Let’s connect on Linkedin!

You’ve probably been in this situation: you talk to AI, it’s confident in what it’s replying to you, and then you say: “Actually, I think you’re wrong.” And the response is: “You’re right. This wasn’t accurate, the right response it ….”

In a healthcare discussion, an example could look like this: you describe your problems and ask for an interpretation. AI gives you a confident response, and then you say: “I forgot to mention, I also have this condition…” and the AI chatbot responds:

”That changes your picture completely ...”

This is called the omission problem.

The initial response wasn’t a hallucination. It was a conclusion made on imperfect information, resulting in a wrong recommendation. I don’t know about you, but when I’ve been in this kind of a situation related to my health, I left the discussion feeling utterly confused, with a sense that AI use… does me more harm than good.

Many developers agree that AI is not yet ripe for diagnosis. And at HIMSS26 Europe, the most commonly raised concern was the challenges posed by the patient use of general purpose models for advice. As Megan Coder, Board Member of The Society for Digital Mental Health elaborated around mental health: General models are optimized for conversational fluency and engagement, not designed for clinical accuracy or patient safety, don’t have domain-specific training on mental health data, lack crisis escalation protocols, are not clinically validated or evidence based.

With this in mind, I decided to re-assess the current state of AI in healthcare software development, how vendors are using AI, what hasn’t changed in terms of data governance needs, and more.

Leave a comment

Subscribe now

How can AI speed coexist with legacy slowness?

Every new AI model seems to be exponentially more capable than the previous one. At the same time, existing medical software such as EHRs, is transforming slowly, due to legacy architecture and components. These can be hard to re-design or refresh with new technology, as too many people are already using the existing product. The integrations in the background that would need to be re-done are time and resource consuming and risky, especially if there is a chance that an update glitch would cause an outage.

Let’s talk about that background on the data structuring level.

Original sources, deterministic models are still key

For healthcare data to make sense beyond the free text clinical notes, what clinicians say and record needs to be coded. ICD-10 codes classify conditions for billing and statistics, SNOMED codes are more detailed and capture deep clinical care details. And there’s many other classifications.

In the US market, IMO Health has for 30 years been translating clinical language in structured, contextualized data, and is present in every major US electronic health record. As mentioned by John Laursen, it’s not enough for data to be structured, it needs to have semantic continuity. This refers to "sugar disease" recorded in the 1990s and diabetes recorded today resolving to the same concept. Without having such a longitudinal layer, an AI model reasoning across a thirty-year chart isn’t accurate. It resembles reading a foreign language it assumes it understands.

At IMO, 60 clinical informaticists and physicians take care of accuracy and ensuring the longitudinal continuity. The point is: yes, AI can tell you what an ICD-10 code or a SNOMED code for a particular use-case is, but if you want to make sure the outputs in your software are consistently correct, you should search for and connect the right code from the original source.

AI can make many processes easier. But original sources, and plenty of manual checking, is still required for the development of scalable solutions.

What now separates a good clinical AI system from a bad one happens underneath the model: how is standardisation done, which terminology services are used, how is the context structured.

Subscribe now

The validation challenge

Robert Tovornik, Innovation Lead at Better, points to another puzzling characteristic of AI solutions. The AI model race is largely settled; in his view frontier models are now performing on a more or less same level.

What the sector hasn’t resolved yet, is validation. Deterministic software can be tested against expected outputs. AI is non-deterministic. Every output, even with the same prompt, will be different. This means that trust becomes experiential - accumulated the way you come to trust a colleague. The more experience you have, the more you can judge what produces quality outputs. This is a challenge for procurement processes and certification regimes, which require demonstrable specifications. Nobody has a good answer for this challenge just yet.

And then there’s omissions

Remember the story from the beginning? Omission has no screenshot. It produces an answer that is fluent, plausible, internally consistent, and wrong in a way that is invisible unless you already knew what was missing. That is the CDS problem of the next few years, and it is a data layer problem before it is a model problem.

AI expert and extrepreneur Elena Ikonomovska, built a company around the challenge of the complete context of patient information. She sees chatbots patients increasingly turn to for answers, as the more dangerous new entrants into medicine, not because they are unhelpful but because they are convincing.

Ikonomovska describes what she calls "faithful hallucinations" — confident, plausible outputs that fabricate detail patients have no expertise to challenge and clinicians have no time to unpick. Generative AI in medicine, she argues, risks widening health inequity, since better-resourced patients will access validated clinical tools while others rely on free chatbots.

Leave a comment

What about clinicians, who increasingly use AI for clinical reasoning? Philippe Habets, CEO of EvidenceHunt, Dutch-based provider of a search engine for the latest medical research sees the omission problem in the prompt. Any published scientific paper draws conclusions from results of a carefully selected group of patients. The patient in front of a clinician is n=1, a puzzle that needs to be put in context. If you leave one comorbidity out of the question posed to AI, the correct answer might change. And the model has no way of knowing which condition it wasn’t told about. It cannot flag the absence of information it never saw.

EvidenceHunt’s response is to refuse case-level advice altogether. If a clinician does describe a concrete patient case and asks for a response, the system reformulates theclinical question into a generalisable research question answered with citations.

Subscribe now

So what’s easier in software development?

According to Robert Tovornik, testing ideas, building prototypes and showing the end customers the art of possible before actual development are enhancing collaboration between IT experts and clinicians. Some clinicians even come better prepared in IT discussions, since they use AI to better understand IT. This doesn’t come without potential inaccurate interpretations, but it’s changing the development dynamic.

What this implies in practice

For anyone evaluating or building clinical decision support, the four conversations suggest a fairly concrete set of questions — most of which are about the data layer, not the model.

• What did the system not use, and can it tell you? A system that can say “I had no medication history for this patient” is more useful than one that scores higher on a test set.

• Which parts of the pipeline are deterministic, and why those parts? If coding, mapping and terminology resolution sit inside the probabilistic layer, ask what failure looks like and how it would be traced.

• How is the record’s quality being maintained, independent of the AI? Problem list hygiene, terminology currency and normalisation are unglamorous and are now the binding constraint on everything built above them.

• What happens to the query? Where does a clinician’s question go, what is retained, and does the vendor’s business model depend on that retention?

Tune in to the full discussions!

Faces of Digital Health — website · Spotify · Apple Podcasts · LinkedIn

Let’s connect on Linkedin!

There is no shortage of challenges with healthcare AI:

• Out of 38 OECD - OCDE countries, only 7 have a formal AI strategy. Furthermore, only 11% have workforce upskilling programs, and 9 conduct Health Technology Assessments (HTA) on AI, with 8 in the process of updating their Health Technology Assessment (HTA) frameworks to evaluate AI’s clinical relevance, cost-effectiveness, transparency, and long-term safety, according to the 2026 Scaling AI in Health OECD Report.

• No internationally comparable indicators exist for AI deployment in health systems.

• Off-the-record, we hear that not just patients, but clinicians sometimes add identifiable personal information in general large language models,

• Policymakers are struggling to create regulation at the pace of innovation, which has accelerated significantly with AI: in the healthtech space, the risk-benefit discourse is benefit-skewed.

• Healthcare providers feel frustrated by patients using AI, asking more questions or taking insights it gives them as the sole truth.

At the same time, which is what I am exploring through the Agentic Patient series, patients don’t necessarily trust AI, but are definitely excited by it, because it can fill the gaps between visits, care coordination and more.

Patients are saying yes to AI.

In 2025 The European Patients’ Forum ran a European survey on patient attitudes toward AI in healthcare, translated into multiple languages, with nearly 1,000 responses. The headline finding is that patient enthusiasm is high: 98% responded positively about AI’s potential.

The dominant concern is bias and the possibility that AI-driven decisions could be wrong.

Patients want direct communication from healthcare professionals about AI’s role in their care, and where appropriate, AI-specific consent forms.

While I disagree, that doctors should be the primary point of information about AI for patients, it is true that patients are most likely to talk about AI in the consultation room, leading to an actionable takeaway for AI deployers and health systems: the patient trust problem is solved less by communication campaigns and more by getting clinicians equipped to talk about AI specifically, in patient-appropriate language.

Subscribe now

Youth aren’t just a future workforce. Why are we treating them as such?

Another aspect not discussed enough, is the impact AI has on youth. As a mother of a young child, I an increasingly finding myself thinking how most adults seem largely unaware that technology is shaping the youngest generations in ways that are fundamentally different from how it has shaped adults.

At HIMSS Europe, Tara Steele, Director, Safe AI for Children Alliance and Marc D. Ritter, CEO of AWE Wellness, emphasized that children are vulnerable to AI systems and AI companions, and can form emotional attachments to them very quickly.

US social psychologist Jonathan Haidt, started a movement that calls for changes in how children use technology and experience childhood. As he explores in his book The Anxious Generation, the rapid shift from a play-based childhood to a phone-based childhood has contributed significantly to rising rates of anxiety, depression, and other mental health challenges among young people.

However, it seem like we are not paying enough attention to this, especially when it comes to policy.

Digital Transformations for Health Lab has been tracking for over a decade how national digital health strategies treat young people. The pattern is stubbornly consistent, says Aferdita Bytyqi, Executive Director and Founding Partner of DTH Lab. Ten years ago, youth health and wellbeing was largely absent from national digital health strategies; today, looking at AI governance frameworks, she sees the same omission reproduced. When young people appear at all, it is through the lens of skills, education, or innovation — a future workforce to be trained — rather than as current users whose health and development are already being shaped by these systems (Read more in the report Shaping AI Governance for Young People).

Aferdita’s most direct argument why this needs to change is numerical: in many lower- and middle-income countries, 50–60% of the population is under 35. Designing AI frameworks without them is not a moral question; it is a sample-size problem.

Share

What is or should be next for health policy and AI?

HealthAI - The Global Agency for Responsible AI in health recently published a paper on translating the EU AI Act in action. Five recommendations include:

• Countries need to optimize AI supervision - Countries have horizontal AI supervision (under an AI authority) and vertical AI supervision (within the health sector). We should explore joint mock exercises across governmental institutions to stress-test the coordination before real deployments expose the gaps.

• EU should address regulatory barriers and cross-nation reimbursement Health technology assessment and reimbursement pathways are still configured for a different category of product. Until that is fixed, AI scales for whoever can afford the regulatory entry cost — which is not most builders.

• We need clarity on the interplay between EHDS, MDR and EU AI Act The AI Act, MDR, and European Health Data Space all touch AI in health, but the different director generals within the Commission need to be coordinating internally, and lessons learned at member-state level need to flow back into framework design.

• Let's formalize knowledge exchange instead of duplicate efforts

• To scale we need multi-stakeholder engagement beyond borders - civil society and patient organizations need to be active actors in the legislative and deployment process — not tokens at the consultation stage.

AI is here to stay. As with any innovation, unwanted side effects occur, and it takes time to mitigate them. As Ilona Kickbuch mentioned at the Responsible Scale of AI in Health conference in Madrid, referring to the tobacco/lead/social-media analogy, “accountability historically arrives via litigation rather than foresight.”

Warnings are lining up. Society for Digital Mental Health cautions that general AI models are optimized for conversational fluency and engagement, not designed for clinical accuracy or patient safety. We need more purpose-built mental health AI developed using domain-specific clinical data, built with clinical expertise and safety-oriented design, clear boundaries around appropriate use, crisis detection and escalation protocols.

Personally, I am hopeful that the lessons learned from the harmful impacts of algorithms and social media will help us respond faster and more effectively to the challenges posed by AI, for a broader predominantly positive impact.

See also…

Recorded live at health.tech in Basel, this panel unpacks the convergence reshaping clinical software: ambient AI scribes, agentic AI in healthcare, on-device LLMs, and the regulatory drag (MDR, EU AI Act, EHDS) that is widening the gap between what clinicians actually use and what hospitals are allowed to buy.

Host Tjaša Zajc is joined byJhonatan Bringas — CEO & Founder, Lapsi Health (Kaiku: FDA-cleared AI stethoscope, ambient scribe and clinical assistant in one device), Blaž Triglav — CEO, Mediately (drug information platform, 1M+ HCPs across Europe), Amanda Herbrand — Clinical data modelling consultant, formerly University Hospital Basel.

The conversation covers:

• Why EHR data fragmentation is the precondition AI hasn’t solved

• Shadow AI: why clinicians trust ChatGPT more than enterprise tools (and the agency hypothesis behind it)

• The asymmetry that will break European medtech: applicants using AI to build, regulators forbidden from using AI to assess

• On-device AI, ambient computing, AGI in clinical workflows — and the de-skilling risk no one wants to discuss

Let’s connect on Linkedin!

I was recently invited to give a talk about patients and AI for AMS Healthcare in Canada. Due to a relapse of my chronic condition in the past year, I have plenty to share on the topic. But since I’m just one patient, I started searching for and interviewing other patients in different contexts and countries on what they use AI for, which tools, and what kind of prompts. I wanted to know exactly what they’re asking and what other patients might find helpful too. Let me start with my story.

The Hype Curve of Living with a Chronic Condition and AI

My relationship with AI in the last 12 months resembles the curve of falling in love: it started with curiousity when my health started deteriorating. I have been living with IBD for 22 years, without bigger issues in the last 15 yeas, when I started suffering from increased problems that required stronger and stronger therapies. I got the first biologic therapy, and had to wait for it to start working. The challenge was that the treatment I decided for, usually required 6-14 weeks to take effects.

While I was waiting to get better, my admiration for AI started growing. When I was between visits, I had no access to my specialist. But I had bloody diarrhea, was loosing a lot of blood and had 100 questions AI was eager to answer. It prompted me to request blood tests from my GP, which revealed that I urgently need an iron IV infusion. It outlined what to expect in my patient journey during those grueling weeks while I waited to see if the treatment would be effective. It wasn’t and we had to change therapy two more times in the next 3 months.

Once my health improved, I started researching AI with more curiosity rather than desperation. I fed different chatbots the same prompts about my health, which was still on the path to stable. However, despite the same prompts, I started getting subtly different outputs from different systems. Gemini suggested I should take it easy with exercise, while Claude implicated that my ability to exercise isn’t problematic for my health. Which one was it? I fell in doubt and confusion about trusting AI.

Then, after months of relative peace, I suffered a sudden complication - unexplained blood in my stool, quite a significant amount of it. Instead of going straight to ER I consulted AI. I wanted to get an idea about the seriousness of my situation, my options, possible outcomes and what doctors would do if I did seek medical care.

After a night in the ER, where much of what had been suggested in an earlier AI conversation became reality, my trust in AI landed somewhere in between. I see the value of chatbots, but also believe caution is essential. AI outputs need careful verification, and patients should use them with that in mind.

One thing is important to mention here: when I used AI to know what to expect in the ER, I used it to be informed. Eight months earlier, when I was between specialists visits, I used it out of desperation and because this was my only consultant and I couldn’t get answers from the healthcare system when I needed them. I think this is crucial: if patients had a choice between chats with clinicians and AI, I know I would choose clinicians. But sometimes that’s not an option and AI is the only voice available.

The Collective Knowledge and Challenges

I believe AI can be very helpful for patients and is changing the information asymmetry between clinicians and patients that was huge in the past. But just changing access to information doesn’t give patients the medical knowledge to interpret AI outputs.

Let me give a concrete example. My father-in-law was worried about ongoing chest pain that hadn’t been explained after an ER visit, where all the standard tests had been done and come back inconclusive. I gathered details: his medications, symptoms, and medical history and gave to to an AI chatbot. But instead of asking AI for a single interpretation, I asked it to assess the situation from five perspectives: geriatric medicine, gastroenterology, hematology, vascular medicine, and cardiology.

The possibilities quickly became complex and overwhelming; far beyond what I could interpret. In the end, the only advice I could give my father-in-law was to:

• monitor his symptoms,

• make a diary of when the pain starts, how long it lasts, what type it is and how it changes and

• take that data to his physician.

