Digital Health Section Podcast- Royal Society of Medicine (RSM Digital Health Council)

Max Jaderberg, Director of Machine learning at Isomorphic labs, shares the Isomorphic approach to drug discovery in one of their first interviews as they emerge from shadow mode.

In 2020 Google DeepMind made history when they took on the grand challenge of biology- the protein folding problem- with their algorithm AlphaFold2.  Their success prompted the creation of a sister company, Isomorphic Labs. Our conversation topics include:   - The concept of “Digital Biology”- How AI can be used model the biological world 

- Bottlenecks in the drug discovery process and how AI could circumvent them

- How AI may unlock 'undraggable targets' to find treatments for incurable diseases  Links:

Isomorphic labs website: https://www.isomorphiclabs.com/

AlphaMissence: https://www.deepmind.com/blog/alphamissense-catalogue-of-genetic-mutations-to-help-pinpoint-the-cause-of-diseases

This episode features a conversation with Dr Saddif Ahmed- Director of Clinical Safety and Accuracy at Flo Health, uncovering how femtech products are providing novel insights into women's health. Discussion topics include: - How period tracking apps are leading to earlier diagnosis of women's health disorders - How femtech data is challenging conventional medical wisdom and offering novel insights into how sleep, mood and exercise vary of their menstrual cycle - The potential for future research in the field using data collected at scale from apps and wearables

This episode features a conversation about bias in healthcare AI models with Dr Xiao Liu- Senior Clinician Scientist in AI & Digital Health, University Hospitals Birmingham NHS Foundation Trust.

Discussion topics include:

• How does AI bias arise? 
• How can we tell whether AI algorithms are biased?
• Who should be held accountable for bias in AI systems?
• Is AI the solution for AI bias?

Papers referenced during the discussion:

Ethical Machine Learning in Healthcare: https://pubmed.ncbi.nlm.nih.gov/34396058/

Standing together recommendations: https://www.datadiversity.org/about

This episode features a conversation with Sara Nelson- Programme Director at DigitalHealth.London focussing on how to pitch healthtech to the NHS.

Sara shares pitching tips based on her experience working within a NHS healthtech accelerator and as deputy CNIO at NHSX

Topics include:

- Navigating the NHS procurement landscape

- Identifying budget holders- why the CEO may not be the best person to pitch to

- Why being too flexible can put buyers off

- The power of effective follow up

- The risk of offering a free pilot

This episode focusses on AI & the Limits of Clinician Vigilance and features of conversation with Prof Julia Adler-Mistein, Professor at the University of California.

Conversation topics include 

• Lessons form the aviation and automobile industry
• Should AI be checking clinicians work rather than the other way around?
• Potential solutions to improve clinicial vigilance ability  
• The importance of considering behaviour science and systems design thinking
• Is there an ethical risk of demanding superhuman performance from AI tools?

Paper reference: https://jamanetwork.com/journals/jama/article-abstract/2816582

This episode focuses on the topic of AI education and workforce transformation with Dr Hatim Abdulhussein, National Clinical Lead for AI and Digital Workforce at NHS England

Key discussion topics include:

- The NHS England strategy for AI education for the healthcare workforce

- 5 healthcare workforce AI archetypes: Shapers, Drivers, Creators, Embedders and Users

- The role of workforce transformation from AI MDT teams to the professionalisation of the digital and data professionals and the creation of new leadership roles in technology and AI. 

Developing healthcare workers' confidence in AI: https://digital-transformation.hee.nhs.uk/building-a-digital-workforce/dart-ed/horizon-scanning/developing-healthcare-workers-confidence-in-ai

Artificial Intelligence (AI) and Digital Healthcare Technologies Capability framework: https://digital-transformation.hee.nhs.uk/building-a-digital-workforce/dart-ed/horizon-scanning/ai-and-digital-healthcare-technologies?

This episode features a conversation with Eric Topol, a renowned cardiologist scientist, and author. He is the founder and director of the Scripps Research Translational Institute where he is a professor of Molecular Medicine . He is editor-in-chief of Medscape and He has published three bestselling books including the highly influential  Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again . 

