Research protocol: An investigation into the presence of financial conflicts of interest on support for screening for atrial fibrillation in the UK: A protocol for a descriptive study Introduction The UK National Screening Committee has recommended that population screening is not performed for atrial fibrillation (AF). This condition predisposes people to stroke, the risk of which can be reduced by anticoagulants. Atrial fibrillation (AF) is an irregular heartbeat. It can occur occasionally or persistently. It is more common with ageing. It may be symptomatic (causing a rapid heartbeat and e.g. palpitations, breathlessness, or lightheadedness) or asymptomatic (detected incidentally, or when it is specifically examined for, via taking the pulse, doing an ECG (electrocardiograph) or a cardiac monitoring over a longer period.) Some sports watches offer to detect atrial fibrillation. AF can cause complications, particularly strokes or transient ischemic attacks, sometimes called ‘mini-strokes’. There are two components of treatment. The first is to ‘rate-limit’, if necessary, slowing down a fast atrial heartbeat, which can reduce the symptoms of atrial fibrillation. The second is to treat with blood thinners (anti-coagulants). Over the last decade, several drugs have been developed (new oral anticoagulants, NOACs) which are more expensive than the older drug warfarin, but blood tests to check the dose are not are routinely monitored, unlike warfarin. This means that people do not require routine blood test monitoring when using NOAC drugs. Side effects include spontaneous bleeding, which is overall uncommon. There is evidence that when atrial fibrillation is treated with anticoagulants, the risk of stroke is reduced, however the risk of bleeding is increased (though to a lesser extent.) The trials showing benefit from preventing stroke from anticoagulants were in patients whose AF was picked up symptomatically, or during assessment for another condition. However AF may also be silent. Because it may be silent (asymptomatic), there has been the call to screen people for AF, i.e. to examine people to detect AF when they have no symptoms and are otherwise well. This has been proposed as a means to reduce stroke through more treatment with anticoagulant medication. Most recently people attending covid vaccination centres have been exposed to screening for AF, supported by the AF association (1) before being introduced into the NHS more broadly (2) Calls to screening assume that benefits will be the same for patients no matter how the AF is detected (symptomatic or asymptomatic). However this is a different population group from that in which benefits have been established in. The amount of AF in the wider asymptomatic population, and the risks of it, are not fully understood. It is currently unknown whether population screening for AF has a net benefit or harm. This is because the group of people in whom anticoagulants are found to be beneficial and who have been studied in research groups (diagnosed when they are symptomatic, or incidentally, for example, at at a follow up appointment for monitoring heart disease where blood pressure is checked) are different from asymptomatic people (who feel well, and are examined to detect atrial fibrillation). This means that the risk profile of this group is different, for experiencing both the benefits and the risks of medication. The UK National Screening Committee makes independent and evidence based judgements about the cost effectiveness of screening interventions in the UK (3). They concluded in 2019 that screening for atrial fibrillation was not recommended for these reasons, and also because it was not certain that earlier detection was beneficial for patients. The SAFER trial, a randomised controlled trial of screening for atrial fibrillation, started in the UK in 2021 and should answer many of these uncertainties. (4) However, the Getting It Right First Time organisation, a set of programmes within NHS England and NHS Improvement, has relied on recommendations from the European Society for Cardiology for screening rather than the UK National Screening Committee (3). Despite this, and the fact that screening for atrial fibrillation is not recommended by the UK National Screening Committee, the NHS has, at various points, proceeded to screen. The UK Government has, via the system of Associated Health Service Networks (AHSN), funded ‘innovations’ into supporting GPs and primary care more broadly into screening for atrial fibrillation. These structures primarily work through the Department for Trade and Industry rather than the NHS. This has been called ‘case finding’ by multiple organisations yet is, by any definition, screening, where asymptomatic people are investigated for a particular condition (5). These AHSN programmes have been sponsored by various technology and pharmaceutical companies. Heat maps have been produced claiming that atrial fibrillation is under diagnosed in some areas. GPs and pharmacists have been encouraged to screen in order to increase diagnosis rates. This has resulted in patients being given screening at covid vaccination appointments, when picking