Caution is essential with AI. But the fact is patients are and will use it. So how can we get the best out of it?

Share

Leave a comment

The agentic patient comes in

As part of Faces of digital health podcast I launched a podcast series called The Agentic Patient that explores how patients use AI. News articles at the moment fall into two buckets: either hyping how AI solves unsolvable medical problems or warning of dangers such as increased anxiety, or patients not seeking medical care when they should.

With The Agentic Patient, I want to get to actionable insights and guidelines. Write down and share which tools patients use, which prompts when chatting with AI, the safeguards they rely on, and what more can be done to ensure its safe and responsible use. This is not a project about glorifying AI, but a project hoping to contribute to health AI literacy.

So far, based on my research, patients using AI fall in three groups:

Subscribe now

• Informed Collaborators: patients who have digital health literacy, think critically and steer AI in the direction helping them structure their data and become a valuable collaborator with the doctor, enabling faster better outcomes.

• Minimizers: AI is tricky: it will go in the direction you steer it to. If you seek reassurance, AI will provide it. These types of patients might delay their treatment and suffer complications.

• Cyberhondriacs: These patients are prone to anxiety, and will go down the spiral of the worst possible outcome even for minor issues. These patients might drive healthcare utilization and add unnecessary cost to the system.

The operational question for health systems in 2026 is not whether to engage with patient AI use. How do we make sure they get the best out of it?

So far, through the interviews with patients, I came to…

6 recommendations on how to use AI as a patient

Share

1. The T-LICC formula

Masterclass recently released a course, prepared by doctors, to help people get the best out of AI. The immediate advice is to think what the doctor might ask you upon a visit. And that is: when did symptoms start, where is the problem, how did it change over time, add relevant context in your medical history.

2. Use a Paid version of a Chatbot

This will enable better data privacy, but more importantly, will give you access to the latest models. They tend to improve in quality exponentially with each version.

Leave a comment

3. Create an archive of your discussions

Russ Read Barrow from UK, author of https://fcancerwith.ai/ created a toolbox for managing his cancer. His advice is to not give too many jobs to the same model. In his journey, he used Claude for symptom tracking, Notebook LLM for searching through history and Google docs for copy-pasting all his chats with AI. He learned the hard way that chatbots have a history cut-off - the chats get lost, while AI retains the context in its memory, to be able to talk to you further.

4. Use guardrails

Using guardrails refers to giving AI concrete boundaries for its research. While in Canada, a clinician said simply the prompt “Don’t hallucinate”, massively improves AI outputs.

A more concrete limit would be to ask the chatbot to “Base your answers on the German clinical guidelines” if you live in Germany. Or, when you work with research, tell AI to “add at least 5 credible sources to any conclusion”. Don’t forget AI doesn’t only hallucinate answers, it can hallucinate links. It’s key to always check the sources AI draws answers from.

As Dale Atkinson, a cancer patient advocate from the UK warns: “AI pulls information from Facebook groups and social-media doctors who may not be credentialed. It uses the logic that the more instances something is mentioned, the more likely it is to be true.This is the opposite of how evidence works. The quiet doctors doing hard work are the ones to look for, and the model will not surface them unless you force it to”

Dale was given a terminal cancer diagnosis, but used a research approach to solve his cancer. He took the research first, AI second direction.

ChatGPT was his literature triage layer. He fed in his diagnosis and medical letters, asking "if you were in my shoes, where would you start?" and getting a reading list. He then manually read roughly 4,500 papers over three-to-six months, initially cover-to-cover, later skimming for the sections that actually mattered.

It is worth mentioning that before getting sick, Dale was a financial crime investigator, which gave him necessary skills to read dense regulated documents, which he was then able to transfer to reading medical papers.

Today, Dale doesn’t have signs of cancer and talks to patients on a daily basis.

One of the things he noticed is that patients use one AI model for prompt preparation and another for execution. For example, they ask one AI chatbot to ask them 20 relevant questions and once they answer those, they use that output to continue research in a different model.

5. Red team analysis

Demetri Giannikopoulos is Chief Innovation Officer at RadAI, overseeing clinical integration of AI in radiology across US health systems. He has lived with multiple sclerosis for roughly twenty years. His wife is a nurse practitioner and a cancer survivor. Demetri doesn’t use AI for his condition management, he did however use it to find the best insurance plan for him.

Demetri’s opening rule: “Never ask a simple question because you’ll often get a simple answer, and that can take you down the wrong course.”

For insurance, a single copy-pasted plan summary missed the loopholes; only pulling the full underwriting document surfaced the real cost structure. His analysis used 50+ pages of source material, cross-compared three plan options simultaneously, and factored in his wife’s expected utilization.

His key advice is also to ask AI to “perform a red team analysis”. This is a term AI will recognize and it means it should create a critical analysis of all the outputs it gave so far and any missing critical points it hasn’t touched upon.

I would add the red team analysis prompt should be used with caution, because it could also give you a grave diagnosis for a minor symptom and send you down the anxiety spiral.

6. Decompress

This tip comes from me personally: when you get too “sucked in” AI discussions, especially if it start confusing you, take a step back. Take a break.

Never forget, you might be going in the wrong direction without even knowing it

The four patient cases above spoke about how AI helped them. But with the true aim of increasing AI health literacy, the Agentic Patient Series covers less positive examples as well. Diana Ferro, a clinician-researcher and data scientist working at the intersection of health, technology, and ethics at Bambino Gesù Children's Hospital in Italy, has seen problematic impact of AI as well:

• parents using AI to deny rare-disease diagnoses,

• adolescents using AI as a pro-eating-disorder coach,

• young people with weak support systems finding AI easier to talk to than a clinician, including, she notes, in contexts tied to self-harm.

She vividly describes: “We are witnessing wars between two people using two AIs to prove each other wrong. The parent’s ChatGPT contradicts the oncologist’s AI decision-support tool, and the child, pays the price.”

I truly believe that by sharing best practices of AI use, we can help more patients achieve better outcomes. However, one should never forget that while AI can assist in organizing information, summarizing data, and helping users prepare for medical consultations, it is not a clinician and does not provide medical advice, diagnosis, or treatment.

I will keep sharing best practices. To go beyond what’s shared in this newsleeter, go to more detailed discussions summaries and recordings of each interview on the Agentic Patient tab on the Faces of digital health podcast website.

Meet Dale and Russ at HIMSS Europe!

Dale Atkinson and Russ Read Barrow will join HIMSS Europe. Dale will share his story in on the 20th May, in the session: Money talks: Financial Impacts of EU Policy, where he will highlight what policymakers should invest in to have the full picture about patients and consequently improved outcomes and lower healthcare costs.

Russ will join the panel discussion How do we define appropriate autonomy vs. full autonomy? in the session Working With AI Agents: Redefining Roles, Autonomy, and Trust in the Workforce, on 20th May 16:30-17:30.

Join us in Copenhagen!

Let’s connect on Linkedin!

In 2026, AI use is widespread. In healthcare, a new unicorn breaking records is OpenEvidence. The startup was valued $12 billion in January 2026, and is the most widely used medical AI among verified U.S. physicians. It’s a free tool for physicians, based on peer reviewed articles from prominent medical journals such as NEJM, JAMA, NCCN, and others.
On the patient side, Dave DeBronkart is gathering patient stories on AI use. Patients, such as Steven Brown, are building multidisciplinary clinical AI agents to help them guide their patient journeys (see him explain how, in the discussion from NextMed Health 2025).

I don’t often get overly excited about new technologies or trends. Most of them pass through the familiar hype cycle, and only a few ultimately deliver lasting impact. But in 2025, through my own patient experience and complications related to IBD, I saw firsthand how profoundly AI can impact a patient story and outcomes.

Patients are already using and will continue to use AI not because they prefer it to their doctors or inherently trust it more, but because healthcare systems around the world are increasingly falling short. In that gap, AI can offer meaningful support: helping people think through problems, generate ideas, ask better questions, and sometimes even find the answers they’ve been struggling to get.

So I will say it: I believe AI agents will transform healthcare to the next level. And we need to start thinking how they will be coordinated and governed.

Subscribe now

How Are AI Agents Helping Healthcare

(If you’re familiar with the basics scroll down to operating systems for agents section)

Ali Parsa, CEO and Founder of Quadrivia sees agentic AI as a solution to a very old problem in healthcare: the structural imbalance between infinitely “elastic” demand (you’ll do anything when your child is sick) and constrained supply. For him, AI, specifically autonomous, real-time clinical agents offers a way to create elastic supply of clinical skills.

He is careful to narrow the scope for Quadrivia: not “everything doctors do,” but the 20–30% of clinical tasks that are repetitive, rules-based, and suitable for automation. The vision is not just co-pilots and scribes, but agents that can autonomously follow structured workflows, talk with patients, record and summarize, and escalate to humans when needed, reliably and within 1–1.5 seconds so they still feel human.

Ali describes AI agents as programmable junior colleagues: configurable, controllable, and anchored in a hierarchical knowledge base that prioritizes clinician- and institution-level sources before general models.

But what happens when suddenly different doctors use several different agents? How do we prevent coordination and data chaos, similar to the interoperability mess we’re still struggling to fix, and may take another decade to resolve?

What we need to think about already today, is an operating system for AI agents.

Share

Risk management for AI agents

Bart de Witte has worked in healthcare for over 25 years. As a European who views access to healthcare as a fundamental right, he has spent nearly a decade advocating for open-source AI and greater transparency in AI models. He believes this is essential to keeping healthcare accessible to all, rather than allowing it to become dependent on income.

Bart recently visited Ljubljana, and I picked him up at the airport. Somewhere between traffic lights and business districts, we ended up having a conversations about where healthcare AI is heading with agents.

Bart calls AI agents the first healthcare AI technology with the potential to deliver real return on investment quickly; especially in areas like scheduling, triage, and administration.

Unlike AI scribes that are summarising notes, drafting letters, agents, with proper access controls, can complete a task from start to finish, by calling EHRs, scheduling systems, clinical databases. They can chain tasks, reflect on outcomes, and trigger workflows without a human pressing “enter” each time.

However, agency also introduces risk. Once an AI system can act autonomously, questions multiply:

• Who controls what the agent can access?

• How do you audit its actions?

• Where does sensitive patient data live?

• Who is liable when something goes wrong?

Healthcare is very good at asking these questions. It’s much less prepared to answer them at scale.

How does healthcare infrastructure need to look like to support agentic AI or OS for agents?

Agentic AI is promising in theory, but healthcare needs to prep it’s IT systems and data for it first. One of Bart’s core concerns is that most healthcare IT systems were never designed for agent-based workflows. Monolithic EHRs embed business logic deep inside proprietary systems. They work well for documentation and billing, but they struggle when workflows need to cross systems, vendors, or even organisations.

Adding AI “features” on top doesn’t solve that. Fo agents to pull information from multiple systems, reasons over it, take action,and learns over time, healthcare systems will need a control layer for AI agents, which is what Bart calls an operating system. Think of it as an environment that governs access, security, audit trails, compliance, and collaboration between agents.

Smaller models, better outcomes?

Another counterintuitive insight Bart believes in is that bigger AI models may not be the future of clinical AI. Emerging research is showing that multiple small, specialised agents, each trained for a specific task, can outperform large general-purpose models when orchestrated together. “Think of it as swarm intelligence: a cardiology agent, a nephrology agent, an oncology agent collaborating on a case,” Bart illustrates.

This matters for regulation. General-purpose AI can’t realistically be certified as a medical product. But narrowly defined agents, doing one thing very well, can. Validation, transparency, and accountability become tractable again. It’s a vision that aligns well with healthcare’s reality: specialisation, boundaries, and trust earned task by task.

Privacy by design through the architecture

Bart is building an OS system with his startup Isaree, which he co-founded together with entrepreneur Mandana Ahmadi. To be GDPR compliant from the get go, Isaree aims to run agents directly on clinicians’ phones, tablets, or desktops, enabling local processing, instead of sending sensitive patient data to cloud APIs. This way:

• data doesn’t leave the device,

• surveillance-based business models break down,

• compliance becomes a design property rather than a legal workaround.

Can small agents scale?

At the moment, most IT systems in healthcare are closed, rigid, hard to change and consequently unfitting to individual doctor’s preferences. With agents build on top of open data specifications, healthcare may move toward marketplaces of certified agents, where clinicians choose what fits their workflow, might even build their own agents, and hospitals control access and governance.

Not every clinician will build their own agent, just as not every smartphone user builds apps. But the ability to customise, combine, and adapt tools lowers the barrier between idea and implementation dramatically. And vibe coding is making building apps increasingly easier, even without coding knowledge.

The real question is not whether agents will be used in healthcare, but under what conditions. And how soon, will solutions such as OpenEvidence be available to patients as well. Why wouldn’t patients build agents on top of their data if they’re getting increasing access to their data through national portals in Europe and vendor patient portals in the US? As written by digital health KOL Hale Tecco for Second Opinion, withholding high-quality AI tools from patients only reinforces a more paternalistic model of medicine.

No one is more passionate and motivated to solve a patient case than a patient herself. So why would they not work together with clinicians, in a true partnership? The very model healthcare has been striving toward for years.

Share

Leave a comment

For more clarity on AI, join HIMSS Europe

The HIMSS26 European Healthcare Conference and Exhibition is the most important European event for digital health leaders, offering unmatched education and bringing together EU policymakers, executives, clinicians and innovators.

AI related topics you will learn about in Copenhagen, include:

• How are AI agents changing the workforce

• How do we define appropriate autonomy vs. full autonomy?

• Operationalizing LLMs in Healthcare: Early Results and Hard Truths

• How are vendors preparing to the EU AI Act?

• How Are EU Countries Building AI Regulatory Sanboxes for AI testing?

and more!

See you in Denmark?

Let’s connect on Linkedin!

Ambient speech tech has exploded in healthcare but mostly as “documentation glue”, helping clinicians with note-taking through speech recognition. But what if the real opportunity isn’t the transcript at all?

I recently spoke with Henry O’Connell, Founder and CEO of Canary Speech. Frankly, because I saw the company claims their tech can diagnose conditions such as Alzheimer’s disease from nothing more than a 45 seconds long discussion. Too good to be true? Not necessarily.

As Henry argues, the last 25 years of “voice biomarker” progress stalled for a simple reason: the field focused on what people say (words) instead of how the nervous system produces speech (signal).

From words to the “primary data layer”

O’Connell traces the roots of the space through decades of speech and language work, from early NLP to Nuance and far-field speech systems. For a long time now, the industry could find correlations by analyzing spoken words and their textual representation. But it never produced widely adopted clinical products. The limitation, says O’Connel, is data density: word-based analysis gives you a few hundred-ish elements; analysing voice characteristics however, yield millions.

Share

Canary Speech’s bet was to turn the industry paradigm upside down: ignore words and instead analyze the articulatory system; how the central nervous system coordinates breath, vocal cords, tongue, and timing to generate language. It’s the same human machinery regardless of language, and it’s the layer clinicians often “read” intuitively like when you can hear someone isn’t okay even if they say “I’m fine.”

45 seconds, 13 million data elements, thousands of features

Surprisingly little audio is needed for diagnosis: 45 seconds of conversational speech captured during a clinician-patient exchange. Canary extracts 2,590 voice features every 10 milliseconds (with a 25ms window sliding by 10ms), which adds up to roughly 13 million data elements for the model to evaluate.

Those features include characteristics tied to vocal cord vibration (e.g., jitter/shimmer), respiratory sounds, resonance, and how those properties change over time; plus derivatives (how fast features change and recover). Many of these features were already known in speech science. Canary focuses on stacking multiple extraction approaches and validating them clinically.