In 2019 he led the Topol review a report focussed on planning for the NHS future workforce, integrating genomics, digital medicine, and artificial intelligence.

Our discussion topics include: 

• How LLMs could make doctors more empathetic
• The potential for ‘keyboard liberation’ using LLMs to repair the damage that EHRs have inflicted on the doctor-patient relationship
• The ‘silicon-vallydation’ issue and the need for high quality evidence generation for AI
• Hallucinations and the ‘frequently wrong never in doubt’ issue

In this episode Stephen Carter delivers a masterclass in all things IP for Healthtech and AI

This includes 

• How to decide whether to apply for a patent 
• Why you need to be wary of copyright when using open-source code
• The myth that you cannot patent AI products. 
• Key pitfalls that companies and organisations of all sizes fall into when thinking about IP

This episode features a conversation with Dr Murray Ellender- CEO of eConsult. It explores how innovative digital triage tools are catalysing new ways of working and improving system efficiency within the NHS.

Discussion points include:

- Digital triage enabled primary & secondary care

- The concept of one unified 'digital door' across NHS services

- The future of AI-enabled triage

This episode features a conversation about Investment in Healthcare with Dr John Lee Allen who is Managing Partner at RYSE Asset Management.

 

John has a background as a physician scientist, entrepreneur, and investor and his ability to bridge the gap between academia and industry has made him an influential figure in the fields of science, technology, and investment.

 

John sits on company boards, is an NHS Board Governor, Grant Assessor at Innovate UK, Expert-In-Residence at Imperial College London, NHS Clinical Entrepreneur Fellow & Mentor, RSM Digital Health Council member and a Fellow at the University of Cambridge.

Conversation topics include:

• Different types of healthtech funders
• How to find your ‘investor tribe’
• Building relationships with the 3 'hums'- humanity, humility and humour
• The AIDA framework for building an effective investor pitch  
• Effective fund raising strategies 

This episode features a conversation with Professor David Lowe- Clinical Director for Innovation at the Digital Health Validation Lab, University of Glasgow and A&E consultant.

Key conversation topics:

- The story of the Digital Health Validation lab

- What validation of Digital Health Technologies means in practice and how it varies depending on the maturity of the technology

- Evaluation challenges and what could be done to overcome them from a policy and governance perspective

- Hints and tips to early stage companies when it comes to evaluation

Digital Health Validation Lab website

In this episode, we delve into the realm of AI regulation and its impact on the future of medical AI innovation in UK. Joining me is the esteemed Professor Stephen Gilbert, a medical device regulatory science academic.

During our discussion, we explore:

• The necessity for innovative regulatory approaches in the AI domain
• The diverging paths of EU and US regulatory strategies
• Challenges posed by the EU MDR approach and its horizontal regulatory system
• Pioneering approaches to regulation, emphasising system rather than product-level oversight

Relevant publicatons:

https://www.jmir.org/2023/1/e43682

https://www.nature.com/articles/s41591-023-02412-6
https://www.sciencedirect.com/science/article/pii/S2949761223000421

This episode features a conversation with Dr Dillon Obika- Senior Clinical Specialist at Google about MedLM: Google's family of medical foundation models.

Conversation topics include:

- Why and how Google have been developing medical foundation models

- How MedLM is currently being used in clinical practice

- The challenge of evaluating model performance for clinical tasks

- How Google are tackling clinical safety risks arising from hallucinations, omissions and bias

This episode features a conversation with Dr Molly Gilmartin- Health Tech investor at AlbionVC about healthtech market trends and trajectories. 

Discussion topics include:

- Recent healthtech market investment trends

- Predicted trends in the next few years

- How to approach market entry

- Breaking down key buzzwords and phrases from product market fit to TAM, SAM, SOM, CAC and CTLV

- Challenges associated with business cases in healthcare and how these vary across regions

This episodes discusses the pragmatic realities of conducting post-market surveillance for AI medical devices. It features a conversation with Dr Dan Mullarkey, Medical Director at Skin Analytics, a company specialising in AI for skin cancer detection.

