The 15th Annual Meeting of the International Society for Medical Publication Professionals (ISMPP) was held in National Harbor on 15–17 April 2019. A summary of the first day of the meeting is provided below to benefit those who were unable to attend, and as a timely reminder of the key topics for those who did.
Day one of the meeting was opened by Chris Winchester (Chair of the ISMPP Board of Trustees and Oxford PharmaGenesis), before a keynote address considering the changing medical affairs landscape in the face of increasing patient engagement and a ‘data tsunami’. Later sessions looked into how blockchain technology could transform medical publishing, before some pertinent updates in ‘news you can use’. Finally, parallel sessions covered the impact of artificial intelligence (AI) on publications and innovations in extending the reach of posters.
Medical affairs 3.0 – back to the future?
The keynote address was given by Gary Lyons (Fishawack), considering how increasing patient knowledge and involvement is changing the face of healthcare and impacting the pharmaceutical industry. Technology and the resulting abundance of patient data is rapidly driving a 2.0 to 3.0 shift from the current patient-centric healthcare paradigm to a collaborative, wellness-focused approach. Our thirst for connected knowledge means that around 69% of adults in the US now track at least one health metric every day and are increasingly willing to share their data: a data and insight tsunami is heading our way. Opportunities for self-tracking mean that individuals can take a more proactive approach to managing their health, with activity trackers, smart watches, smart clothing and smart implants, patches and ingestibles representing key technology impacting healthcare in 2019. This abundance of data could be combined with AI and machine learning to create ‘digital twins’ in the future, helping to provide health guidance tailored to individuals.
Alongside the popularity of self-tracking, we are seeing an explosion in health tracking start-ups and investments. The race is on between pharmaceutical companies (the ‘incumbents’) and major tech companies (the ‘disrupters’) to embrace HealthTech. Lyons highlighted two recent cases of disruption:
- The Apple Heart Study: Apple and Stanford Medicine collaborated to monitor the heart rhythms of 419,297 individuals in the US using the Apple Watch and an iPhone app. Participants notified of irregular heart rhythms were prompted to contact a study doctor and could be sent an electrocardiogram (ECG) patch to confirm the atrial fibrillation findings. In data presented at ACC 2019, the positive predictive value for receiving a notification from the app was 84%.
- Tech firms are increasingly focusing on healthcare systems: in the last year, several partnerships have been struck to direct technology towards improving healthcare systems and patient experiences. Amazon, Berkshire Hathaway and JPMorgan Chase have established Haven, aiming to leverage data and technology to improve patient experiences and healthcare systems; Microsoft and Walgreens Boots Alliance have partnered to develop retail innovations and improved healthcare delivery models; and Apple have launched AC Wellness, aiming to integrate technology to deliver exceptional employee healthcare.
Healthcare professionals (HCPs), pharmaceutical companies and the medical communications field need to be mindful of the future impact of technology. In the UK, the Topol review was recently published, recommending how the National Health Service can take advantage of innovative technologies – and how these will affect current clinical staff. In pharmaceutical companies, the future will see medical affairs focusing more on bringing patients and HCPs together and developing technology collaborations. Excitingly, technology will also have role in addressing current issues in publications: AI may help to reduce bias and the human resource burden of peer review, improving the speed of publication, while blockchain could enhance peer review transparency and promote interaction between authors, reviewers and the wider scientific community.
Innovation insights: the transformative power of blockchain
The second session of the day, moderated by Gary Burd (Caudex), delved deeper into the potential of blockchain to solve current challenges in medical publishing.
Joris van Rossum (Digital Science) highlighted that how the key concepts of blockchain might relate to medical publishing:
- Blockchain brings us from an internet of information to an internet of value, establishing ownership of information and enabling cryptocurrencies – some companies are looking to develop a cryptocurrency for science, which could provide credit for peer review or data sharing, for example, and could then be spent elsewhere
- Blockchain creates immutable timestamps, establishing provenance and therefore intellectual property rights for novel content, ideas and data – this might encourage data sharing
- Blockchain can provide distributed, decentralised and transparent but pseudonymous data storage, meaning that blockchain can be used without the need for a third-party company or database – this could enable advanced research metrics and allow access for validation, transparency and reproducibility
Blockchain could also be used to create a neutral infrastructure supporting blind and promoting open peer review processes across the industry. This would facilitate finding suitable, available reviewers, enable reviewers to receive recognition, allow management of the openness of peer review, and improve the insight that funders and institutions have into review activities.