How the “ground truth” gets built

Canary partners with clinical teams (neurology, psychiatry, specialty services), captures patient audio during clinical interviews in medical practice, and uses clinician diagnoses and test batteries as labels. This is then the ground truth. Machine learning then learns correlations between the diagnostic labels and the extracted features.

The result is an additional insight for the clinician. Tech is meant as clinical decision support, not a standalone diagnostics tool.

Accuracy: very high for progressive neuro, harder for behavioral health

Progressive neurological diseases like Huntington’s, Parkinson’s, and Alzheimer’s can be detected with 98%+ accuracy, says O’Connell. Behavioral health conditions such as anxiety or depression are more nuanced. The accuracy there is “in the 80s.” A central point: the system can measure multiple dimensions at once e.g., cognitive decline plus depression, fatigue, and anxiety without extra time burden on the clinician. Still, an insight can prompt the clinician to ask questions they otherwise wouldn’t.
One example O’Connell shares: “A doctor is meeting with a young mother six weeks after her birth. The mother is being very much like every mother on the planet Earth is often. Focused on her child, focused on the responsibility of caring for that child, feeling that obligation, a good one, and a responsibility to do so. The doctor says to her, how are you doing? Her focus is all over here and she says, I’m doing fine. Canary’s speeches assessment came back at virtually the same time and said her depression levels were clinically as high as they could get. So the doctor said: ‘Why don’t we explore that a little bit more?’ And by the time they had a few more questions in, the woman related how deeply depressed she was.“

Why adoption lagged and why it’s moving now

If the results can be strong, why hasn’t this type of tech become routine? O’Connell points to earlier research norms that relied on structured read speech (scripts) borrowed from speech pathology. Since, as claims O’Connell, the brain regions involved in reading vs conversing differ, script-based studies can be counterproductive for building conversational biomarkers.

Canary avoids read speech entirely.

In addition, AI and infrastructure shifts, such as real-time streaming, fast compute, better transcription and ambient tooling are accelerating development outside of text analysis. Canary’s integration with ambient documentation workflows (e.g., a clinical copilot listening anyway) turns voice biomarkers into “no extra workflow” intelligence.

Subscribe now

Share

The ethics question: screening, incidental findings, and trust

What if a clip of a conversation reveals more than an individual would like to know? Similarly as with genetics tests, where, clinicians can face a dilemma of finding and sharing information outside of the scope of the initial diagnostic purpose. Canary’s framing is that voice biomarkers describe what’s present in the moment, not a future risk. That matters in primary care screening, where objective signals can prompt better follow-up without relying on subjective questionnaires alone.

Beyond diagnosis: workforce safety and operational signals

The most unexpected use case for voice biomarkers may be in-room monitoring for aggression risk. Unfortunately, in U.S. hospitals, violence against nurses is a serious operational and safety issue. Canary can contribute to minimizing incidents by indicating a specific room has aggressive individuals, enabling teams to enter these environment with more caution (for example, two people enter, not a nurse alone) and prepared to de-escalate conflicts.

Remote monitoring and clinical trials: faster screening and outcome measurement

For pharma and trials, Canary can support:

• Pre-screening participants (e.g., who is or isn’t demonstrating mild cognitive impairment)

• Longitudinal tracking to detect changes over time (e.g., whether cognitive function improves with a therapy)

• More direct outcome measures than proxies like plaque reduction alone

The company is planning to expand the number of diagnoses it covers, adding models for conditions such as PTSD, postpartum depression, ADHD and autism, pain, and COPD. It also plans to expand into additional languages, ensuring each one is properly validated before launch.

What’s next: global rollout and consumer wellness

Canary is already deploying in multiple regions (U.S., Canada, Japan, UK/Ireland, parts of Europe, and expanding in South America and Arabic-speaking markets). They plan on deploying a wellness product in 2026, which will still be built on clinically developed models rather than scraping uncontrolled public sources.

Bottom line: “Voice in healthcare” isn’t only about transcription. If clinical validation scales, voice biomarkers could become a quiet layer of decision support: always on, non-intrusive, and potentially transformative for early detection, behavioral health, and even workforce safety.

Leave a comment

Subscribe now

Share

2) Time stamps for the podcast version (chapter-style)

00:12 Intro: voice biomarkers beyond ambient documentation
01:03 Potential vs reality: what voice can (and can’t yet) prove clinically
01:54 Why earlier voice-biomarker work focused on words—and why it stalled
06:34 The “intuition” problem: we hear mood without words
07:33 Canary’s clinical-first approach and global clinical partnerships
08:39 Do you need long-term data? (Agatha Christie example)
09:41 Method: ~45 seconds of conversational speech, ambient capture
10:36 Scale: 2,590 features every 10ms (~13M data elements)
11:02 Where the features come from (vocal cords, respiration, resonance)
14:36 How models are built: IRB, clinician ground truth, ML correlation
17:30 Accuracy + adoption: why it’s not standard practice yet
18:22 Reported performance: 98%+ neuro, ~80s behavioral health
22:44 Culture/language bias concern: why validation per language matters
24:22 Guardrails: validate every new language and population testing
31:07 Why read-speech scripts misled the field; conversational-only stance
34:16 What changed: AI, compute, real-time streaming, workflow fit
37:10 How it’s used: screening vs suspected disease; incidental findings
39:24 Primary care example: postpartum depression flagged despite “I’m fine”
47:56 Wellness/employee use: de-identified dashboards and ethics
53:42 Technical requirements: device capture, signal-to-noise checks
57:10 In-room monitoring: aggression risk signals for staff safety
1:02:22 Clinical trials: pre-screening + measuring therapeutic impact
1:07:14 Global rollout: regions, languages, and partnerships
1:09:42 Consumer access: wellness product planned in 2026
1:12:32 Wrap-up: why this matters as cognitive health needs grow

Let’s connect on Linkedin!

My vacuum cleaner doesn’t have it’s one name. Apparently, that’s rare, since 80% of Roomba robots have names.

In 2026, for $20,000 or a $500 monthly subscription, you can get your own home robot – Neo. When I first saw this, I was thrilled. I would love to have a machine that does the dishes three times a day instead of me, makes sure the vacuum robot doesn’t get stuck somewhere, and can wash and fold the laundry. It seems this is close to becoming a reality.

But hold the excitement just for a little bit.

What is NEO, the robot you can buy already, capable of?

If you watch the video review by Wall Street Journal, NEO is still very basic and not autonomous. Many tasks are done by a human in the background and the robot’s “eyes” are cameras that help the human operator perform movements. Early adopters will essentially help the company, 1X Technologies, gather data for AI training and support innovators building the road toward robotic independence.

By buying Neo, company employees can see you and your home through the robot’s cameras. However, as emphasized by the CEO in the WSJ interview, each device includes several layers of privacy controls: rooms can be off-limits, and privacy concerns are presented as a small trade-off for the benefits. Especially over time, as updates are rolled out based on field data.

If you look at the December 2025 robotics exhibition in Japan, robots from various vendors still seem to have very limited capabilities. Nonetheless, the race to build a general purpose, autonomous humanoid robot is on. Tesla is building its version called Optimus with the aim of producing 10-million units annually by 2027.
Boston Dynamics is focusing on humanoid robots for manufacturing, and China’s TechnologyX AI is making waves with humanoid robots capable of astonishingly human-like movements. These are just a few of the players.

This made me wonder: is robotics the next big change we’ll introduce in healthcare, and what do we already know about its potential impact?

Robots in healthcare

Robotics in healthcare is not new. For decades, hospitals have relied on machines—from surgical robots, to robots that quietly deliver medications from pharmacy to ward, to automated cabinets that manage and safeguard medication stock behind the scenes, to robots that interact with patients and keep them company.

But let’s take a closer look at humanoid robots and animal-shaped robots, which we tend to anthropomorphise (that is, attribute human characteristics, emotions, intentions, or behaviors to non-human entities such as animals, machines, or software).

The healing effect of robots

In her book The new breed, anthropologist Kate Darling shares extensive research on our attitudes toward robots. As mentioned earlier, 80% of robotic vacuum cleaners have names. As long as robots aren’t too human-like—which can feel creepy—we tend to be intrigued by them. History shows that we become attached to animal-shaped robots quite quickly, especially when they look cute.

Share

If you were born before 1990, you might still remember owning cheap and affordable Tamagotchis and Furbies, which are still available today for less than $50. More sophisticated consumers with deeper pockets could opt for more expensive robots such as Sony’s Aibo dog, which today costs around $3,100, plus a subscription for AI and cloud enhancements.

As a granddaughter with a grandmother living with dementia in a care home, I’m particularly intrigued by Paro the baby seal ($8000), Tombot dog (price unknown) and Joy for all cats and dogs ($200 USD). These animal-shaped robots are designed specifically for environments unsuitable for live animals, such as hospitals and care homes. With life-like features: purring, barking, and responding to voice or touch their therapeutic effects include psychological enrichment and joy for people with dementia, emotional and behavioral distress, PTSD, social isolation, and for patients in palliative and hospice care.

Subscribe now

Knowing all these positive impacts, the question remains: why aren’t we seeing robots everywhere—in care homes, or supporting the understaffed caregiving workforce?

As with any technology, integration and implementation are key.

What does anthropology say?

Tanja Ahlin is an anthropologist of health and technologies and author of Calling Families: Digital Technologies and the Making of Transnational Care Collectives.

She is currently researching the use of robots for people with dementia and the impact they might have on mental health.

In a recent discussion for Faces of digital health she shared her findings so far. We discussed the international use of robots, the surprising lack of adoption, why we trust robots and AI, and what this means for the future.

Surprises from Japan, Netherlands and Denmark

Let’s start with Japan, often treated as the default reference point, when it comes to robots. James Wright’s Robots Won’t Save Japan book argues that the image of Japan as a robot-saturated society is partly strategically constructed. As part on an economic plan, Japan itself has been working a lot on coming up with an image of itself as highly technologically developed. A less-discussed but crucial policy dimension of care robotics emerges in the Japanese context, where robots were not only imagined as a technological solution to population ageing, but also as a political alternative to immigration. “The debate was often framed in stark terms: do we bring immigrants or do we create robots?” says Tanja. Robotics promised a way to preserve cultural homogeneity while addressing severe workforce shortages in elder care. Yet everyday care realities proved far more complex than this binary suggested. Despite heavy investment and global visibility of care robots, the approach did not deliver the level of support needed in practice.

Inspired by promising research, particularly around therapeutic robots like PARO, Denmark introduced social robots into elderly care homes. But that did not pan out successfully. “One of my colleagues, who is an ethnographer, went to check where are these robots and how are they doing, and she discovered that they were nowhere. She really didn’t find them. In the end, she discovered that they were put into cupboards and even into toilets, they were put away. It turned out that the carers were really worried about the robots. They saw them as very expensive technology and they didn’t want them to be destroyed or damaged. So they felt they had to take care of the robots,” Tanja explains.

What this shows is that, as with any technology, introducing robots doesn’t mean they will automatically be used. When not introduced thoughtfully, they don’t replace care work, they often create new forms of care work.

Share

The Netherlands presents a contrasting approach. Instead of large national rollouts, adoption is increasingly bottom-up and situational. Robots enter care homes through local initiatives, enthusiastic staff, or even as gifts from family members. This allows experimentation without pressure and lets care teams decide whether a robot actually fits their daily routines. As Tanja notes, usage depends heavily on workplace culture and continuity: when staff are curious and engaged, robots are integrated; when staff change and knowledge disappears, robots can become invisible again.

Why do we trust robots and AI?

2025 saw a rapid proliferation of AI-assisted therapists and chatbots, alongside growing media reports highlighting the harmful consequences of using ChatGPT as a substitute for professional mental health care. In several cases, these incidents were linked to suicides, prompting multiple wrongful death and negligence lawsuits against OpenAI. The question is: what is it about robots and AI that makes it easier for us to trust them with our deepest secrets and concerns?

Tanja sees trust as something we co-create with technologies. Robots invite projection. “We project a lot onto them… this phenomena of anthropomorphization is very common.” And while robots and chatbots aren’t truly empathic, the interaction can still feel emotionally meaningful: “they’re not empathic, but we feel empathized with when we talk to chatbots.”

That perceived safety can make disclosure easier. Tanja acknowledges a more uncomfortable implication: trust in machines can be a mirror, revealing gaps in human systems (time, attention, judgment, stigma) that make people prefer non-human listeners.

Is the future bright or dark?

Personally, I would love to have a robot that does chores instead of me. I would love to see a future where my parents can stay independent in old age, even with mobility challenges, because they have a robot that helps them live at home and call for help in emergencies.

The reality is that birth rates are declining and people are having children later in life. As a result, children are often still building their own independence when their parents already need care. While longevity and healthspan narratives are booming, societies are not getting healthier, and healthcare systems are not supporting ageing well at scale. I don’t believe relying on human labor alone will suffice.

As healthcare systems though leader Bogi Eliasen likes to emphasize: “Decision makers often think: how can we have the best healthcare system, instead of asking themselves: how can we create the healthiest population?” He believes dignity as a concept should become central in healthcare.

With AI becoming increasingly capable and sophisticated, the unanswered question is whether we’ll know how to manage AI once it turns into AGI. Will robots become terminators—or will we achieve bliss and radical abundance, as predicted and hoped for by Nobel laureate Demis Hassabis?

Tanja Ahlin bets on awareness, not panic: being conscious of “what we are doing with technologies and what the technologies are doing back to us.” She rejects both techno-utopianism and doom. Instead, she frames the future as a collective balancing act. Humans are incredibly flexible and creative, and we have survived many transformations before. We will find a way to live with technologies as well.

Leave a comment

Share

Let’s connect on Linkedin!

“I just saw… many things I can’t afford,” said an Australian attendee of WHX Tech conference, we ran into with Peter Birch, Host of Talking HealthTech Podcast while we were roaming around the enormous Dubai mall, wondering if there were any affordable shops in this huge complex full of posh, luxury brand boutiques where a simple T-shirt could easily be priced beyond 500 USD. Dubai, the 3,5 million people city of expats, is a dream place for those that already have, or those that want to make a lot of money. The air here is infused with business ambition and opportunities across different sectors.

Digital Health Landscape in the UAE

The UAE’s digital health market was valued at approximately USD 619.3 million in 2023 and is projected to grow at a compound annual growth rate (CAGR) of 23.3% from 2024 to 2030, according to Grand view Research.

Looking back at the digital health evolution, Mazin Gadir, regional expert in digital health strategy, Director with Alvarez & Marsal Healthcare and Life Sciences in Dubai describes two main eras:

From 2005 to 2014 developments were focused on EMRs and Health Information Systems, followed by attention to Health Information Exchanges (HIE) and Interoperability. Just as global luxury brands dominate storefronts in The Dubai Mall, leading Western health-tech vendors, such as Epic for electronic health records, are the preferred choice when adopting technology.

After 2020, consumer convenience, portals, and value-based care came to focus. The system now wrestles with burnout and the need to rationalize a “complex ecosystem” of data and tools.

Interestingly enough, while burnout among the healthcare workforce due to IT systems is a global challenge, self-reported burnout rates in the Middle East (UAE and beyond) are surprisingly low, according to KLAS research.

As presented by Coray Tate, VP of Clinician Experience and Interoperability at KLAS Enterprises, at HIMSS25 Europe (see his slide above) and explained in a Faces of digital health episode, collaboration-focused culture contributes to this good result in the region.

Share

Innovation and Investments

Reenita Das, Partner at Frost&Sulivan lived in several different countries in the course of her more than 20 years long career in healthcare. She describes Dubai’s health system as “one of the most sophisticated” she has used, with tech adoption near-universal. High digital literacy makes the environment a good test bed to solutions.

Innovation in UAE thrives due to deregulation and visionary leadership, with a clear differentiation between the two largest cities. While Dubai is celebrated as a proving ground for innovation, Abu Dhabi distinguishes itself with research depth and a biotech emphasis. As on any market, the success of newcomers depends on local customization, partnerships with regulators. Long-term sustainability will require stronger academia–industry collaboration.