In our conversation you’ll hear Dan explain how Skin Analytics have tackled some difficult questions such as:

How do you decide the appropriate sensitivity levels of an AI device?

How you can ensure you’re given access to the ongoing outcome data you need from NHS organisations to conduct appropriate PMS 

What are the challenges of dealing with real world data analysis compared to traditional research approaches?

This episode features a conversation with Professor Colin Espie Professor of Sleep Medicine at the University of Oxford and co-founder and Chief Scientist at Big Health. It focusses on the topic evidence generation for digital therapeutics

Discussion topics include:

- Why evidence generation is so important for digital therapeutics

- The distinction between evidence-informed and evidence-based therapies

- The challenge of designing placebo controlled trials for digital therapies

Links to papers:

https://www.healthaffairs.org/do/10.1377/forefront.20220223.739329/

https://www.jmir.org/2023/1/e43682

In this episode Giuseppe Sollazzo shares his insights from the NHS AI lab skunkworks programme - an initiative to find new ways to use AI for driving forward the early adoption of technology in the NHS.

Discussion topics include:

- The value of an open-source approach to AI development in the public sector

- The importance of being transparent about both successes and failures

- The role of MDT AI teams in the NHS and upskilling the workforce in relation to AI

More information about the skunkworks projects here: https://transform.england.nhs.uk/ai-lab/ai-lab-programmes/skunkworks/

This episode explores software as a medical device regulation on both sides of the Atlantic.  It features two incredible guests, Johan (ex-MHRA) and Brendan (ex-FDA) who share their unique perspectives from their respective experience at regulatory agencies. Discussion points include: The key differences between the EU, UK and US approaches to regulation - How these differences impact market entry strategy - Approaches to regulating AI including the EU AI Act and what it means for AI medical device manufacturers - International regulatory harmonization efforts - Key Dos and Don’ts when interacting with regulators

Who will be held to account legally for harm caused by AI use in Healthcare? This episode delves into this knotty question with Majid Hassan- a solicitor and parter at Capsticks LLP specialising in Healthcare litigation.

The conversation covers:

- Key legal concepts such as negligence and and product liability

- How the litigation process would take place practically

- Why you could be more at risk from NOT using AI than using it

This episode features a conversation with Dr Hugh Harvey- Managing Director at Hardian Health exploring the regulatory, ethical and logistical challenges involved in developing Large Language Models (LLMs) as medical devices.

Discussion topics include:

- A step-by-step guide to the regulatory approval process for LLM medical devices

- Should users be allowed to ask general purpose LLMs (like chatGPT) medical questions?

- What needs to change in current regulation in order for it to keep pace with recent AI developments

Blog posts mentioned in podcast:

Hardian Health blog by Hugh Harvey: https://www.hardianhealth.com/blog/how-to-get-regulatory-approval-for-medical-large-language-models

MHRA LLM blog post: https://medregs.blog.gov.uk/2023/03/03/large-language-models-and-software-as-a-medical-device/

RSM Ethics and Regulation for AI in health event: https://www.rsm.ac.uk/events/digital-health/2022-23/ter04/

This episode features a conversation with Dr. Bob Wachter, a prominent physician and academic known for his significant contributions to the field of healthcare and technology.

Dr Wachter has authored several influential books including the brilliant "The Digital Doctor: Hope, Hype, and Harm at the Dawn of Medicine's Computer Age" (2015)

In 2016, Dr. Robert Wachter was commissioned by the UK government to deliver Wachter Review aimed to assess the state of health information technology (IT) in the National Health Service (NHS) and provide recommendations for its improvement. 

As we mark seven years since the Wachter Review, Dr Wachter joins me to reflect on the progress made, future potential, and address the persistent challenges facing the NHS.