Sean Manion (Science Distributed and Blockchain in Healthcare Global IEEE/ISTO) then spoke about the implementation of blockchain in healthcare. Currently, 42% of blockchain investment is in the finance industry, compared to only 2% in healthcare. The US Department of Health and Human Services has just implemented a blockchain-based procurement system, with real time data sharing enabling faster strategic purchasing. However, alongside use in administrative or financial capacities within the healthcare industry, many potential applications of blockchain are being investigated: verifying provider identities, medical devices, the pharma supply chain, clinical trials, publishing, clinical support tools and electronic health records. However, many of these are pilot projects and might not be implemented for a decade. Blockchain in Healthcare Global IEEE/ISTO has identified 16 categories of barriers to adopting blockchain technology, including access, ethics, scalability, cybersecurity and education, which will require multi-stakeholder collaboration to resolve. Manion finished by directing us toward further information, including a library of resources created by the HIMSS Blockchain Task Force. He also highlighted that a special issue of Frontiers, Blockchain for Distributed Research, will be published in autumn 2019.
News you can use
Up next were three fast-paced presentations on noteworthy topics in medical publishing.
First off, Caroline Halford (Adis, Springer Healthcare and Springer Nature) spoke about ‘Can reporting of randomised controlled trials (RCTs) be improved?’. Caroline highlighted that researchers and journals are consistently under fire for inadequate reporting of RCTs and there is pressure to improve reporting standards. One analysis of RCTs published in 2016–17 found that the policy of some journals, requiring submission of a CONSORT checklist, may not have been very effective: inconsistencies between checklists and manuscripts are rife. Now, an RCT is underway, investigating whether providing feedback on how consistent CONSORT checklists and manuscripts are, alongside the standard peer review process, improves the overall completeness of RCT reporting.
Secondly, Amy Foreman-Wykert (Amgen) gave us a preview of what’s changed in the upcoming 11th edition of the AMA Manual of Style, due to be published in December 2019. Key changes include:
|Requirement to define some common abbreviations||Guidelines on citing social media, apps, podcasts and other digital media|
|Full stops following a digital object identifier (DOI) in journal references||Guidelines on terminology for socioeconomic status and addiction|
|Publication’s location from references||Guidelines for grammar in social media|
|Capitalisation of internet and hyphenation of email||Use of the singular “they”|
The final news snippet was delivered by Brian Falcone (Oxford PharmaGenesis), discussing the Pharmaceutical Information Exchange (PIE) Act. Historically, healthcare manufacturers have been reluctant to share clinical and economic information with payers prior to FDA approval, due to unclear regulatory guidance surrounding these communications. This delayed access to information in turn delays patient access to medications. As the US healthcare system becomes more value-focused, payers have shown increasing interest in receiving the information sooner, ideally 12–18 months prior to approval. The PIE Act, which would clarify the scope of permitted communication, was introduced as a bill in Congress in 2017, but has not yet been approved. Falcone emphasised that if it were, it could have a considerable impact on publication professionals: publication plans might be restructured to prioritise health economic publications, literature support for communication with payers might be requested, and outcomes-related measures might be introduced earlier in clinical research.
Attendees then broke off into parallel sessions, covering various topics, including current roles of AI in publications and extending the reach of posters with ‘publication extenders’.
How can AI be utilised in publications?
This session, moderated by Todd Parker (MedThink SciCom), examined how AI can be implemented in the pharmaceutical company context.
Ken Truman (MedThink SciCom) kicked off the session, looking at the current applications of AI. The availability of additional computational power and big data sets, and progress in machine learning, is rapidly enabling AI. AI encompasses machine learning, algorithms that learn from and make inferences based on data, and natural language processing (NLP; machine-based processing that allows for human-like interpretation of text). At the intersection of machine learning and NLP is natural language generation (NLG), generating natural language from data. A couple examples of AI use in healthcare were provided. Firstly, machine learning is being applied to detecting breast cancer metastases in tissue samples. Secondly, in publications, algorithms can simulate which aspects of a publication attract most reader attention and these predictive analytics can then be used to create effective poster designs.