Not everything is perfect: with its only 10 million people, the UAE is a small, high-cost market with multiple authorities and a mobile expat workforce. Equity isn’t solved either: high-end care is often insurance-dependent, and many lower-wage workers still travel home for treatment. “These are solvable problems, but only if we carry the same ambition we’ve shown on infrastructure into coverage design, payment reform, and cross-border standards,” says Reenita Das.

Subscribe now

Data Fragmentation is a Challenge in the Region

For all the wins, the region’s leaders are clear-eyed about what comes next.

When searching for care outside your own country in the region of Middle East, patients still hand-carry data - sometimes literally via WhatsApp, when they move between Gulf states. Interoperability across borders remains unfinished. Concrete use cases are pushing change: the Hajj season now anchors pilots to let core health data follow pilgrims, a first step toward an EHDS-style Gulf framework. The ask is simple: if Europeans can refill e-prescriptions abroad, outside their residential country, why not the GCC?

(These insights are based on the discussions with Dr. Mahmood Adil and Dr. Sam Shah - both advised various governments, healthcare regulators and public bodies on digital transformation strategies - Reenita Das, Partner at Frost&Sulivan, Dr. Sam Shah, clinician, advisor, and former startup founder, and Mazin Gadir, a regional expert in digital health strategy, Director with Alvarez & Marsal Healthcare and Life Sciences in Dubai. See the full interviews on Youtube.)

There are (at least) three concerns I have with AI in healthcare:

1. Over-reliance on AI insights by both clinicians and patients

2. Added pressure on patients to be informed, “empowered” and self-managing

3. Patients falling through the cracks due to over-automation, with no human intermediary to intervene when something goes wrong

Leave a comment

In this newsletter I will present these concerns through personal examples. I’m complimenting the lived experience with expert insights about ethical considerations with AI and current regulatory ideas, which were covered on the Faces of digital health podcast. Scroll down based on your interest!

Concern 1: Over-reliance on insights from both clinicians and patients

Two years ago, I experienced intense chest pain. After my GP ruled out cardiovascular causes, I was referred to a gastroenterologist since reflux other conditions cause heartburn (a burning sensation in the chest). The wait time for my appointment was two months and during that period, the pain persisted in waves. I was scared, to put it mildly. I turned to Google and ChatGPT to find clarity, and from what I read, gastritis seemed to best match my symptoms.

When I finally saw the specialist and he asked, “What brings you here today?” I blurted out, eager to be efficient and informed, “Gastritis!”
He looked at me with a raised eyebrow and replied, “Says who?”

That question hit me. It was a fair one. It made me realise how shaky my self-diagnosis actually was, despite the fact that I knew AI can hallucinate and you should never take its responses for granted.

I appreciated the doctor’s calm follow-up: that we’d need to run several tests before drawing any conclusions. We never did find a definitive cause. But it wasn’t gastritis and eventually, the pain faded away. Two years later, my skills for searching answers seemed to have improved, at least on the situation I’m about to describe.

Concern 2: Added pressure on patients to be informed and “empowered”

I’ve lived with inflammatory bowel disease for over 20 years, and right now I’m in the middle of one of the worst flare-ups I’ve ever had. Since first line of medications didn’t help, we had to significantly change my medications. When I spoke to my doctor, I knew I will likely need to get biologics (a class of medical products derived from living organisms or their components, that target specific pathways in the immune system or other biological processes) but I didn’t know much more than that.

When I finally got a specialist appointment, it left me stunned. Not because of what the doctor said, but because of what he didn’t say. He didn’t give a lecture. He didn’t tell me what he was going to prescribe. Instead, he looked me in the eye and asked, “So, which medications do you want?”

I was caught off guard.

Treatment protocols have changed significantly over the past two decades, and I’m not a doctor. However, before the appointment, I had done my homework. I’d had a long “conversation” with ChatGPT, feeding it several prompts that described my situation. This is how it went down.

How to prompt for patient research

I asked ChatGPT for guidance through the following prompts, constantly keeping in mind that for valuable responses you need to be as detailed as possible in prompting. This was easy, since who knows the patient’s history, bad experiences and problems better than the patient? These were my prompts:

1. “You are a gastroenterologist. This is your patient: [input patient details – current problems, current therapies, past therapies]. What escalation steps would you take?”

The result was a detailed list of diagnostic procedures, lab tests, and medication options. I followed up with another prompt:

2. “Create a timeline of suggested next steps, with expected duration for each potential therapy. Keep in mind the patient previously discontinued [therapy] due to [side effect], and stopped [therapy name] after 8 years of use. Provide instructions for the patient including diet, exercise, and other relevant recommendations. How do you choose between biologic therapies? What background can you give the patient, and what choices can you offer?”

That gave me an overview of available therapies, a comparison table, and loads of useful context. I asked one more thing:

3. “Based on the discussion above, create patient instructions.”

This turned details into an easy to digest lay language.

This 15-minute AI “consultation” gave me a solid understanding of where treatments currently stand. I also did some fact checking with Evidence Hunt - a reliable source of medical information meant for clinicians.

Regardless of this research, I kept in mind my chest pain experience, reminding myself that I’m not a doctor and shouldn’t feel too confident about the results I got. But I felt prepared for the doctor’s appointment, ready to ask follow-up questions, and fact check my research with the specialist.

So back to the doctor’s office and his question, which medications I want.

I cautiously suggested the therapy I liked most based on my AI research. The doctor briefly walked me through other options, but confirmed that my suggestion was worth giving a try.

I left the appointment feeling hopeful about the new therapy, proud about my research that seemed to have been solid this time around, but also wondering: Is this what medical appointments are now? Do patients show up with pre-conceived ideas based on AI research? Are doctors expecting us to do that level of prep in advance?

Subscribe now

Patients Need Project Management and Digital Skills

Years ago, Graham Prestwich, Patient and Public Engagement Lead at Health Innovation Yorkshire & Humber, wrote a worthwhile read about “patients being project managers.”
This means that if you’re a good project manager, your chances are far better for good outcomes then if you’re not. If we expect patients to be a more important player in improved outcomes, we also need to equip them with the right tools - like project management skills, and in the age of digital health, strong digital and AI literacy.

I worry that patient outcomes increasingly depend on how engaged, informed, and assertive people are in advocating for themselves—right down to reminding clinicians about tests or procedures that may have been overlooked.

I am eternally grateful for the clinicians that take care of me, but I also see how increasingly overburdened the healthcare system is. I’m glad I’m still communicating with the human GP, not an automated system. Here’s why.

Concern 3: Patients falling through the cracks due to over-automation, with no human to intervene

Already back in 2021, WIRED reported the story of a woman who was denied pain medication for severe endometriosis-related chronic pain. She later discovered that an algorithm, which was part of a system designed to combat opioid abuse, had flagged her as a “drug shopper”: someone trying to obtain multiple prescriptions for controlled substances from different sources.

However, the reason for the apparent overuse wasn’t what the algorithm assumed. Her dogs had been prescribed opioids, benzodiazepines, and even barbiturates by their veterinarians. And prescriptions for animals are typically issued under the owner’s name.

The algorithm didn’t know that. It only saw multiple prescriptions in her name—and nobody intervened to double-check.

These examples point to a growing tension: as AI tools become more integrated into healthcare, we need to make sure they’re supporting—not replacing—human judgment, empathy, and care. Empowerment should never become a burden. And automation should never mean abandonment.

Subscribe now

Share

The Ethical Perspective on AI in healthcare

Dr Jessica Morley is Associate Research Scientist at Yale University Digital Ethics Center. At HIMSS25 Europe she had a short lecture titled: When Data Turns Dangerous: The Risks of Bias and Misuse in Healthcare AI. Two of her concepts stuck with me:

• AI is not enabling personalized medicine, instead what’s actually happening resembles targeted advertising, where people are clustered into groups based on their characteristics.

• Bias is not just a data imbalance, and can’t be solved only by addressing the diversity of datasets. Even with more data, systemic issues like racism, mistrust, and underdiagnosis will still show up in how data is recorded and interpreted.

I asked Jessica for a longer discussion to unpack these issues.

1. Why AI resembles targeted advertising?

Instead of treating people as individuals, AI systems often cluster patients into demographic or behavioral categories resembling how advertisers target consumers based on attributes rather than personhood. Because AI lacks semantic comprehension and cannot understand complex, contextual expressions like “I don’t feel like myself” or “I’m in pain” it makes it poorly suited to genuine personalization.

The open question here is: We know AI isn’t perfect. If we accept the thesis that it resembles targeted advertising and cluster people based on available data, is this bad? Is it making care worse?

Dr. Morley believes it might. People with fewer digital records (e.g., marginalized populations) are poorly represented in the datasets used to train AI systems. Therefore, AI doesn’t “treat” them but it treats a partial data shadow that may not represent them accurately at all. The clinical problem therefor is that over-reliance on this model could lead to misdiagnosis, underdiagnosis, or inappropriate care.

If you want to read about concrete example of AI causing harm, head to STAT News. In the last few years they extensively covered how the US largest health insurance company pressured its medical staff to cut off payments for seriously ill patients in lockstep with a computer algorithm’s calculations, denying rehabilitation care for older and disabled Americans, or how a study had found that Epic’s artificial intelligence tool to predict sepsis, a deadly complication of infection, was prone to missing cases and flooding clinicians with false alarms. Among the latest research, Declan Grabb, a forensic psychiatry fellow at Stanford and Max Lamparth, a postdoctoral fellow at the Stanford Center for AI Safety and the Center for International Security and Cooperation shared research about the use of chatbots for mental health and found that all but one language model was unable to reliably detect and respond to users in mental health emergencies.

2. Let’s talk about bias

It’s easy to think that by getting more diverse data sets, AI accuracy problems will be solved. Here, Dr. Morley warns of two things, we don’t often think about. In healthcare, “bias” isn’t always a bad word. Accounting for biological differences across sexes or ethnicity may be essential for effective treatment. The goal is transparent, justifiable bias, not neutrality. By focusing solemnly on creating a "perfect" data representation of everyone risks surveillance creep, meaning more data collection, more invasive monitoring, all in the name of fairness.
Bias should be seen as a diagnostic tool for identifying where systems fail patients, and be managed with warning labels. Which brings us to regulation.

Current ideas for global regulation for AI in healthcare

Up until recently, Coalition for Health AI (CHAI) seemed to be leading the way in regulating AI, through a network of quality assurance labs to evaluate AI models in healthcare. CHAI wanted AI innovators to provide some sort of AI food labels that would include information about AI model, data sources and more. However, end of May 2025, CHAI pivoted from assurance labs to “training assurance resource providers”, as reported by STAT news. This in essence means, that instead of governing assurance labs that provide credibility and accountability for assessing AI solutions, CHAI shifted towards empowering organisations to assess solutions themselves.

In the US a lot has changed with the current administration, with the focus on speed of innovation and deregulation of AI, not just in healthcare. The FDA is becoming both a power‑user of generative AI and a lighter‑touch regulator of external AI products. This is how some regulations compare based on Dr. Morley’s research:

Can we build a global regulatory network with an early warning system?

Beyond the US, HealthAI, a global non-profit that champions responsible AI in Health, is trying to build a global collaborative system for AI regulation. The agency works with governments to help them set up appropriate governance structures to enable the adoption of AI for good.

“ One question, I ask myself many times is how ethical is it that a technology that can improve outcomes is not being put in the market because we haven't been capable to put in the right governance in place, and that's happening today,“ says Health AI CEO Ricardo Baptista Leite, a Portuguese medical doctor, manager, author, university professor, analyst and politician. The key challenge in his eyes is the financing equation. If the payers, the governments and the insurance companies are not going to chip in the payment models, scalability will be an issue.

Health AI has several ideas for managing AI globally, with heavy reliance on governments. Rather than relying on external actors, Ricardo Baptista Leite advocates for investment in sovereign capacity, which includes building local regulatory and technical expertise to assess, validate, and monitor AI tools. He also emphasizes the importance of collaborative governance, encouraging countries to participate in global initiatives like Health AI’s regulatory network and community of practice, where shared knowledge and real-time incident data can strengthen oversight.

To make adoption viable, governments must learn to evaluate AI’s clinical and economic impact and establish evidence-based reimbursement pathways. Crucially, Ricardo insists that regulation must be adapted to each country's unique context—avoiding rigid, imported frameworks that ignore cultural and structural realities.

Health AI’s Vision for Governance

Health AI envisions a globally coordinated but locally adaptable model of AI governance in healthcare, centered on trust, transparency, and scalability.

Key elements include:

1. Global Regulatory Network: A coalition of 10 "pioneer countries" (to expand further) working together to build and share AI regulatory capacity, tools, and best practices.

2. Maturity Model for AI Oversight: A framework—based on WHO models—to help countries assess and improve their readiness to govern AI, from incident tracking to market approval.

3. Early Warning System: A platform to share post-market incident reports across countries in real time, improving safety and transparency across borders.

4. Global AI Directory: A centralized listing of AI tools approved by country regulators, helping governments, innovators, and payers understand what technologies are validated and where.

5. Health AI Navigator (coming soon): A digital tool tailored to different audiences (innovators, regulators, policymakers) offering implementation-oriented guidance on AI governance, reimbursement, and data management.

6. Flexible Technical Assistance: Health AI works as a partner, not a policymaker, helping governments co-develop governance models without imposing rigid structures.

 “One thing that in a way breaks my heart sometimes is seeing some countries saying to us, listen, we're not ready yet to address this challenge. Those are the countries that need to step up first, because if they're too far behind, the longer they wait, the harder it is to keep up and the technology is moving in the speed that while it can be either the greatest equalizer of our time, elevating people out of poverty towards better health, or it can be the greatest divider leaving an important part of the population behind,” Ricardo concludes.

So how harmful is AI in healthcare?

My current patient journey has reminded me that while we often speak about patient-centricity and interdisciplinary care as the north stars in healthcare, we’re still far from seeing those principles fully realized in practice.

I remember a doctor colleague from my time in healthcare IT once saying, “Doctors are trained to make fast decisions.” But as healthcare systems become increasingly strained, and clinicians are pushed to move even faster, more responsibility, more project management, more cognitive load is shifting onto patients. Especially if you want the best care.

AI is often presented as the solution to these pressures. In fact, Anthony Chang, a leading authority in the field, argues that it will soon be unethical not to use AI in healthcare. In his view, the biggest challenge with AI in healthcare is that “mature AI tools aren't being used efficiently or effectively; tools that could reduce caregiver burden and improve outcomes.”

Personally, I see tremendous potential in AI particularly in research and system-level optimization. But if we want AI to be an equalizer rather than a divider, as Ricardo Baptista rightly cautions, we have serious work ahead.

We need to design systems that don’t just assume every patient is digitally literate, medically informed, and self-confident enough to challenge their clinician. We need AI that supports, not replaces, human connection. Otherwise, we risk creating a healthcare future that’s efficient on the surface but deeply inequitable underneath.

Where do you stand on the topic? Leave a comment and share your perspective.

Leave a comment

Share

Share Faces of Digital Health Newsletter

Share

Europe is building the European Health Data Space, which presents a vision for a connected healthcare across Europe, interoperability and healthcare data exchange among different EU countries, and with this digitalization, new opportunities for data research through secondary use. For years Europe has been looking at Finland as the exemplar of secondary use of health and social data. There, since 2019, a dedicated authority Findata has been granting permits for the secondary use of social and health care data. Fees apply, depending on where the interest is coming from.

Germany is now following suit, granting researchers access to statutory health-insurance claims data it has amassed over the past decade. Because 90 % of residents are covered by public insurers, the dataset offers an almost comprehensive view of the nation’s healthcare billing.