The conversation explores the journey that the NHS has come on since 2016 including the widespread introduction of electronic health records, patient health record access and development of NHS CCIOs. It also explores the future of AI enabled healthcare, the challenge of the so-called productivity paradox and whether AI is likely to result tin evolution or revolution of care delivery.

This episode features a conversation with Malte Gerhold- Director of Innovation and Improvement at the Health foundation.

It discusses Implementation Insights from the research and policy undertaken by the Health Foundation on Health technology and AI.

Discussion areas include:

- Why we need to focus on funding change rather than the technology

- The public discussions that are needed about trade-offs patients are willing to make when it comes to AI 

- Why we need a problem driven approach healthtech adoption 

- The importance of avoiding an AI aristocracy in the NHS so that everyone can benefit 

Papers discussed:

Priorities for an AI in health care strategy

How would clinicians use time freed up by technology?

This episode features a discussion with Mohammad AI-Ubaydli- CEO and Founder of Patients Know Best on personal health records.

Key discussion points include

- Do patients really know best?

- How personal health records are driving clinicians to rethink the way they write their clinical notes

- Changes that are needed within clinical training to better prepare the workforce for treating patients using personal health records

In this episode Prof Pearse Keane- Consultant Ophthalmologist at Moorfields and Professor of Artificial Medical Intelligence at University College London discusses AI in Ophthalmology.

Conversation topics include:

• The story of the Moorfields- Google Deep Mind collaboration and the use of AI to detect retinal diseases
• The birth of 'oculomics' as a field: Using the eye as a window to the rest of the body
• The practical challenges of moving from code to clinic when it comes to Healthcare AI
• The potential for the UK to be a world leader in healthcare AI by linking NHS data with top universities
• Medical foundation models and their role in the future of healthcare

In this episode focusses on the future of regulatory approvals for AI software medical devices in the UK (and EU).

Featuring guest James Dewar- Co-founder of Scarlet a EU Notified body and UK approved body that specialises in certifying software medical devices. 

Key discussion topics:

• The current regulatory position of the UK post-Brexit and the opportunities that this could present
• The practical impact of the EU AI act for medical device manufacturers within the EU. 
• What makes regulatory submissions uniquely challenging for AI devices
• Getting novel technologies such as Large Language model and other foundation models regulatory approval

In this episode Dr Kath Halliday- President of the Royal College of Radiologists discusses how AI is impacting radiology.

Discussion points include:

- The AI technologies that are being used in NHS radiology derpatments

- Challenges faced with deploying AI in the NHS

- How AI is likely to change the nature of radiologists' work

- The use of autonomous AI in radiology

Links from the Royal College of Radiologists:

- AI Registry

- AI education course

- AI Code of Conduct

In this episode Haris Shuaib- Head of Clinical Scientific Computing at Guy's and St Thomas'​ NHS Foundation Trust shares his vision of what is needed to build an AI-ready NHS.

Haris shares the story of how he created the first-ever NHS specialist medical AI team and discusses the unique value that individuals with a dual skillset of clinical and technical expertise bring to the NHS, allowing it to become:

- an intelligent AI customer

- a learning healthcare system

- a developer of in-house algorithms to solve problems that wouldn't be tackled by commercial AI companies.

Haris also shares the importance of 3 key factors in the success of AI at scale: People, Policy and Platforms

This episode features a conversation with Dr Minal Bakhai - National Director for Primary Care and Community Transformation and Improvement at NHS England. She discusses how technology is enabling system change in General Practice through the Modern General Practice model.

Drawing on her experience of product development, user research and design from working with the digital health industry she shares how she is using an evidence based test-and-learn approach to developing new ways of working in primary care and national standards to ensure that solutions are as useful and accessible as possible.

Minal shares the most important enablers to shifting behaviour across General Practice, the potential of novel technologies like ambient voice transcription and what you need to consider if you want to introduce a new technology into your practice.

In this episode Dr Chris Kelly- Clinical Research Scientist at Google discusses key barriers to the widespread adoption of AI in healthcare.

In 2019 Chris wrote an article titled ‘Key challenges for delivering clinical impact with artificial intelligence’ and the episode covers how Chris might approach this article if he was writing it again 5 years on.

