allowing comparison against a metropolitan setting and a control group of standard practices. The 49 50 project will assess if screening will efficiently fit within an existing workflow to quickly and accurately 51 identify AF. The study will also explore the effect of GP and practice nurse education, electronic 52 screening prompts and EDS to facilitate evidence-based treatment. 53 54 In Australia, the Heart Foundation’s Guideline for the Diagnosis and Management of Atrial 55 Fibrillation will be released in August 2018, the draft guideline recommends opportunistic screening 56 for AF in general practice for people aged ≥65 years. This study will potentially have important 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 9 of 24 BMJ Open

1 2 3 implications for the implementation of a large scale opportunistic AF screening program for those 4 aged ≥65 years. It could also inform the design of a future randomised trial of AF screening. It is 5 anticipated that the study will demonstrate that the general practice setting, provided with 6 additional electronic tools, can meet the challenges of large-scale opportunistic screening with a 7 8 cheap, convenient, scalable solution. 9

10 11 Figure legends 12 13 Figure 1 - Kardia device and iECG reading 14 15 Figure 2 – Screenshot showing test patient file and TopBar apps 16 For peer review only 17 Figure 3 – entering an iECG result in TopBar AF app (test patient shown) 18 19 Figure 4 – Screenshot showing EDS recommendation for test patient 20 21 Figure 5 - Process flow of AF screening 22 23 24 25 Author contributions: LN, SBF, NL and JO conceived the original concept of the study. JO, LN, NL, 26 SBF, CMH CF, RG and NZ were involved in the design of the study. LN provided qualitative analysis 27 expertise and NL provided cost-effectiveness analysis expertise. RW and AP were involved in the 28 29 design and testing of the electronic study apps and provided statistical expertise. JO, LN, SBF, JL and 30 NL were involved in drafting the article. CMH, and NZ provided input on general practice aspects of 31 study design, including in relation to educational materials for practices. JO, LN, SBF, RW, AP, RG, JL, http://bmjopen.bmj.com/ 32 CF, NZ and NL contributed to the critical revision of the drafting and final approval of the manuscript. 33 34 Funding statement: This work was supported by a National Heart Foundation of Australia/ New 35 South Wales Health Cardiovascular Research Network Research Development Project Grant 36 (101133) with top-up funding to complete the study in rural NSW through an investigator-initiated 37 38 grant from Pfizer. The funders have no influence on study design, data collection or the decision to 39 submit reports for publication. AliveCor have provided free Kardia devices for study purposes. JO is on September 28, 2021 by guest. Protected copyright. 40 supported by an Australian Government Research Training Program (RTP) Scholarship. RW is funded 41 by a NHMRC Early Career Fellowship (APP1125044). NL is funded by a NSW Health Early Career 42 Fellowship (H16/ 52168). 43 44 Competing interests statement: JO, JL, CH, RG and NZ have no disclosures. NL and SBF report grants 45 46 from BMS/Pfizer during the conduct of the study. RW and AP report other financial interests from 47 George Health, outside the submitted work. CF reports personal fees from Pfizer outside the 48 submitted work. LN reports grants from Pfizer/BMS and Boehringer Ingelheim outside the submitted 49 work. BF reports prior fees and advisory board honoraria from Bayer Pharma AG, Boehringer 50 Ingelheim, and BMS/Pfizer but for the past year has removed himself from pharmaceutical advisory 51 52 boards and receives speaker fees only for accredited educational meetings. 53 Data sharing statement: Data will not be made available as data sharing is not permitted by our 54 55 ethics approval. 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from BMJ Open Page 10 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 11 of 24 BMJ Open