Understanding healthcare Data Challenges in Germany

Subscribe now

Germany has been observed with mixed feelings within the digital health space: on the one hand, it deserves admiration for its efforts in digital therapeutics reimbursement and normalization of prescribing digital treatment solutions (read more about DTx in Europe in the dedicated newsletter). On the other side, there’s been no lack of frustration with the slow introduction of national electronic healthcare records, ePrescriptions and more. The challenges accompanying delays compared to many other European countries are multifold:

• lack of national digital infrastructure, that had to be build,

• size of the country and its fragmentation to 16 federal states, each with its own rules and desires,

• while Germany initially created a system where citizens had to opt-in to get their data in electronic health records, the system was later turned into opt-out, where citizen have to actively exit the the system. Many other countries in the past, such as Australia with its My Health Record, learned the hard way that an opt-out approach is the only way for large scale adoption of electronic healthcare records.

Hence, the prospect of opening claims data for 74 million people marks is an especially momentous achievement. “Starting next year we will receive pseudonymised data from the electronic health records as well and that data will come to us already linked to claims data,” says Steffen Hess, the head of Health Data Lab, which is in charge of granting access. Side note: quite some data still needs to be structured in an interoperable format hence not all data will be added to this data pool immediately.

Who, When, How?

The application process will begin in the of summer 2025. Interested parties should follow Health Data Lab’s website to see when the digital door will open. Below are the basics of what you can expect.

1. What Data is Available?

• Claims Data: Main source, covering 95 statutory health insurance providers—representing 90% of the German population.

• Electronic Health Records (EHRs): Pseudonymized EHR data will also become available and linked to claims data as the next step.

• Scope: Broad national coverage (approx. 74 million people), ideal for research into rare diseases, health system improvements, epidemiology, etc.

2. Who Can Apply for Access?

• Anyone operating under GDPR-compliant jurisdiction.

• Applicants must propose projects with approved public interest purposes (e.g., medical research, health system improvement).

• Commercial interests like market access research or harmful product development are explicitly excluded.

3. How Does Access Work?

• Applications expected to open Summer 2025 via the Health Data Lab.

• Data access is granted through a secure virtual processing environment—data cannot be downloaded.

• Results can be exported only after clearance to ensure no personal data is exposed.

• Expected approval time: < 3 months.

• Access fees apply, based on use and duration (similar to models like Finland’s Findata).

Share

Tune in to the discussion about the new system for secondary use of data in Germany with Steffen Hess - Head of Health Data Lab.

Secondary Use of Data: The Basics

When talking about secondary use of data its worth starting with a clarification, which data we are referring to. We often speak about secondary use of data in a general sense. However, as classified by Fidelia Cascini, Digital Healh expert engaged by the Italian Ministry of Health on European and national initiatives related to data use, who recently wrote a book on secondary use of data, healthcare data categories include:

• Real-World Data (RWD): Derived from EHRs, registries, and wearables, used for observational research and regulatory support.

• Administrative Data: Billing and claims data used in healthcare delivery analysis.

• Biobanking Data: Genetic and biospecimen data linked with personal health data for research.

• Clinical Trial Data: Structured data for regulatory approval and hypothesis testing.

• Electronic Health Records (EHRs): Central to modern health informatics, used for research, quality, and public health.

• Genomic and Proteomic Data: Used in precision medicine and integrated research.

• National Registries: Provide longitudinal health insights across populations.

• Patient-Generated Health Data: Self-tracked data from apps and devices.

• Social and Behavioural Determinants of Health (SBDH): Contextual data influencing health outcomes

As mentioned in the case of Germany, researchers will have access to billing data first, later in the process to electronic EHR data.

If you’re interested in detailed understanding of secondary use of healthcare data, Fidelia Cascini book is open-access publication, that provides an overview of approaches and challenges in Europe.

One thing is clear: the European Health Data Space, a legal framework that took effect in 2025, has raised expectations for future advances in research and the use of healthcare data across Europe. Only anonymized data will be available for secondary use and individuals will be able to limit or restrict certain data categories. “The barriers are time and limited resources but in the next eight to ten years the picture will be completely different compared to today. Larger countries will reach good results in interoperability and allow the secondary use of data,” believes Fidelia Cascini.

Subscribe now

Learn more at HIMSS25 Europe!

Fidelia Cascini, Steffen Hess and many other policy makers and health data authorities representative will be present at HIMSS Europe 2025 Conference and Exhibition, taking place in Paris from 10-12 June 2025.

Secondary use of data and EHDS will be addressed through several discussions. By coming to the conference, you can also obtain a digital health literacy certificate! Read more on the website.

What else can you expect at HIMSS?

Participants will be able to learn about AI, women’s health, longevity, European legislation, bridging policy and practice, the future of workforce and more.

You can still join the conference remotely through the live stream: https://www.nextmed.health/live-2025. See the agenda of the upcoming days: https://www.nextmed.health/program 

The most gripping presentation on Day 1 of the NextMed Health conference 2025 did not come from an ambitious health-tech innovator, business analyst or a doctor, but from a coder, filmmaker and AI developer Steven Brown. A month after undergoing all possible medical tests, from blood draws to colonoscopy, and doctors guessing if the cause of his unexplained symptoms was just stress or gas, a sudden stab of abdominal pain during dinner sent Steven to the ER, where a new set of tests led to a diagnosis of a rare form of multiple myeloma. That prompted Steven, who comes from a family of physicians, to explore AI’s potential in diagnosing and managing diseases. He created MedicalAdvocate.ai - a cutting-edge platform for AI-led medical consultation.

After uploading medical results on the platform, multiple AI agents act like a tumor board with different specialties (oncologists, radiologists, etc.). Each AI agent provides insights from their specialty, contributing to a comprehensive review. Then, a higher-level AI agent ("Hippocrates") consolidates expert opinions and offers a holistic recommendation.

This is not a first story of a patient solving his medical case with the help of AI. While AI is not a standard of care yet, it has proven to be instrumental in some edge cases and hasn’t ceased to amaze with its life-saving potential.

What does this mean?

• AI isn’t replacing doctors but can enhance patient understanding and decision-making.

• AI agents can help bridge the gap between patients and specialists who have limited time.

• AI enables personalized medicine by analyzing patient data more thoroughly than human doctors can in routine consultations.

MedicalAdvocate.ai is at the moment only a personal project and a closed platform.

Share

Subscribe now

The move towards agentic AI

In 2023 generative AI was all the rage. 2 years later, we are moving way beyond that with agentic AI, AI agents. What are they?

• Generative AI (e.g., ChatGPT) functions like a coach on the sideline—it provides advice, generates content, and suggests actions but does not take direct action itself.

• AI Agents perform specific functions within an AI system but do not necessarily operate independently. For example, AI agents act like a virtual research team: data analysts, statisticians, literature reviewers, scientific writers. In Steven’s case, AI agents represent doctors from different specialities that each comment on medical results from their own perspective.

• Agentic AI functions like a skilled project manager - not just making strategic decisions, but delegating tasks, tracking progress, and adapting in real time to meet its goals.

Data remains a key challenge for AI development

While increasingly more tangible in it’s promises, AI is still facing many challenges because the quality of AI depends on data it’s fed. At the moment 30-40% of data in electronic health records are inaccurate. Most published studies can’t be replicated or scrutinized because datasets used for those studies are not available. Bayo Curry-Winchell, Founder of Beyond Clinical Walls and Urgent Care Medical Director at Saint Mary’s Health Network, reminded us that simply increasing access to healthcare will not eliminate bias. Instead, we must fundamentally transform how bias is recognized and addressed within the system. Despite her leadership role, Curry-Winchell experienced pain dismissal while in labor at the very hospital where she works — a failure in care that nearly cost her life. Her call to action to developers and users of AI is simple:

1. Check your bias.

2. Consider diverse patient groups – involve them in their process, do not focus on the narrow population

3. Champion equity.

Why AI is the only hope for healthcare

AI adoption and development are advancing at a rapid pace, and it’s doesn’t seem crazy to predict that AI will in many cases become mainstream in well before 17 years, which is the usual timeline for new innovations to become a standard practice in healthcare. 1017 devices with AI have been approved by the FDA to date, most of them in radiology.

With current trends in healthcare sustainability, aging and workforce shortages, we don’t really have a choice but to try to get as much as possible from AI to optimize costs and care delivery. WHO predicts a 10 million shortage of clinicians by 2030. As shared by Anthony Chang, Founder of AIMed and Chief Intelligence & Innovation Office, Children's Hospital of Orange County, U.S. health expenditure is 4.2 triillion USD per year. Because 25% of the cost is administration, AI can lead to cost savings of 150-250 billion per year in United Stated. With 90% of time spent on technology and 10% of time in clinical practice on interactions with the patient, AI has the chance to rehumanize medicine.

At some point it’s not going to be ethical to NOT use AI in healthcare, said Chang in a 3-hour workshop. At the same time, Chang reminded us that “change doesn’t happen at the pace of technology — it happens at the speed of people.”

Subscribe now

Share

“My bigger concern is that mature AI tools aren't being used efficiently or effectively”

Here is a short interview with Anthony Chang, Founder of AIMed and Chief Intelligence & Innovation Office at Children's Hospital of Orange County, after the first NextMed Health presentation and his 3-hour workshop on AI in healthcare.

Mr. Chang, you just had a really engaging session on AI in healthcare. One of the strongest points you made was that it may soon be unethical not to use AI in healthcare. Can you expand on that?

Mr. Chang: Of course. That statement needs to be interpreted carefully. I mean in specific situations where AI has already been proven effective. We're reaching a point where the standard of care will require the timely use of AI to make accurate diagnoses.

Radiology is probably the prime example.

Particularly brain and cardiac scans. We're already seeing some accountability emerging when hospitals don't use available AI tools in these areas.

You also mentioned that when you consult hospitals on their AI approach, it's not uncommon to find that an actual strategy doesn’t exist. What’s your advice to hospital leaders in that situation? If they already have a data strategy and governance framework, what’s typically missing—and where should they begin

Often, when we’re asked to advise on an AI strategy, we find the hospital doesn’t even have a solid data strategy or data governance in place. That’s the foundation. Once that’s established, you can build an AI governance framework and create a dedicated committee.

As I mentioned during the session, the AI agenda in healthcare must be driven by human-to-human interactions and relationships. Change isn't technology-driven—it’s human-driven. And that’s one of the beautiful things about AI in healthcare. The technology is robust and mature, but it takes people to learn it, adopt it, and use it effectively.

Where do you currently see the biggest dangers with AI? One participant mentioned concerns about non-diverse datasets, which may not improve in the near future. Her call to action was for people to collect more data. How do you view that issue, and what are the broader challenges AI will face before it becomes more widely adopted?

AI is going to hold us accountable for ensuring balanced datasets. There are technological solutions, like the use of synthetic data, that can help address this. But my bigger concern is that mature AI tools aren't being used efficiently or effectively—tools that could reduce caregiver burden and improve outcomes.

Why do you think these tools aren’t being used effectively? You mentioned earlier that of the $4.2 trillion spent annually on U.S. healthcare, around 25% could be saved through AI-driven automation of administrative tasks.

One major reason is a lack of education in this area—which we’re trying to address through sessions like this. Another reason is misaligned financial incentives among payers, providers, hospitals, and patients. Even the most elegant AI solution can’t overcome those human-driven issues unless incentives are aligned.

Finally, is there anything you think people should understand right now about agentic AI, which came up at the end of the workshop? You mentioned it’s often confused with AI agents.
Yes, it’s an easy distinction to miss. I see agentic AI as a capable team captain on the field. It has autonomy, but still operates under the guidance of a coach—which, in this analogy, is generative AI. The individual players could be seen as AI agents, each acting autonomously, but within the broader autonomy of the entire system.

Subscribe now

Let’s connect on Linkedin!

Ideal healthcare data universe is clear: having the right data at the right time, regardless of where patients seek care. This involves seamless data flow with robust privacy protections, ensuring sensitive information remains secure. The European Health Data Space (EHDS) aims to achieve this through several projects in the past and present exploring ways to unify diverse data standards and healthcare IT systems. The ultimate goals: optimizing healthcare and contributing to its sustainability.

Subscribe now

Source: EHDS Factsheet.

Why is this so hard to achieve in real life?

Fragmentation in European Healthcare

Europe’s healthcare landscape is a patchwork of 27 countries, each with a different language, financing, and reimbursement schemes. Larger countries are further divided into autonomous regions, complicating healthcare IT architecture. Patients often face confusion when seeking care outside their regions due to inconsistent digital systems.

For example, Italy has 20 regions with varying levels of digitalization. Northern regions like Lombardy are advanced, while southern areas lag behind. As Marco Foracchia, CIO of Local Health Unit in the Emilia Region, IRCCS, describes, patients struggle to make appointments as each region uses its own digital tools.

Healthcare providers may be digitized, but they are not always connected. Historically, software procurement has been fragmented, with institutions independently deciding on solutions without broader regulatory mandates for standardization. When buying software to date, healthcare institutions have often seemed as of 3 year olds who say “I will decide for myself” and potentially “I don’t to have what Mary has".

Up until recently it was legally forbidden to share data in the Netherlands. With a new law two years ago, the direction has shifted and is aligned with the current broader mindset across the world, to create ecosystems where data follows patients.

Decisions in the past created legacy systems, which left patient data fragmented among several patient portals for different services patients are using. Slowly however, we are seeing consolidation progress on regional levels.

Share

Case Study: OneLondon

Though the UK is no longer in the EU, its healthcare data consolidation efforts offer valuable lessons. In London (population: 8.8 million), 24 different EHR systems exist. The OneLondon project, launched in 2018, aimed to connect 35 NHS Foundation Trusts across five Integrated Care Systems.

Since its inception in 2018, OneLondon has evolved significantly, with increased data-sharing capabilities that now extend beyond London. This interconnected system allows various healthcare providers to access critical patient data at the point of care, facilitating better-informed treatment decisions.

The bigger picture is care that’s integrated not just between healthcare institution but social care as well. Many police interventions are related to mental health issues, and police is not equipped to dela with them properly.

Progress happens one segment at a time

Improving interoperability in London is incremental. To support end-of-life patients, OneLondon centralized urgent care plans on a platform accessible to all EHRs. Additional care plans are being introduced based on these learnings.

Queen’s University Belfast did an independent economic evaluation of the London Care Record, and found that based on 27 million uses of the London Care Record (up to March 2023), the value of time saved is estimated to £2.1 million per month, with 20,000 hours per month saved.

Public trust is everything

Interoperability projects are very technical and hard to translate into lay language. But in the end it’s the human factor that is essential for success. Luke Readman, the CIO of OneLondon says building public trust has been central to OneLondon’s success. The project engaged in annual public consultations to understand and align with the public’s expectations for data usage and privacy. This inclusive approach helped strengthen public support and confidence in the system.

Similarly, before France designed their digital health ambitions and investments, they did a thorough on the ground research to see the differences between regions and populations and also understand what people expect from healthcare digitalization.

Share

The Small Countries Problem

To support EHDS, large countries first need to set national strategies to harmonize regional data with EHDS requirements. On the opposite end, smaller countries have a different problem: lack of choice.

Healthcare IT solutions aren’t like consumer products you can order from wherever to wherever. Healthcare IT vendors may deprioritize smaller markets due to the high cost of localization and low scalability. This lack of competition can result in outdated systems, as vendors have little incentive to modernize their offerings.

Leave a comment

Subscribe now

End of data standards wars?

On the up side it seems that collaboration is coming in the front row of data standards debates and data standards bodies. The Annual International openEHR conference in November 2024 showcased leaders from Karolinska University Hospital, Catalonia region, University Hospital Basel, and vendors such as AWS advocated for combining openEHR (for clinically granular data storage) with FHIR (for messaging).