Chris breaks down the barriers into 6 buckets:

- People factors & change management

- Machine learning and technical factors

- Safety and monitoring of AI once deployed

- Governance

- Regulation

- Business challenges

Mammography AI paper discussed during the episode

This episode centres on the evaluation of digital health solutions. It features a conversation with Dr Saira Ghafur, Consultant in Respiratory medicine and co-founder of Prova Health.

Discussion topics include

- The different types of evidence required at different stages of product development

- The use of innovative approaches such as simulation to produce evidence for digital health technologies

- The importance of linking evidence and reimbursement and different international approaches that are being implemented to tackle this

This episode features a conversation with Jeanette Kusel- Director for Scientific advice at NICE the national institute of health and care excellence, and addresses the role that NICE  has to play with regards to digital health technologies

Discussion topics include:

• The difference between the evaluation required for regulatory approval (CE/UKCA marking) and NICE approval
• Innovative new approaches to assessing healthtech innovations including:
• The NICE META tool
• Early value assessments
• The AI and digital regulations service
• NICE Evidence Standards Framework
• What is involved in a NICE approval process for digital health technologies 
• Common mistakes that companies make when evaluating their products and how these can be avoided

Links:

META tool: https://meta.nice.org.uk/

Early value assessments: https://www.nice.org.uk/about/what-we-do/eva-for-medtech

AI and digital regulations service: https://transform.england.nhs.uk/ai-lab/ai-lab-programmes/regulating-the-ai-ecosystem/the-ai-and-digital-regulations-service/

NICE evidence standards framework: https://www.nice.org.uk/about/what-we-do/our-programmes/evidence-standards-framework-for-digital-health-technologies

NICE Real World Evidence framework: https://www.nice.org.uk/corporate/ecd9/chapter/overview

In this LIVE recording of the RSM Digital health podcast from the HETT conference guests Darrien Bold and Dr George Harston discuss the key enablers that supported AI for stroke to be deployed in 100% of NHS stroke units in England

Discussion topics include: 

• The power of top down focus on specific areas of unmet need
• Efficiencies gained from centralising due diligence and governance processes
• Resourcing focussed on dedicated central teams, financing local procurement and enabling evaluation 
• Peer to peer learning and the need for shared learning networks 
• Guidelines and the importance of demonstrating evidence of value at a system level

Exploring Google's consumer mental health efforts with Megan Jones Bell, Clinical Director- Consumer and Mental Health at Google & ex-Chief Strategy and Science Officer at Headspace.

This episode is themed on mental health to accompany the upcoming Digital Health section event 'Digital mental health in adults and older people' https://www.rsm.ac.uk/events/digital-health/2022-23/ter52/

Find out more about Google's consumer mental health efforts here: https://blog.google/technology/health/mental-health-month-2022/

This episode features a conversation with Dr Satya Raghuvanshi- VP at Accurx exploring the development of healthtech communication tools in the NHS. Discussion points include - How Accurx is solving entrenched communication problems in primary & secondary care - Accurx’s approach to product development and market entry that have contributed to its success - How Accurx is now tackling elective recovery

This special edition of the RSM digital health podcast is a joint episode with the hosts of the Dev & Doc podcast focussed on Explainable AI in healthcare.

Key discussion topics:

- The confusion around what explainable AI is

- The question of whether it is really necessary in healthcare

- Exploring different explainability techniques

- The future of explainability and what’s needed to get there

If you like this episode then I'd love to hear your comments and and do check out the Dev & Doc podcast where doctors and engineers unpick healthcare AI related topics

On this episode Dr James Ray National clinical advisor for urgent and emergency care at NHS England shares how digital technologies are being introduced to improve care delivery in 111 and the challenges faced and lessons learned along the way He also shared his vision for how a single digital front door could help to simplify and standardise triage and provide a potential solution to the NHS navigation problem.