up prescriptions from a pharmacy, or in waiting rooms of GP surgeries using ‘instant’ , hand held, screening devices. The Apple watch is another example of AF screening which has been in the media with often uncritical comments about the function. There are at least two UK charities who are active in promoting screening. Previous concerns have been raised about pharmaceutical companies paying for pharmacists to make recommendations for patients with AF for anticoagulants, mainly in changing the older drug warfarin to NOACs (6). There has been recent publicity over screening for AF in Covid-19 vaccination centres in 2020/2021(7) , supported by charities who receive funding from industry. Associated Health Service Networks set standards and activities (2018-2020) for screening for atrial fibrillation together with ‘heat maps’ of areas with what they saw as underdiagnosis of AF and made suggestions, such as screening people attending pharmacies for AF (8). It is accepted that financial conflicts of interest are associated with bias towards the funder (9). A previous study of twitter activity and financial conflicts among haemato-oncologists has found a bias towards tweets regarding drugs for which the doctor had a financial conflict in and recommended better disclosure practices (10). At this time it is not shown that screening the adult population for AF is beneficial to patients, and may cause net harm through side effects of anticoagulant medication. It is important that people have high quality information, regarding benefits and harms of screening and the treatments this can lead to, in order to make informed choices to participate or to be treated. There has been widespread acceptance that screening, in inviting well people to be tested, requires to be performed in an ethical framework where shared decision making occurs, with accurate and meaningful information, and where patient autonomy is respected (11, 12). There is therefore a concern that the information being presented to the lay public may not reflect the underlying uncertainties of whether screening for AF is advantageous, and may be related to financial conflicts. By investigating and describing conflicts of interest in this area, we can assess whether uncritical promotion is more likely or not when financial conflicts of interest are present. We have noted that some stakeholders commentating on the desirability for screening have substantive personal financial conflicts of interest. These can be difficult to find and often require prolonged searching to identify. It is well recognised that financial conflicts can result in bias towards activities which benefit the funder. We seek to systematically examine the support for screening for atrial fibrillation in the UK and its relationship to financial conflicts of interest. The objective of this study is to describe the association between financial conflicts of interest and supportive messaging for screening for atrial fibrillation in the UK. This has not been investigated before, as far as we are aware, and we aim to delineate the link between financial conflicts and promotion of screening for atrial fibrillation. METHODS AND ANALYSIS Key question Is information on screening for atrial fibrillation, from individuals or health charities or pressure groups, in the public domain, subject to bias through financial conflicts of interest, and if so, to what extent? Overview We will conduct a descriptive cross-sectional study of UK mainstream and social media coverage of screening for AF from start January 2018- end July 2021. We will conduct a content analysis of the UK’s National Health Service (NHS) information about screening from AF between the same dates. We will trace the financial interests of individuals and organisations making positive (recommendations) negative (caution or discouragement) or neutral (fact based) commentary on screening for atrial fibrillation and draw conclusions on the likelihood of predominant and positive media commentary being associated with financial conflicts of interest. Search strategy Media coverage. We will identify relevant UK news coverage through searching the Proquest, the first twenty pages of Google News (pragmatic decision based on previous research using the same eg (13,14) and , PressReader electronic databases, with using explicit keywords,in the specified dates. The keywords will be, in quotes (for google search specificity) : ‘atrial fibrillation’, ‘afawareness’, ‘’globalafawareweek’ ‘knowyourpulse’ ‘af’ ‘irregular heartbeat” ‘afib’. A librarian/information specialist with expertise in systematic search design in the University of St Andrews will assist with the search strategy. We will expand the search as we discover links to individuals so that we have a comprehensive oversight of the most cited or active individuals in this area. Syndicated articles will be counted only once. Social media, Twitter, Facebook, Youtube and LinkedIn will be used with the same explicit keywords during the same time period. If we note other key terms with high search positivity we will consider adding these to the strategy. In order to minimise bias with Google News algorithms