The focus has shifted from debating which standard is superior to exploring how standards can work together.

Starting at the source

Regional and national perspective are a few levels above the source of data capture challenges which start in healthcare institutions. In the era of cyberattacks being a weekly occurrence, AI, medical devices and digital health rapidly advancing, it may seem easier and cheaper for healthcare institutions to stick to paper. In many cases in Europe, they still do. Software implementations take years of time and millions of euros.

Several challenge are present in this process for vendors and providers:

1. Providers may have unrealistic expectations or lack clarity on their needs.

2. Vendor claims in tenders are difficult to validate.

3. To avoid incorrect purchases, healthcare providers are embedding clauses in contracts that stipulate payment is made only after successful installation. This puts vendors in a tough spot of planning their revenue, especially since delays are common in implementations.

Helping hand in digital transformation

The EU’s Resilience and Reconstruction Program (PNRR) has driven funding and deadlines for nationwide EHR implementations. Ministries had to design strategies, designate funding, write tenders. In 2026, we will see the successes, failures and EU funding that will need to be returned if implementations are not in place.

One helping hand hospitals and countries can use when considering where to even begin with healthcare digitalization and what to invest in first, are HIMSS maturity models. These are self-assessment questionnaires, that enable hospitals to benchmark their current capabilities and outline their roadmap for digital transformation, aligned with global standards. With this they can enhance decision-making for procurement and strategy, and avoid installing the wrong solution that will be hard to replace. Before Germany divided 4 billion euros to hospitals, they assembled a DigitalRadar project. Among other things, they assessed the digital maturity of hospitals with 65% of the evaluation based on HIMSS EMRAM criteria.

HIMSS offers six maturity models, each focusing on different aspects like EMR adoption, infrastructure, and community care. Each model has seven stages helping hospitals understand where they are. Recognition of the highest levels of digitization (e.g., Stage 6 or 7) involves on-site assessments by HIMSS experts.

At HIMSS Europe a dedicated track called Digital Maturity Roadmaps will give hospital an national representatives an opportunity to meet and learn from hospitals across EMEA and beyond about their digital transformation journeys.

In 20 years ….

The EHDS framework for cross-border data exchange and secondary data use has spurred necessary changes across Europe. Even in smaller countries vendors will need to upgrade their solutions to meet European expectations.

We’re still far away from the idea of the perfect healthcare data universe described in the beginning. Progress being made will be measurable in the next 10 to 20 years.

Leave a comment

Share Faces of Digital Health Newsletter

Share

Let’s connect on Linkedin!

Subscribe now

If 2023 was full of excitement related to generative AI, 2024 is sliding towards maturity, partnerships and practical applications of AI. Rock Health tracked 3.3 billion USD in investments to healthcare AI from January to October 2024 in the U.S., representing 36% of all funding in digital health.

Incorporating AI into every solution has become a necessity to capture investor interest. Which means that now, solution will require innovative elements to truly stand out, and have AI just to appear up to date.

Expectations for AI's benefits are high, but we're still in the early stages of realizing its full potential. The 2024 AI Adoption in Healthcare Report, prepared by HIMSS and Medscape, reveals that while clinicians are increasingly leveraging AI to enhance efficiency, it has yet to reduce their workload or lower costs.

In old news: clinicians are still concerned about privacy and ethical dilemmas in using AI.

Subscribe now

Share

For patients and providers to feel safe about AI, it needs to be regulated and wear a stamp of approval. The question is, who can do the approval with credibility?

Several organisations popped up for AI regulation in the last few years:

• CHAI - Coalition for health AI - is an industry membership organisation which is setting up assurance labs across the USA to validate AI solutions,

• TRAIN - Trustworthy & Responsible AI Network is an intiative giving guidance on the practical steps to responsibly operationalize AI,

• HealthAI: The Global Agency for Responsible AI in Health - is a Geneva-based non-profit organisation establishing global standardization to advance the development and adoption of Responsible AI solutions in health,

• and many more.

Share

Leave a comment

In Europe, the EU AI Act defines 4 levels of risk for AI systems, and different requirements based on the level.

Assurance Labs for Healthcare AI in the U.S.

The further along with efforts to validate AI solution is CHAI - The Coalition for Health AI, which has over 1,300 member organizations, including tech giants like Microsoft, Google, and Amazon. CHAI is establishing a network of quality assurance labs to evaluate AI models in healthcare. According to the CEO Brian Andersson, two assurance labs will be open by the end of 2024.

Leave a comment

Subscribe now

These labs aim to provide independent, transparent assessments of AI models' performance across different populations. To get the details, I interview Andersson and asked who will pay, how will model cards work and more.

Model Cards and Evaluation

With the increasing complexity of new tech, we need approaches to simplify quality recognition. A very common idea is to have food nutrition type labels for technology. The European project Label2Enable promotes a quality and reliability label for health apps. This is a European project, and not a standard of practice in European countries just yet. ORCHA (Organisation for the Review of Care and Health Apps) is working globally to provide digital health apps assessments and offer clinicians clarity on what to recommend, and what not. It’s therefore no surprise that CHAI also wants to introduce labels for AI models.

AI model card presented by Brian Andersson at Digital Health and AI Innovation Summit in Boston 2024.

Model cards are filled out by vendors and need to describe their training data, methodology, indications, and limitations. Assurance labs provide independent evaluation reports.

Assurance labs aim to balance innovation with safety in AI development. They can help identify model performance issues across different populations and accelerate improvements.

Open questions with AI assurance labs

The introduction of AI assurance labs is very welcome and intended to build trust in AI among healthcare providers and patients. However, several question remain open: will healthcare providers trust the assessments, given that CHAI is industry led? Looking at the other side of the coin - the industry knows healthtech development best and is best positioned to assess technology, especially in the challenging workforce shortages in healthcare IT.

The second question is: how expensive will assessments be and will this cost be an additional barrier to success for early-stage innovators? According to Anderson, CHAI is creating a competitive marketplace of quality assurance labs to keep costs reasonable. “The cost of evaluations is expected to be in the range of thousands of dollars, not millions,” he said.

Share

Leave a comment

Future expectations

Evaluation reports will be published in a public registry for transparency.

While vendors won't need to disclose the metadata their models are built on, the added transparency regarding AI model quality is a significant step forward. It will uncover new challenges, providing developers with valuable insights and a clearer direction for future advancements.

CHAI is also exploring partnerships with health systems and NGOs to establish quality assurance labs in the EU. The initiative aims to be scalable and adaptable to different geographic regions and populations.

Share

Leave a comment

In other episodes …

This is what was covered on Faces of digital health in the last two months.

Podcast website

Spotify

How Can Hospitals Advance Their Digital Transformation With HIMSS Maturity Models? (John Rayner, Senior Director Analytics - EMEA at HIMSS)

The future of data in in collaboration among standards bodies (Rachel Dunscombe)

How does remote patient monitoring look like in South Africa and Nigeria?

Zayna Khayat: Language, Mindset and the Future of Care

Don’t miss!

Holiday Edition: Future Trends in Digital Health & AI (Daniel Kraft, Lucien Engelen, Zayna Khayat)

In this special holiday edition of Faces of Digital Health, healthcare futurist and strategist Zayna Khayat, futurist and founder of NextMed Health Daniel Kraft and digital health expert, CEO of Transform.health Lucien Engelen discuss key trends in digital health and AI. The panel also touches on the role of new players in the health space like supermarkets and tech giants, the importance of prevention in healthcare, and the promising technologies poised to revolutionize the industry.

Mark your calendar on Linkedin

Or click “Notify me” on Youtube.

Let’s connect on Linkedin!

Let’s start with a glossary:

• Vendor clutter - When vendors sell solutions they have, or think the market needs, instead of solving real problems.

• Pilotitis - Doing pilot projects that don’t translate to everyday practice.

• Patient odyssey - Many patients don’t experience a typical patient journey; instead, they go through an odyssey of navigating complex care systems.

  Share

As a chronic patient, I feel immense gratitude for the innovations in drug development. I’m alive thanks to Pharma and it is amazing to see how far we’ve come in gene therapies, biologics, drug repurposing to address more than one need, or “simple” things such as oral laxative formulas that you need to take before your colonoscopy (trust me, there's a big difference between having to drink several litres of what tasted like salt water the day before the procedure 20 years ago, and drinking a relatively palatable solution the evening before colonoscopy in 2024 …). I’m cautiously optimistic about the rise of platforms for in-silico (AI-driven) drug discovery.

But under this promising progress, Pharma struggles to fully embrace digitalization. As Craig Lipset, a former Pharma executive and leader in clinical research and medicine development, mentioned at DHAI, even the simplest digital solutions, such as eConsent (digitized consent forms for clinical trials), were only used in 49% of clinical trials in 2023 (see his slides with sources below).

Despite regulatory guidance and patient preferences for eConsent, its use has declined since the pandemic. Digital data enables better compliance monitoring, yet Pharma isn't fully capitalizing on it.

Share

Subscribe now

Craig Lipset believes regulation is often used as an excuse for delaying innovation. Many digital solutions are already feasible within regulatory frameworks—what’s missing is bold leadership.

A company showing just that, might be Otsuka Pharmaceuticals. Otsuka is notable for two reasons: first, it invested in Proteus, a digital health unicorn that developed a sensor to track adherence to schizophrenia medications. In 2017, this led to the first FDA-approved digital pill, which sparked debate about its long-term success, particularly for this indication.

To surprise of many, in early 2020, Otsuka pulled the plug on financing Proteus, forcing the company to file for bankruptcy. Now, in 2024, Otsuka has launched a digital therapeutic (DTx) for major depressive disorder. According to the company, the DTx comes with a low co-payment to enhance accessibility. Otsuka acknowledges that this therapy will initially incur financial losses, but we will have to wait to see how successful the support for this therapy turns out to be.

Moreover, Otsuka made another move worth keeping an eye on: the company partnered with Amalgam. In October 2024 they released an app for caregivers of patients with Alzheimer’s disease. This is a free app available in the US market, but more importantly, it doesn’t address patients directly, but caregivers - a market niche that is continuously overlooked in healthcare, even though we know that ‘when a patient gets sick, the whole family gets sick’.

How are wearables used for research and in clinical trials?

Pharma is utilizing digital tools in several ways to measure clinical endpoints. For example, wearables that continuously track gait can detect motor function decline faster than standard in-office tests, such as the 10-meter walk test (a widely used measure of gait speed in Parkinson’s disease). Real-life mobility data from smartphone apps can also identify fall risks, which could serve as an early warning for fall risk prevention measures.

Subscribe now

Real-world data (RWD) is often seen as the true test of clinical innovation, as it reflects less-controlled, everyday conditions. Health monitoring sensors embedded in rings, jewelry, watches, bands, smartphones, and IoT devices collect RWD passively, making it a potential goldmine for future discoveries. However, as one speaker at DHAI pointed out, RWD in research often isn’t used for it was gathered for. When patients monitor their health they might be inconsistent in data gathering. RWD is messy and requires careful interpretation. For instance, motor function on the right wrist can’t be measured accurately if the wearable is on the left wrist.

From the user’s perspective, simply collecting data without meaningful feedback can induce more anxiety than improve well-being (see the humorous take from Gary Monk below).

The lack of consumer insights is a well-recognized problem, and AI is stepping in to address it. For example, sleep tracking apps can now offer insights into sleep quality and readiness scores, while apps can nudge users to move when they’ve been sitting for too long. And now, generative AI is also playing a role.

It seems patients are making the most out of AI

Patients use generative AI in several ways: to interpret their medical results, to translate medical jargon, to prepare for medical checkups and more.

That said, most patients aren’t scientists, and they may not recognize when AI-generated responses are inaccurate. It’s essential to use AI cautiously, keeping data security in mind and consulting healthcare professionals before making any decisions. Patient advocates Grace Vinton and Grace Cordovano recommend these steps for safely using generative AI:

1. Start with Questions: Use AI tools to prepare for medical visits by asking them to generate lists of questions based on your symptoms or diagnosis. This can help you make the most of your time with healthcare professionals.

2. Refine Prompts: Developing good prompts is crucial for getting the most useful responses from AI. Be specific, and don’t hesitate to refine and adjust prompts to get better results. For example, if you are using AI to interact with an insurance to appeal a denied claim, make sure to describe the situation and your goal (convincing appeal) to AI.

3. Check for Errors: AI can help patients review their medical records for potential errors or duplications, which can improve the accuracy of their care.

To add to that: keep in mind that AI has a specific writing style, and many can now recognize when content is AI-generated. Use AI as a tool for information but always consult professionals (doctors, lawyers, etc.).

Share

Subscribe now

What to be mindful of when using AI:

1. Security and Data Privacy: Patients should be cautious about sharing personal health information with AI tools, especially with open platforms like ChatGPT. Avoid sharing detailed medical results unless you're sure about the platform’s privacy policies.

2. AI Limitations and Hallucinations: AI can sometimes produce incorrect information (known as hallucinations).

3. Confirmation Bias: Be mindful of relying too heavily on AI or the internet for self-diagnosis, as this can lead to confirmation bias. Always consult with a doctor to verify the information.

How are we getting along with AI regulation?

Ever heard the joke, “The good thing about data standards is that there are so many to choose from”? (Cue sarcasm.) The same issue is cropping up in AI regulation. In the U.S., there’s CHAI - Coalition for responsible use of AI, which will provide Assurance labs across the US to test and validate AI models:

The overview of AI assessment process in Assurance labs under CHAI. Source: CHAI

There’s TRAIN - Trustworthy & Responsible AI Network (TRAIN), which was announced at HIMSS Global 2024, and aims to operationalize responsible AI principles to improve the quality, safety and trustworthiness of AI in health.

Working globally is Health AI - The Global Agency for Responsible AI in Healthcare, and HIPPO AI Foundation, which fosters and advocates for open AI models in healthcare and much much more.

What’s universally agreed upon is that AI must be regulated, tested, and made transparent to some extent. As Brian Anderson, CEO of CHAI, remarked, AI should have “food labels” that inform users about model details, intended use, data used for training, ethical considerations, and more.

Slide by Brian Anderson, CEO of CHAI, at DHAI.

In conclusion, and shifting the focus back on Pharma: if eConsents are a challenge for Pharma, I guess we don’t have to worry about AI being adopted at scale in the industry soon.

A longer interview with Brian Anderson about Assurance labs and CHAI will be published soon on Faces of digital health podcast. Untill then, tune into the episode about AI development globally with John Halamka (iTunes, Spotify).

See you in March?

San Diego will welcome medicine, healthtech and innovation trailbraizers between 30 March and 2 April. Program will cover topics such as Exponential Technologies Update, AI & GPT Infused Health & BioMedicine, Hospital to Home, Catalyzing Healthspan, Where Technology meets Policy, Planetary Health, and much more.

Learn more on NextMed Health Website or Watch last year’s recording on their Youtube Channel.

In 2010, Portugal, a country of 10 million, centralized healthcare IT decision-making at the national level by establishing SPMS—a shared services organization within the Ministry of Health. SPMS not only sets the direction for digital health transformation but also builds and manages IT solutions.

In 2018, NHS England started several projects aimed at creating single care records in different areas around the UK. The projects called Local Health Care Records Exemplars are run in London, Manchester, Yorkshire and Humber, Wessex, and Thames Valley and Surrey. For example, One London, connects 42 NHS trusts, five integrated care systems, and 1,400 GP practices.

Recognizing the need for streamlined decision-making, the NHS replaced over 200 commissioning groups with 42 Integrated Care Systems (ICS) in 2022 to enhance healthcare procurement.

Consolidation drives better data exchange.