This episode focusses on PIFU (patient initiated follow up) and how technology is helping to reimagine outpatient care pathways. It features a conversation with Tom Whicher, CEO and co-founder of DrDoctor Discussion topics include: - How technology enabled patient initiated follow up pathways are helping patients, clinicians and the healthcare system  - How data from appointment bookings is enabling a personalised approach to patient appointment booking and outpatient care  - How PIFU technology can help to even out the variability in care experienced in the NHS

This episode features a conversation focussed on copilots to assist clinicians with Dr Dom Pimenta who is a cardiologist and CEO of Tortus the creators of OSLER, an AI agent for clinicians 

Conversation topics include:

• Clinical liability for AI-augmented decision making
• How can we protect clinicians from the impact of AI misdirection and automation bias
• The practical realities of clinically evaluating LLM tools
• The surprising use of clinical communication skills in LLM prompt engineering

This episode features a conversation with Jess Morley- Health Data Academic at Yale Digital Ethics Centre, discussing ethical implications of digital health transformation.

Jess shares a framework for thinking about some key areas where technology is likely to impact on healthcare and some difficult questions we are going to need to face as a healthcare system. 

Conversation topics include questions about trust, accountability, equity and the changing doctor patient relationship.

This episode focusses on the topic of confidence in healthcare AI technologies. My guest is Dr Mike Nix, Lead AI Clinical Scientist at Leeds NHS trust.

Discussion topics include:

• How AI errors and human errors are fundamentally different, and why that’s so important in the context of healthcare AI 
• The concept of explainable AI and its potential limitations in the context of human cognitive biases and decision making 
• The role of regulation in supporting AI confidence 

Link to reports on Healthcare workers' confidence in AI:

https://digital-transformation.hee.nhs.uk/building-a-digital-workforce/dart-ed/horizon-scanning/understanding-healthcare-workers-confidence-in-ai

https://digital-transformation.hee.nhs.uk/building-a-digital-workforce/dart-ed/horizon-scanning/developing-healthcare-workers-confidence-in-ai

This episode features a conversation with Dr Azeem Alam, member of the RSM Digital Health Council and co-founder of Bitemedicine, the creators of Bitelabs, a healthtech and innovation fellowship for clinicians  Discussion points include: - The bitelabs fellowship - Ways clinicians can gain experience in digital health - How you can combine tech experience with clinical training  - Roles for clinicians within healthtech companies Resources/opportunities discussed during the podcast: Bitelabs: @BiteLabsHealth https://bitelabs.io Fellowships/programmes: - Topol Digital fellowships:⁠ ⁠⁠Topol Digital FellowshipsHealth Education Englandhttps://topol.hee.nhs.uk › digital-fellowships⁠Fellowships in Clinical Artificial Intelligence : https://www.aicentre.co.uk/fellowships - National medical director's clinical fellow scheme (digital roles available): https://www.fmlm.ac.uk/programme-services/individual-support/national-medical-directors-clinical-fellow-scheme - Royal College of Physicians Flexible portfolio training (clinical informatics stream): https://www.rcplondon.ac.uk/projects/flexible-portfolio-training - Clinical entrepreneur programme: https://www.england.nhs.uk/aac/what-we-do/how-can-the-aac-help-me/clinical-entrepreneur-programme - Digital health leadership programme https://digital-transformation.hee.nhs.uk/digital-academy/programmes/digital-health-leadership-programme - Harvard health tech fellowship https://projects.iq.harvard.edu/healthtechfellowship/fellowship-overview - Stanford biodesign innovation fellowship:https://biodesign.stanford.edu/programs/fellowships/innovation-fellowships.html - Faculty fellowship: https://faculty.ai/fellowship-fellows/ - Entrepreneur First: https://www.joinef.com/ Events: - Royal Society of Medicine Digital health council events - HETT - Digital health Rewired - SomX events - Intelligent Health  - Med-tech World Blog article about different roles for clinician in digital health: https://medium.com/@annabellepainter/what-is-a-digital-doctor-db14522ec398 Range- Generalism book: https://davidepstein.com/the-range/