we will use an ‘incognito’ or private search tab. We will search initially by search words and test date ranges in pilots for the most relevant hits. NHS information. We will search for NHS press releases, Google News, and NHS branded websites (e.g. Associated Health Services Network, healthcheck.nhs, NHS RIghtcare), online UK news media and social media content only for information originating within the NHS. We will only include content that refers to screening for atrial fibrillation (e.g. using tests in asymptomatic individuals). We will exclude news and social media content that only focus on tests for symptomatic people, media stories about patent approval or business issues only. We will include syndicated news stories only once. We will do this by searching for : All previous search terms + NHS For NHS branded websites: NHS Digital, NHS Scotland, Wales, healthcheck.nhs, NHS Rightcare, NHS Accelerator, NHS Longterm Plan, NHS Screening and Associated Health Service Networks. Google: Search term = Site:WEBSITE NAME “SEARCH TERM 1” OR “SEARCH TERM 2” Patient information websites. We will search for the top 10 most popular google hits, including adverts/promoted sites for patient facing information websites regarding atrial fibrillation which include screening as a route to diagnosis e.g. “You should see a doctor for diagnosis but you may also detect it by feeling your pulse at your wrist. A normal heart rate is steady and usually between 60 and 100 beats a minute when you're resting.” We will do this by searching for: ‘’Atrial fibrillation’ and ‘AF’ Google: Search term = Site:WEBSITE NAME “SEARCH TERM 1” OR “SEARCH TERM Screening process Sets of two independent investigators will be involved in performing the screening of stories, tweets, and facebook posts. We will exclude exact duplicates (same title, same outlet and same date) before starting the screening and will keep track of the number of duplicates. Investigators will independently assess the eligibility of news and social media content for potential inclusion according to the predefined selection criteria. Any disagreements in judgement will be resolved by discussion to reach consensus or by consultation with a third reviewer. Get list of articles Export results from a given database e.g. google news Turn off adblock Use verbatim search/depersonalised search Exclude quickly based on title, subtitle, or skimming full text if necessary For websites where Researcher 1+2 disagree, discuss and agree if to include or exclude. Researcher 3 or 4 to act as final say if can’t agree. Data extraction and coding We will use a structured template to extract and code the relevant data. An iterative design process will be used to refine the tool for the purpose of our study. Sets of two independent investigators will extract data and code the media content. Any disagreements in extraction or coding will be resolved by discussion to reach consensus or by consultation with a third reviewer. The percentage of disagreements on each coding variable requiring resolution through use of a third reviewer will be recorded. Before formal data extraction and coding, the sets of independent investigators will apply the data extraction tool to code 30 statements about screening for AF. Disagreements in data extraction and coding will be resolved by discussion and this will act as our pilot. We will take each example and record: professionals who have been quoted or who make statements; and charities or organisations who make statements or are quoted regarding screening for atrial fibrillation. We will record their statement. We will record whether these are human interest stories, research findings based stories, or ‘awareness’ weeks or events, and what type of media they have appeared in: eg social media, local newspapers, national newspapers, or magazines. We will take screenshots of each. When stories have been reported identically or near-identically in multiple venues via a single press release, we will analyse only one example. We will record all declarations of interest made in the text, and their attribution (source/journalist/other). We will pilot this with 30 news stories initially and then review our processes to ensure they are fit for purpose. We will report and explain deviations from protocol. Analysis We will develop and agree a method for scoring quotations from individuals or representatives of organisations as positive (in broad support of AF screening) and negative (‘there may be many false positives’) or neutral (comment contains both potential benefits and harms). These will use key words and phrases (for example: ‘everyone should know their pulse’, or ‘screening for atrial fibrillation will prevent stroke’ or ‘found 20 people with atrial fibrillation….will save lives’ as positive examples, and ‘we need more research on whether this will help’ or ‘we don’t recommend screening yet’ as neutral examples.) We will test these with the first 30 or so mentions of AF screening by each of us independently marking them and then reviewing the results side by side. In the case of multiple views being expressed (eg by author and commenters) we will extract data for each