Let’s look at one more example in Europe: In 2017, CatSalut - Catalan Health Service, published a Master plan detailing the next steps in the development of healthcare digitalization for Catalonia’s 7.6 million residents. It mandated the use of openEHR data specification for all future applications to create a future-proof, data-persistence-focused layer. This decision took away some autonomy from healthcare providers who were up until that point, free to choose what type of healthcare software they wish to use. Variety in software means variety in healthcare data standards used and hence challenges with interoperability. The decision to mandate openEHR aims to enable patients to have a comprehensive healthcare record with all their health data in a structured format. That’s the aim for Catalonia by 2030, says Jordi Pierra Jimenez, Director of the Digital Health Strategy Office at the Catalan Health Service. Further more, Spain aims to take a similar path on the national level (see the full explanation in the video).

A common thread in these three examples is a recognition that the market can’t regulate itself and if we want to connect and enable easier flow of healthcare data more broadly, on regional, national, and cross-border levels, it has to be done by policy first.

Why does data standards unification matter?

Technically, systems can integrate using messaging standards for data exchange. However, in practice, each integration is time consuming and somebody needs to pay for it. Vendors don’t integrate just because it’d be good for healthcare.

Second, the examples above show that a broader change requires defined decision makers. You know how when you have 10 colleagues involved in an email chain, and no one reacts because everyone expects someone else to do the first step? That’s how healthcare collaboration and data exchange can easily get stalled because everyone is waiting for others to start interoperability efforts. This is the commonly mentioned role of EHDS - it’s the usual suspect that gives governments an excuse to mandate specific healthcare data strategies on national levels, dedicate funding to support healthcare IT investments, etc.

Subscribe now

Share

Why is digital transformation unpopular for decision makers

Healthcare IT is very complex. Development, especially when done with clinical engagement, takes time. Hospitals need robust infrastructure: good network coverage, high-speed internet, and reliable hardware. Without these, software can be slow or buggy, frustrating users and impeding adoption, and in addition frustrating vendors, because they can’t do anything about the suboptimal end result.

Why does this matter in the broader political context? For policymakers, tangible investments like new hospital buildings or CT machines have more visible impact than digital transformation, which is why digitalization often isn’t a priority. This is especially true in smaller countries, where fear of failure can deter long-term investments.

France, Catalonia and even Germany are ambitiously stepping on new paths with grand ideas and digitalization promises. But ask any Germany about digitalization in healthcare and they will grunt with dissatisfaction. Digitalization is a slow process.

All the challenges countries and ministries need to overcome

Governments face numerous challenges in digital healthcare transformation. They need to have good experts who care more about public service then their own pockets. Ministries need to know what they want and articulate that in detail. This is where consultants often come in. They prepare a strategy… and then leave. As mentioned by Mehdi Khaled in a discussion about consultancy in healthcare, one of the bigger challenges for governments and ministries is the lack of knowledge transfer between consultants and local workforce that would make strategies more easily executable. Strategies can end up shelved, or governments might need to keep paying consultants to make progress. Other challenges for governments, include:

Defining the problem correctly: Often, the problem is ill-defined, leading to suboptimal solutions. If consultants come in when the problem is already defined, they can have a harder time making meaningful impact on a project.

Measuring impact: There's often a lack of clear metrics to measure the impact of digital health initiatives on overall health outcomes.

Vendor management: Governments often struggle to evaluate vendor claims and ensure that technologies are appropriate for their specific context.

Limited budgets and timeframes: Governments often have to work within electoral cycles or limited budgets, which can lead to short-term thinking.

Lack of accountability: Many healthcare systems lack clear accountability structures, making it difficult to implement and sustain changes.

Subscribe now

Healthcare data potential vs. reality

It’s fairly easy to understand the dream version of the healthcare of the future: data available to patients and providers regardless of where they seek care, personalized care based on all past patient experience and medical history. The challenge is in HOW to get there. August and September episodes of Faces of digital health focused on several aspects of digital transformation strategies in healthcare: from hospital level with Anne Forsyth, Director of Clinical Applications and Support at Women’s College Hospital in Canada, exploring regional care coordination with Jordi Pierra Jimenez, and national strategies and considerations with Cátia Sousa Pinto, Head of Global Digital Health and International Affairs at SPMS - shared services of ministry of health of Portugal. And to add a broader view to that Mehdi Khaled, Managing Partner at Seha consultany, shared what governments struggle most when creating new digital health strategies.

Key things Ann Forsyth recommends hospitals is this:

1. Have a data governance structure in place. Data standardization, defined data stewardship, data access and control, quality management and cybersecurity are fundamental for success.

2. Relationships are key to success. This refers to successful implementations, which requires buy-in from the end users, training, mindful workflow design. Properly designed workflows and effective training help reduce the administrative burden on already overworked healthcare providers. Workflow issues need to be addressed upfront, since it is easy to put in place a bad workflow but incredibly difficult to change it once thousands of users already adopted it.

3. Look for gradual improvements. Large-scale change can be overwhelming. Start with a smaller scope, and always compare the before and after effect. When mapping workflows, Anne’s team started with registration and check in. They divided the project in smaller chunks to be able to test the best practice and training. Further moto is: keep improving, keep optimizing, keep standardizing.

This advice is valuable for other organizations too. FHIR has been instrumental in driving excitement about interoperability. In Europe, openEHR data specification is gaining momentum in the NHS, Spain, Sweden, Slovenia, and beyond.

I’ll be further reporting about this from the openEHR conference, taking place in November in Reading, UK. :)

Subscribe now

Share

Don’t miss!

Want to learn more about the future of health? Join the one conference attendees come to focus on listening to the program, not networking during sessions. NextMed 2024 is happening in December in San Diego.

Here is my summary from last year’s program.

See you there?

Who will be there among others:

Get a sneak peak in the discussion for Faces of digital health, Founder & Chair of NextMed Health Daniel Kraft.

Currently, there are 64 applications in the DiGA repository in Germany. By January 2024, six applications were removed due to various issues faced by the companies. In 2023, France introduced the PECAN framework for the approval and reimbursement of digital solutions, including telemedicine and DTx. By June 2024, four solutions completed the approval process, with one solution for telemonitoring being approved. Three were rejected for various reasons, resulting in a 75% rejection rate, as noted by Elsa Duteil at the DTx France 2024 conference in July. In the U.S. companies stopped waiting for reimbursement codes and are focusing on value-based care and innovation in business models. Why is progress with DTx so slow?

Subscribe now

First, a history lesson.

I still remember the confusion around digital health apps in 2017: Are innovators onto something? Is it all just hype? The number of digital health smartphone apps kept rising and we stopped counting them once their count surpassed 350,000. It became clear the industry needed more differentiation and new labels in this sea of solutions.

Digital health categories according to the Digital Therapeutics Alliance.

That marked the birth of digital therapeutics - digital solutions that, by definition, should have a therapeutic impact on patient outcomes, proven in clinical trials. We celebrated the first FDA-approved prescription digital therapy reSET, a product for treating substance use disorders, in 2017, designed by Pear Therapeutics - the rising star in the digital health space… until it filed for bankruptcy in 2023 and sent shivers through the industry.

Many other solutions were subsequently approved by the FDA, and a lot of these were categorized as requiring a prescription. The idea was, that this would encourage regulators to speed up development of reimbursement models. Some companies chose to go public, but are now facing low valuations and declining stock prices. (Read this in-depth analysis by investor Marko Kuisma for more details and a VC perspective).

As a patient, I often regretted the decision for prescription-only approaches because it meant limited availability. A few years later, due to lack of success with prescriptions, companies are moving away from not being available directly to consumers.

In my opinion, this is good because many digital therapeutics are aimed at mental health and are based on cognitive-behavioral therapy principles. With the shortage of mental health professionals, patients are scrambling to find solutions on their own. Having the option to try DTx as a consumer is a ray of hope.

The Challenge With Innovation in Healthcare

As with any innovation, it soon became apparent that we need a pathway to include digital therapeutics in healthcare provision. Several things usually need to happen to enable this:

1. The solution needs solid evidence of efficacy and impact on patient outcomes.

2. Medical experts need to agree the solution has medical merit, and potentially be excited about it. The impact is presented in the medical community, and hopefully, the use of an innovation will gradually spread. With use comes more data and insights into the real-world applications.

3. If successful, the solution then gets reimbursed by insurance companies, which first want the green light from the medical community, a cost-benefit analysis on treatment outcomes, and often a push from the legislators. Easy.

Ironically, there is a general consensus in healthcare that digitalization is an integral part of healthcare delivery because it will bring transparency, optimization, and more. We want digital data, audit trails, and data pools for data mining, algorithms, and AI development. We often get stuck, however, with the resistance to change the existing delivery models.

Look at Belgium!

Belgium is rarely mentioned in the DTx discussions but a country that made great progress in setting up the environment for digital therapeutics to thrive even before Germany. In 2016, the Belgian Minister of Health launched a call for mobile health pilot projects. 24 subsidized projects were supported and it became clear after evaluation that a validation framework was necessary. The Mhealth Belgium initiative started as a joint effort by industry and government, with the aim of raising awareness and providing a government-approved listing of medical apps through the mhealthbelgium.be portal. How awesome is this? All the stakeholders together, on a joint mission to give innovation a chance. The validation pyramid with three layers was established by the end of 2018.

Source: Mhealth Belgium.

A unique validation pyramid classifies apps into three levels, starting with basic CE certification and culminating in potential reimbursement, following stringent ICT and economic value criteria. However, success didn’t follow. “The difference between Belgium and the two other countries (Germany and France) who are also embracing the innovation is that they have budgets for this innnovation direction. Sometimes large budgets and they also finance the products themselves,” says Dr. Steven Vandeput, Advisor for Digital MedTech and Services & Technologies Home Assistance in Belgium.

Politics is Everything

Any digital health transformation project needs to start with someone who is willing to set the course of progress and who is willing to fight the windmills. This is what the digital health space will remember Jens Spahn for—he was the German minister responsible for setting up the DiGA framework for digital therapeutics in Germany in 2019. I doubt there is a digital health conference, where DiGA wouldn’t be mentioned when talking about digital therapeutics. The success of that framework in 2024 is debatable.

By the end of 2023, 374,000 prescriptions were issues and costing the German healthcare system 113 million euros. Some argue the uptake is too slow, while optimists point to the positive trend of rising numbers of prescriptions. A positive recent change for patients is that in the past, they had to wait up to two weeks to get the reimbursement code from insurance companies. In 2024, this has now been reduced by a two-day rule.

The reality check in 2024

As nicely shown by the overview below, provided by Galen Growth, four countries in the world have a clear framework for digital therapeutics.

Source: Galen Growth.

But we still have a long way to go. The remuneration amounts in Germany agreed between the National Association of Statutory Health Insurance Funds (GKV-Spitzenverband) and the manufacturers, or determined by the arbitration board are on average 221 euros, which is half less then manufacturers initially imagined.

Manufacture prices (gray) vs negotiated reimbursement prices (red) with GKV-Spitzenverband for Digas. Source: National Association of Statutory Health Insurance Funds (GKV-Spitzenverband)

In France, only one solution has been approved so far through PECAN. But to be fair: the pricing range for digital therapeutics has only been introduced in a few months ago. After the experience in Germany, where solution providers set their prices and then re-negotiated them, France wanted to give innovators predictability with pre-set reimbursement rate. DTx gets a flat rate of €435 including VAT for the first three months for a digital therapy. This includes a 20% tax, causing the actual amount to be even lower. The maximum cost of DTx that is reimbursed is €780 per patient per year - a limit set to manage healthcare expenditure, and a limit that was developed through discussions with many industry stakeholders, was explained at the DTx France 2024 conference.

One obsticle companies face is providing clinical evidence for their solutions. As mentioned at DTx France 2024 conference, there is a need for funding well before the start of PECAN to finance this clinical study, which is a significant investment for startups.

Subscribe now

At least two challenges for DTx exist on all markets:

1. Patients and doctors need to be aware of DTx, and startups don’t have marketing budgets the size of Pharma to make enough noise.

2. Doctors need to not only be aware of DTx, but are in addition expected to explain them and train patients in their use. Is anyone surprised they want to be paid for that effort?

What seems to be working?

1. Creativity in business models.

2. Partnerships with consumer brands and pharma.

3. Value-based contracting.

Let’s look at examples.

1. FibriCheck is a medically certified detection and monitoring application that can be used to detect or monitor an irregular heart rhythm, such as atrial fibrillation. FibriCheck can be downloaded from the App Store or the Play Store, or can be prescribed by your physician. You can measure your heart rhythm by placing your finger on the camera of your smartphone. It was developed in Belgium, and when it became clear, that the insurance and DTx frameworks path will make it difficult to scale, the company partnered with Fitbit smartwatch to detect heart rhythm irregularities.

2. In 2021, Vicore Pharma decided to develop a DTx for anxiety specifically designed for patients with pulmunary fibrosis (PF). It was meant to be a complimentary but stand-alone treatment. Instead of developing it in-house, Vicore Pharma AB partnered with Alex Therapeutics - a DTx software design specialised company. “We started from nothing, setting up interviews with pulmonary fibrosis patient groups as a first step. Alex Therapeutics have designed and built the front and back end and cognitive-behavioral therapy expertise, but Vicore Pharma contributed the clinical information on pulmonary fibrosis, connected with key-opinion leader in the PF space and PF Community, designed and financed clinical study study. I think a division like that works really well,” says Jessica Shull, Director of Digital Health at Vicore Pharma AB. In March 2024, the solution called Almee received FDA Breakthrough Designation.

3. Digital therapeutics, with their potential for exponential scalability and reduced marginal costs, fit well into the value-based contracting model, says Aaron Gani, Founder and CEO of RealizedCare, formerly known as BehaVR. As he says, DTx appeal to risk-bearing, value-based entities by promising enhanced member engagement, improved health outcomes, and overall savings, without the entities worrying about specific reimbursement codes. Focusing solely on CPT codes and fee-for-service reimbursements overlooks the broader opportunities presented by value creation. In July 2024, the U.S. Centers for Medicare & Medicaid Services (CMS) issued new changes to their Shared Savings Program for 2025, which opens up new possibilities for DTx companies in the US.

Now what?

To answer the question in the title of the newsletter, the story of DTx is far from over.

Investor Marko Kuisma believes that from a VC investors’ perspective, this might be a good time to invest in DTx, as valuations have returned to a meaningful level after the 2020-2021 hype. “For VCs, taking risks on various factors is business as usual, and the key question is, which risks are you comfortable taking – e.g. are you willing to take an efficacy / evidence risk, technological risk, or market / GTM risk,” he says.

It takes over ten years to bring a drug to market. It takes between 3-7 years to bring a medical device to market, and a lot longer for it to become a standard of practice.

Should we give up on DTx? Not quite yet.

Tune in!

Faces of digital health podcast related to DTx.

• DTx in 2024: Where Are We With Business and Reimbursement Models? (Andy Molnar, CEO of Digital Therapeutics Alliance)

• DTx in Belgium: It's Time to Embed Digital Therapeutics in Clinical Pathways (Steven Vandeput - Medtech Advisor)

• How Do Pharma and Digital Health Converge in 2024? (Amir Lahav, Pharma Consultant)

• Innovation at Charité, The Problems With DiGAs and the German Market (Dorothée Marie-Louise Döpfer)

It’s almost impossible to summarize a conference that had more than 6 stages (Main, Sound, Sight, Health, Touch, Aroma, Forum, Startup …), coupled with tons of parallel events and partner programming. You can access recordings of all sessions on the HLTH Community platform.

The topics below are only a fraction of the content presented at HLTH Europe.

Art Gallery at HLTH Europe showcased art installations from three patient groups from across Europe, alongside digital submissions from clinicians, patients and people working in health.

Europe: Diverse and Fragmented

When examining the European healthcare market from a business perspective, one finds few positive connotations: the market is fragmented. Every country is different, each insisting on its own specific regulations. Do you speak English? Great, but in Germany, you’ll need a local German-speaking team member. In France, ensure you have a French-speaking team. In Switzerland, be prepared to offer your solution in four languages. On the positive side, healthcare is relatively accessible. However, even this is under increasing sustainability pressure – in the NHS, 8 million people are currently (data from April 2024) waiting for hospital treatment.