1 2 3 16. Freedman B, Camm J, Calkins H, et al. Screening for Atrial Fibrillation: A Report of the AF-SCREEN 4 International Collaboration. Circulation 2017;135(19):1851-67. doi: 5 10.1161/CIRCULATIONAHA.116.026693 [published Online First: 2017/05/10] 6 17. Mairesse GH, Moran P, Van Gelder IC, et al. Screening for atrial fibrillation: a European Heart 7 Rhythm Association (EHRA) consensus document endorsed by the Heart Rhythm Society 8 (HRS), Asia Pacific Heart Rhythm Society (APHRS), and Sociedad Latinoamericana de 9 Estimulacion Cardiaca y Electrofisiologia (SOLAECE). Europace 2017;19(10):1589-623. doi: 10 10.1093/europace/eux177 [published Online First: 2017/10/20] 11 18. Tieleman RG, Plantinga Y, Rinkes D, et al. Validation and clinical use of a novel diagnostic device 12 for screening of atrial fibrillation. Europace 2014;16(9):1291-5. doi: 13 10.1093/europace/euu057 [published Online First: 2014/05/16] 14 19. Morgan S, Mant D. Randomised trial of two approaches to screening for atrial fibrillation in UK 15 general practice. The British Journal of General Practice 2002;52(478):373-80. 16 20. Orchard J, FreedmanFor SB, peerLowres N, et al. review iPhone ECG screening only by practice nurses and 17 receptionists for atrial fibrillation in general practice: the GP-SEARCH qualitative pilot study. 18 19 Aust Fam Physician 2014;43(5):315-9. [published Online First: 2014/05/06] 20 21. Orchard J, Lowres N, Ben Freedman S, et al. Screening for atrial fibrillation during influenza 21 vaccinations by primary care nurses using a smartphone electrocardiograph (iECG): A 22 feasibility study. European Journal of Preventive Cardiology 2016;23(Supp 2):13-20. doi: 23 10.1177/2047487316670255 24 22. Britt H, Miller GC, Henderson J, et al. General Practice Activity in Australia 2015-16: BEACH: 25 Bettering the Evaluation and Care of Health. Sydney: Sydney University Press 2016. 26 23. Media Release, Prime Minister and Minister for Health. Ground-breaking flu vaccines to protect 27 millions of Aussies 19 February 2018. 28 http://www.health.gov.au/internet/ministers/publishing.nsf/Content/health-mediarel- 29 yr2018-hunt020.htm, accessed 9 March 2018 30 24. NSW Government. Shingles Vaccination Program 2018. 31 http://wwwhealthnswgovau/immunisation/Pages/Shingles-programaspx, accessed 9 March http://bmjopen.bmj.com/ 32 2018 33 25. Australian Government Department of Health. Health assessment for people aged 75 years and 34 older (2014). 35 http://wwwhealthgovau/internet/main/publishingnsf/Content/mbsprimarycare_mbsitem_7 36 5andolder accessed 19 March 2018 37 26. Australian Government Department of Health. Chronic Disease Management - Provider 38 39 Information 2016. on September 28, 2021 by guest. Protected copyright. 40 http://wwwhealthgovau/internet/main/publishingnsf/content/mbsprimarycare-factsheet- 41 chronicdiseasehtm, accessed 9 March 2018 42 27. Gavino AI, McLachlan CS. Review of screening studies for atrial fibrillation in rural populations of 43 11 countries. Proc (Bayl Univ Med Cent) 2017;30(3):280-85. 44 28. Australian Institute of Health and Welfare (AIHW). Cardiovascular Medicines and Primary Health 45 Care: A Regional Analysis. Canberra, 2010. 46 29. Allenby A, Kinsman L, Tham R, et al. The quality of cardiovascular disease prevention in rural 47 primary care. Aust J Rural Health 2016;24(2):92-8. doi: 10.1111/ajr.12224 [published Online 48 First: 2015/08/11] 49 30. Orchard J, Lowres N, Freedman SB, et al. Screening for atrial fibrillation during influenza 50 vaccinations by primary care nurses using a smartphone electrocardiograph (iECG): A 51 feasibility study. Eur J Prev Cardiol 2016;23(2 suppl):13-20. doi: 10.1177/2047487316670255 52 [published Online First: 2016/11/29] 53 31. Peiris D, Usherwood T, Panaretto K, et al. The Treatment of cardiovascular Risk in Primary care 54 using Electronic Decision supOrt (TORPEDO) study-intervention development and protocol 55 for a cluster randomised, controlled trial of an electronic decision support and quality 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 13 of 24 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 Figure 1 - Kardia device and iECG reading 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 15 of 24 BMJ Open

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 http://bmjopen.bmj.com/ 32 33 34 35 36 37 38 39 on September 28, 2021 by guest. Protected copyright. 40 41 42 43 44 45 Figure 5 - Process flow of AF screening 46 47 46x68mm (600 x 600 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-023130 on 31 October 2018. Downloaded from Page 19 of 24 BMJ Open