representative/commentator. We will examine the financial conflicts of interest of those whose views or recommendations are represented in each story/tweet/comment. We define a conflict of interest as a financial tie to a device company actively marketing screening technology for AF or a pharmaceutical industry or medical device company actively marketing medicines or devices for the treatment or prevention of AF. This will be determined through searches of industry/company websites and relevant product information material by a single reviewer. Each episode will be ‘marked’ by two individuals for positivity, neutrality and negativity. If ratings cannot be agreed with discussion, they will be arbitrated by the another researcher. We will include all financial ties that have been recognised by the International Committee of Medical Journal Editors (ICMJE): grants (funding for research study), personal fees (consulting, advisory, speakers, honoraria, travel), patents/copyrights/royalties and miscellaneous. We will focus on the most active organisations or individuals, aiming to be as comprehensive as possible and working within the parameters of the resources we have. We aim to analyse 30 of the most cited organisations/individuals. We will complete a search for each person/organisation. For academics, we will conduct a search of academics’ publications in the Pubmed database using their names and affiliations over the four years before the publication of media story/tweet/comment. WHO guidance suggests a period of four years prior to publication is relevant when disclosing financial ties. Full texts will be read. Full text reading will be stopped once one potentially relevant financial tie was found within the four year search period. The URL and full text of the declaration will be extracted. For health professionals, Disclosure UK will be searched using health professionals names and location; we will also a search on youtube for presentations/lectures, examine conflicts of interest registers if they exist for their workplace, and for publications in journal articles for declarations of interest together with FOI requests if needed, Youtube for conference presentations and declarations, Companies House, and google searches, which may reveal educational sponsored material. For medical societies or patient/consumer organisations, we will visit their websites and look for relevant payments that would be defined as a potential financial tie in their annual financial report on the website, the presence of medical device or pharmaceutical companies’ (that actively market medicines or devices for the treatment or prevention of AF) logos on their homepage; the presence of relevant pharmaceutical or medical device industry sponsorship in the programme of the medical society's last annual conference; the presence of relevant industry sponsorship of satellite symposia during the last annual conference/any patient/consumer event, disclosures in annual reports or at Companies House, or any references to relevant medical device or pharmaceutical industry sponsorship or payments on the website. We may have to write to them or sponsors. We will stop once a conflict has been found and use time-limited searches focussed on the last 10 years. We will record the financial conflict, the date, and the nature of the financial transaction if recorded (e.g consultancy to company making drugs for anticoagulation). We will upload screenshots of all of these to Figshare, excepting FOI responses if we do not have permission to publish. Once obtained, we will analyse the financial information for connections with interventions for the diagnosis, treatment and screening for AF. If there is no obvious financial bias towards the earlier diagnosis of AF, we will complete the detailed search in order to minimise non-detection of related interests. We will then have a set of the most frequent commentators /organisations regarding screening for AF in the UK, and whether this was likely to be positive, negative, or neutral, mapped to their financial interests. This will allow us to draw conclusions about the strength of positive or negative associations of financial interests to positivity, neutrality or negativity regarding support for AF screening. Investigation of internal NHS activity Of the NHS organisations promoting AF screening, we will request information via FOI for industry funding, including donations/cover of staff costs, lecturing, technology, educational events, data collection, and publicity involving industry in relation to it. We will locate professional and patient facing resources produced by these organisations and search for notification of sponsorship on them. We will search for up to 5 of each. We will record each. This will allow us to establish the type of industry funding for AF screening within the NHS. We will also request details of sponsored posts, or industry support, which cover cardiology or atrial fibrillation management. Data analysis Descriptive statistics (means, SD, counts and percentages) will be used to summarise the extracted data (eg, number of stories, number of tweets, number reporting benefits/positive comments