Next, Dheepa Chari (Pfizer) discussed the operational applications for AI in the setting of a pharmaceutical company. Chari outlined a four-step operational approach for the selection of AI initiatives: ask a relevant question, acquire data needed to respond, analyse to find the solution, and act on key learnings. Opportunities for AI initiatives to solve business problems can fall under one of three categories:
- Insights: detecting key patterns and relationships from data in real time to derive deep, actionable insights, e.g. gap analyses and systematic literature reviews
- Engagement: using mass personalisation to improve customer, patient and employee engagement, e.g. digital assistants to answer scientific and clinical questions, bots to coordinate logistics of HCP interactions
- Automation: automating repetitive, knowledge- and natural language-rich human-intensive decision-making processes, e.g. detecting adverse event reporting in medical information requests, document review for medical/legal/regulatory purposes
Chari noted that in implementing AI, you should consider the availability of data and support for the project, identify the value that the initiative will bring, and ensure that expectations are aligned with the current possibilities of technology.
Jenny Ghith (Pfizer) rounded off the session by discussing the impact of AI in publications. Currently, an overload of information exceeds human capacity. AI might be able to help with time-consuming publication gap analyses, particularly in terms of identifying and selecting relevant articles. Extracting meaning from publications still needs human input, for the time being, due to the risk of false positives (AI including articles that are not relevant) and false negatives (AI removing articles that are relevant). However, the potential for AI in future time- and cost-savings in gap analyses and other resource-intensive, repetitive or mundane tasks, is huge. Existing AI technology is already being investigated in the industry and leveraging AI can significantly increase efficiency if utilised in current practices. However, we must evaluate AI through the lens of business needs and be mindful of its strengths and limitations.
At the end of the session, attendees were directed towards further resources including a Nature Toolbox article on future interactions between AI and scientific literature, and an online course from MIT: Artificial Intelligence: implications for business strategy. The session was summarised by emphasising the potential that AI has to rapidly provide answers to medical questions, ensuring that publication and strategic planning is more accurate. AI will require significant investment in software and will take time to get right, but when it does, the impact will be considerable.
Case study: extending the reach of your poster with digital publication extenders
In this session, Heather Abourjaily and Alia Bucciarelli (both from Biogen) discussed how moving from the traditional approach to posters to a more digital approach aligned with how HCPs want to receive information. According to a survey of neurologists, HCPs want to receive accurate, regular, prompt, up-to-date, non-promotional, unbiased information and materials. This is not new. However, how they want to receive information is changing. By 2020, two-thirds of HCPs will be digital natives (they qualified during or after the internet went mainstream, and have relied on digital interactions throughout their professional career). In a study by EPG Health, 81% of HCPs surveyed would prefer access to congress materials year-round, not just during the event, and want continued discussions beyond the congress setting. This shift requires a different, digital approach to communicating scientific information.
‘Publication extenders’ are helping the traditional poster to evolve to better share data. Biogen have developed a range of ways for HCPs to interact with posters: QR codes on posters link to a ‘microsite’ that contains the poster PDF, a mobile poster, poster slides and an author video. The microsite is optimised for viewing across mobile, tablet or desktop devices, and is available indefinitely. Author videos are short, 2–3 minute videos, which may be developed with the aid of free teleprompter apps.
Developing publication extenders presents challenges, including the short timeframe between finalising a poster and the congress, particularly considering that all materials require review and approval (including by all authors), meaning that perhaps efforts should be restricted to key posters. However, they can significantly increase the reach of a poster. For example, during Biogen’s last congress and the following week, users accessed high-priority microsites about 1200 times. This introduction of microsites and author videos increased the reach and lifespan of the data: for some posters, viewing time doubled following the congress. However, future opportunities include exploring Google analytics and introducing additional functionality to better meet HCP needs.