and harms/negative comments, etc.). Categorical data analysis will be used to investigate potential associations between overall impression of the media content (positive/beneficial, negative/harmful, mixed/neutral) and presence of financial conflicts of interest among the commentators. We will examine how many of our top cited individuals/organisations commenting in a negative, neutral, or positive way regarding screening for AF had financial conflicts with industry. Ethical considerations We discussed this project with the convener of the University of St Andrews Teaching and Research Committee. Because all the data we were searching for was in the public domain, it was agreed that there was no need for formal ethical review. We took the decision not to publish names of any individual healthcare professionals in the published article, and names would be removed from appendices. Patient or public involvement Discussions took place with members of a patient panel who had already been convened to assist with another research project examining conflicts of interest in medicine. Members supplied information and citizen views were sought in writing the protocol. Dissemination The results of this study will be published in a peer reviewed journal and presented at relevant medical conferences. Results We will have created a framework and recording of positivity/negativity of statements regarding screening for atrial fibrillation, and the association between these and financial conflicts of interest. We will assess their relationship. We will describe the current activity in the NHS in relation to atrial fibrillation screening, given that it is not recommended by the UK NSC at this time. We will describe the financial conflicts of interest present in these organisations. We will publish these results in a peer reviewed journal. Appendix Draft Coding Story Origin Type of story URL Date Quote Positive Neutral Negative Owner of quote/status Status: Academic/Charity/NHS organiation COIs Can a £245 heart-check patch diagnose atrial fibrillation and save thousands of people a year suffering a stroke? Daily Mail News https://www.dailymail.co.uk/health/article-9231461/The-high-tech-patch-save-stroke.html 6/2/21 “With the Covid pandemic, using wearable technology like this to track patients’ health remotely is a no-brainer.” P Professor Martin Cowie Prof Cardiology Imperial College London A Research grants administered by Imperial College London from Bayer, Boston Scientific, Abbott, Medtronic, and ResMed • Consultancy and speaker fees from ResMed, Servier, Novartis, Pfizer, Bayer, Medtronic, Boston Scientific, St Jude Medical, Alere, Daiichi-Sankyo, Bristol Myers Squibb, Roche, Amgen, MSD, Respicardia, Sorin, Torrent Pharmaceuticalshttps://www.rcplondon.ac.uk/file/13336/download Joynson 23/3/21 AF Association. Oxford AHSN Project review https://www.oxfordahsn.org/wp-content/uploads/2021/05/Project-review-West-Suffolk.pdf Targets AF Detection in COVID-19 vaccination centres. GRIFT/ Oxford ASHN https://www.oxfordahsn.org/wp-content/uploads/2021/05/COVID-vax-clinic-guidelines-FINAL.pdf UK National Screening Committee. Atrial Fibrillation https://legacyscreening.phe.org.uk/atrialfibrillation#:~:text=The%20UK%20NSC%20does%20not,in%20people%20found%20through%20screening Safer Study. https://www.safer.phpc.cam.ac.uk/ McCartney M, Fell G, Finnikin S, Hunt H, McHugh M, Gray M. Why ‘case finding’ is bad science. Journal of the Royal Society of Medicine. 2020;113(2):54-58. doi:10.1177/0141076819891422 McCartney M. Partnerships: pharma is closer than you think BMJ 2015; 351 :h3688 doi:10.1136/bmj.h3688 Resources for detection of AF at COVID-19 vaccination centres https://www.heartrhythmalliance.org/aa/uk/detection-of-af-at-vaccination-centres ASHN Atrial Fibrillation. National Programme April 2018-March 2020 https://www.ahsnnetwork.com/about-academic-health-science-networks/national-programmes-priorities/atrial-fibrillation Nejstgaard C H, Bero L, Hróbjartsson A, Jørgensen A W, Jørgensen K J, Le M et al. Association between conflicts of interest and favourable recommendations in clinical guidelines, advisory committee reports, opinion pieces, and narrative reviews: systematic review BMJ 2020; 371 :m4234 doi:10.1136/bmj.m4234 Kaestner V et al. Conflicts of interest in Twitter. The Lancet Haematology 4/9/17 https://broomedocs.com/wp-content/uploads/2019/08/Twitter-COI.pdf Forbes et al. Offering informed choice about breast screening. J Med Screen 2014 Vol 21(4) 194-200 Joynson C. 23/3/21 Embedding ethics at the UK National Screening Committee. https://phescreening.blog.gov.uk/2021/03/23/embedding-ethics-at-the-uk-national-screening-committee/ Steve Woloshin, Lisa M. Schwartz, Sara Dejene, Paula Rausch, Gerald J. Dal Pan, Esther H. Zhou & Aaron S. Kesselheim (2017) Media Coverage of FDA Drug Safety Communications about Zolpidem: A Quantitative and Qualitative Analysis, Journal of Health Communication, 22:5, 365-372, DOI: 10.1080/10810730.2016.1266717 O'Keeffe M, Barratt A, Maher C, et alMedia Coverage of the Benefits and Harms of Testing the Healthy: a protocol for a descriptive studyBMJ Open 2019;9:e029532. doi: 10.1136/bmjopen-2019-029532