Subscribe now

Despite these challenges, the European values of solidarity and healthcare accessibility remain strong. Governments are investing in digitalization as a means to improve efficiency and sustainability, and startup founders still aspire to heal the market with their innovative ideas. Most companies in Europe still focus on the European market, with 17% aiming at the U.S., according to Sara Schmachtenberg, Digital Health Venture Expert and Head of Data and Analytics for Europe & USA at Galen Growth.

DiGA, PECAN, here we come!

At HLTH Europe, a hot topic was the reimbursement and regulation of digital therapeutics. How does one become a DiGA app in Germany? If you successfully undergo the approval process with the regulator BfArM, your solution needs to be reimbursed by insurance companies. As of June 2024, there are just over 50 apps in the national repository, a possibility available since 2019.

In 2023, France unveiled their PECAN program, a new reimbursement pathway for digital therapeutics and remote monitoring solutions. By June 2024, one solution for telemonitoring was approved. Reimbursement rates for DTx were defined in April this year and will range between 435 EUR and 780 EUR in the first year after approval and will be reassessed later.

Share

6-7 solutions are expected to be approved by the end of 2024 through PECAN, said Stephane Tholander, Co-founder of AppTherra - the digital therapeutics distribution platform.

Getting through the approval process is loooong. In Germany, it never takes less than a year and a half, said Dorothée Marie-Louise Doepfer, Deputy Head of Digital Health Accelerator at the Berlin Institute of Health at Charité, who chaired a session on the German market.

If certification seems demanding, getting through the approval process is actually the easy part of a DTx company’s journey. People in the digital health space may know what digital therapeutics are, but most clinicians don’t. Getting your distribution channels right and educating doctors about the clinical validity and usability of your solution is the real Herculean task. Oh… And then the doctors also need to educate patients about the proper use of these new treatments. If implementation is 60% of success in digital transformation, the same is probably true for patient education and the success of digital therapeutics.

Share

EU AI Act, EHDS and credibility scores for generative AI

Most of the discussions around AI stayed high-level and focused on the potentials of technology. Hype around generative AI has settled down in 2024, with the ongoing race, especially in the US market, around which will be the number one clinical note taking and summarization solution provider.

According to Galen Growth, by the end of 2023, 2% of all digital health ventures were using generative AI. AI more broadly is a different case with over 50% were using some form of AI.

In a short discussion (which will be published as a podcast in July), John Halamka, President of Mayo Clinic Platform shared his view on the European regulation of AI and data, the latest on generative AI testing and cross-border collaboration for data exchange.

In short: GDPR is not a restriction on the secondary use of data, it’s a guideline on “the how”, he says.
In the field of generative AI regulation, companies are beginning to develop credibility and quality measures for generative AI. Since the quality of AI-generated answers can fluctuate, assigning a numeric credibility score will help determine when an answer is reliable or if further prompting is necessary.

Subscribe now

Generative AI will always produce some inaccurate and subpar responses, posing varying levels of risk depending on the use case. Hence healthcare requires consensus on what this technology should or shouldn’t be used for, says Halamka. For instance, using AI to summarize historical records is low-risk, while allowing it to decide on treatment plans is high-risk and potentially harmful. Therefore, prioritizing low-risk, high-yield administrative use cases is recommended.

He sees the Netherlands, Sweden and Denmark as the European frontrunners in their willingness to take risks and implement AI.

Girls, Ladies, we need to invest

One thing that got noticed by many at the conference, was a dedicated women’s health hub, standing at the Heart stage in the middle of the show floor.

Women’s health was front and center, emerging as a key topic set to shape digital health in 2024. This focus is unsurprising given that 80% of patients with autoimmune diseases and 66% of Alzheimer’s patients are women. Women also experience a higher incidence of migraines, and cardiovascular diseases present differently in women than in men. Many of these statistics still lack clear explanations.

“Women’s health is not a niche. 55% of the population is not a niche. So don’t invest in women’s health because it’s the right thing to do, but because it’s the smart thing to do,” said Priya Agrawal, VP of Health Equity & Partnerships at MSD, during a panel on women’s health, highlighting the significant business potential in this field.

The discussions and booths covered a wide range of topics, including methods for gathering DNA after a sexual assault (women have a 72-hour window to collect DNA from the perpetrator, which can be crucial evidence in court), sanitary tools for use after intercourse, wearable devices with sensors to monitor fetal health during pregnancy, wellness at-home blood test kits, and more.

“We put the patient at the center”

…says almost every company targeting patients... Yet, they often forget to include patients in the discovery process and product design. By only involving patients after the final product is ready, companies shouldn't be surprised if patients find their solutions useless.

This is what Jen Horonjeff, PhD, Founder and CEO of Savvy Cooperative, and Ronnie Sharpe, co-founder and COO at Savvy Cooperative, say one shouldn’t forget about patients and product development:

• Patient-Centered Outcomes: Involve patients from the start in drug development to determine which outcomes and endpoints matter most to them, and iterating on their experiences throughout the trial and commercialization phases.

• Feedback and Co-Design: Incorporate patient feedback in the development of digital health tools, avoiding the assumption that sponsors or doctors know best, and ensuring tools are co-designed with patients to meet their needs effectively.

• Integration of Tools: Ensure that digital tools integrate seamlessly with each other and other devices patients use, to avoid adding frustration and burden from managing multiple, disconnected tools.

• Access to Data: Provide patients with access to their data, as many track their health using various devices, and addressing concerns about data sharing to empower patients in managing their health.

• Iterative Development: Utilize usage data to inform the iterative development of digital tools, engaging with users to refine and improve tools rather than abandoning them if initial usage is low, similar to product development in other sectors.

If you’re a podcast fan, tune in to the HLTH Europe recap episode with some of the speakers mentioned above and others:

• John Halamka, President of Mayo Clinic, about AI regulation and future,

• Sara Schmachtenberg, Digital Health Venture Expert / Head of Data and Analytics, Europe & USA at Galen Growth, about who ventures partner with,

• Simon Phillip Rost, the Chief Marketing Officer at GE Healthcare, about patient engagement,

• Jen Horonjeff, PhD, Founder and CEO of Savvy Cooperative, and Ronnie Sharpe, co-founder and COO at Savvy Cooperative, about including patients in product development,

• Christophe Jauquet, Author & Professional Keynote Speaker on how business & technology shape a healthier, happier, more sustainable future, about values to consider when designing for digital health,

•  Dorothée Marie-Louise Doepfer, Deputy Head of Digital Labs / Program Management Digital Health Accelerator & Community Building at Berlin Institute of Health at Charité, about women’s health and entering the German market.

This is it for this edition. Feedback is always welcome, feel free to share your thoughts in a comment or email.

Enjoy the summer!

Leave a comment

Share

Before you leave…

🎥 🎞️ [VIDEO] 🎥 🎞️

Head to Youtube to hear the following discussions:

➡ John Halamka on AI in healthcare, regulation and country-specific differences.

➡ Sara Schmachtenberg, Galen Growth | Insights You Can Trust about who digital health companies partner with, and which markets different continents focus on.
➡ Zayna Khayat, Ph.D. and Lucien Engelen about their observation of carification of retail, phlebotomy automation, hope and upstream health.

➡ Andrew Lacy, CEO of Prenuvo, about influencer marketing, whole-body MRI scans and prevention.
➡ Peter Birch, Host of Talking HealthTech and Dr. James Somauroo, Host of The Healthtech Podcast about LGBTQ healthcare, women's health, communities and social media and healthcare.

Rome was buzzing with digital health at the end of May as the HIMSS Europe conference took place at La Nuova Congress Center. Delegates from 80 countries and representatives of health ministries of 20 countries marked the event with discussions about the implementation of the European Health Data Space (EHDS), the EU AI Act, and more.

EHDS and the EU AI Act have been setting the framework of rules and hopes we want to see in Europe in future development of technology. The key aim of EHDS is to enable cross-border health. The discussions now resolve around the HOW and practical implementation of relevant health standards.

Different countries in Europe are at different levels of digitalization and giving patients access to their data. Countries such as Portugal, Slovenia, Croatia, Estonia, Poland and Hungary, enable patients to access national portals that contain their prescriptions, referrals, discharge letters, etc. In other countries such as Belgium and Spain, patients can access regional portals. In the Netherlands, Germany and Switzerland, individual healthcare providers offer their own patient portals. This means if a person visits multiple facilities, they must use different portals to access their health data. 

Subscribe now

Let’s start with data

European diversity can be great and inspiring if you’re on a backpacking trip across countries. But it becomes a nightmare when you try to harmonize expectations around health data standards and privacy on the European level.

While some consensus has been achieved, it’s time to be more ambitious, says Jordi Piera-Jiménez, Director of the Digital Health Strategy Office, Catalan Health Service.“To date, a lot of work has been done to define the European electronic health record exchange format and other standards. These promote I would say the exchange of PDFs translated into English language. This can be a starting point for the European health data space. But while talking about exchanging PDFs translated to English language can be a starting point, I believe that in the 21st century European citizens deserve more.”

He believes information systems of the future should be based on re-built foundations, and translation of medical language in technical language. This is what the openEHR specification, built and continuously improved by clinicians, aims to do: define clinical concepts to a very granular level, to make sure all the relevant information is captured in a structured format. FHIR messaging standard is the leading standard for data exchange, but doesn’t necessarily capture in data all the relevant clinical concepts.

Speaking of connectivity and digital transformation: HIMSS has several digital maturity models:

• EMRAM - Electronic Medical Record Adoption Model,

• AMAM - Adoption Model for Analytics Maturity,

• INFRAM - Infrastructure Adoption Model,

• C-COMM - Community Care Outcomes Maturity Model,

• DIAM - Digital Imaging Adoption Model.

These models are used by healthcare institutions as guiding principles in digital transformation. Many participants shared their experience with reaching different levels of maturity. Cleveland Clinic London, for example, reache EMRAM Stage 6 (out of 7 stages). The hospital doesn’t use paper."This includes closed-loop medication management and connecting all devices, such as infusion pumps and vital signs monitors, to the EHR. This integration allows for seamless data transfer and minimizes the risk of errors from manual transcription or data entry. By reducing the need for manual data transcription, we enhance efficiency and patient safety. However, the connectivity between numerous devices introduces new requirements and requires new caution due to increased cybersecurity risks. When it comes to advice on how to best connect all the devices in the hospital for the seamsless data transfer and make sure staff knows how to navigate the tech-driven environment, Nana Odom, Director of Clinical Engineering at Cleveland Clinic London advises the following:

• Always consider a standardized approach,

• Reduce the number of interfaces and middlewares as much as possible,

• Design should be collaborative - Ensure involvement of all stakeholders (Clinical, Technical, Infrastructure, other relevant stakeholders) from the start - design, testing phase,

• Provide support as the technology is implemented,

• Support technology adoption effectively,

• Clinical engineers should be part of the initial process of requirements gathering and design.

  Share

Can’t AI do it?

When we talk about automation with AI, it’s worth noting, that data structuring and coding still requires a lot of manual work and double-checking if for example the right codes were assigned to the right clinical concepts. Think about it in the context of automatically generated subtitles. They can easily be applied to videos, but you still need to do your part in language checking and corrections, if you want the subtitles to be accurate, especially with names and abbreviations. Expecting generative AI to do all the work in structuring free text, is a popular topic, but not a solution. “Generative AI can be a way to to ease the task of conversion. It’s an effort that it's very costly. It requires a lot of human power. But again I think that we should put efforts in rebuilding the foundations of our information structure,” says Jordi Piera Jimenez.

With AI, let’s start with operations

Most current applications in healthcare are centered around optimization of non-clinical workflows with AI in healthcare.

The EU AI Act classifies AI systems into four different risk levels: unacceptable risk, high risk, limited risk, and minimal risk. Each class has different regulation and requirements for organizations developing or using AI systems. As named by Rachel Dunscombe - CEO of openEHR International and former the UK AI council member leading on AI in digital health and care, the key discussions among the experts and regulators at the moment are:

• the design of national sandboxes for testing data models,

• the necessity for corporate indemnity insurance to protect clinicians who currently take individual risks when using AI,

• an ongoing debate about the pace of policy development: when “the next ChatGPT revolution comes" how will policy-making respond faster?

How is healthcare delivery transforming and what it means for digitalization

WHO estimates a shortage of 10 million clinicians by 2030. With healthcare workforce shortages, aging population, increases in chronic disease incidence, it’s clear that care delivery needs to be thoroughly redesigned to obtain healthcare sustainability.

Value-based care is the prime word in the shift towards optimized care and financial sustainability of healthcare systems.

Prevention doesn’t mean just healthy living and well-being, but more tangible for healthcare systems - the focus on secondary prevention and keeping an eye on existing patients at risk of costly complications. How does this transfer to digital transformation? Svava María Atladóttir is the Executive Director of Development at the Landspítali National University Hospital Iceland which is moving to a completely new hospital by 2030. Taking into account the trends toward at-home care, virtual wards and hospital at home, she sees new care delivery planning includes focus around patient journeys and multi-discriplinary work to ease transitions for patients from one setting to another. “Building the physical building is the easy part. Even choosing the technology is somewhat easy. What is challenges is ensuring that we are changing the processes, how we provide care and that we are really communicating with our patients and with the whole community about the changing approaches to care-provision,” says Svava María Atladóttir.

Navigating the healthcare systems is complex. Through their own experience of patient and caregiver, Luke and Monika Evason pointed out the need to improve transit between pediatric and adult care, because at the age of 18, patients are suddenly on their own, while much of the care was taken care for them just a week before their birthday. I like to compare this transit as falling of the cliff for chronic patients: in pediatrics care is more risk-averse and connections with clinicians closer. In adult care you fall into a much larger pool of existing older patients, which may mean longer waiting times, less contact with clinicians and more self-management of care. Less convient and less safe. Not what we want in 10 years time.

Subscribe now

Keep an eye on

Several interesting EU-supported projects were presented at HIMSS:

• EDIHTA - harmonizing healthcare technology assessment for digital health (ends 2028).

• FLUTE - improving prostate cancer discovery through AI and cross-border data utilization.

• Gravitate Health - improving access to medication information in one’s native language and improving adherence to medication.

• Label2Enable - adding “food labels” to digital health solutions.

• AI4Health.Cro - non-profit public-private consortium of experts in the field of AI, healthcare and startups that see artificial intelligence (AI) as the key to advancing healthcare and medicine.

• Warifa - personalized recommendations for managing NCDs-related chronic conditions such as cancer, cardiovascular diseases, diabetes, and chronic respiratory diseases.

Also, tune in to the podcast episode with full statements of the speakers above!

These include:

• Jordi Piera-Jiménez, Director of the Digital Health Strategy Office, Catalan Health Service, who shared his view and experience on data management and data standards, 

• Rachel Dunscombe, CEO of openEHR foundation and former the UK AI council member leading on AI in digital health and care, who presented the discussions on AI,

• Svava María Atladóttir, Executive Director of Development, Landspítali National University Hospital Iceland, about what it takes to build a new hospital which is what she needs to do by 2030, 

• Nana Odom, Director of Clinical Engineering at the Cleveland Clinic in London, who shared her advice for clinical engineers,

• Diana van Stijn, Chief Medical Officer and co-founder of Lapsi Health and resident of Pediatrics at UMC Amsterdam, who talked about the doctor’s perspective on the use of tech and upskilling, 

• Luke Evason, Consultant and Monica Kleiijn Evason, Leadership Coach and Book Author, who both talked about the patient perspective on gaps in healthcare, particulary at the transfer of children going from pediatric to adult care. 

What’s up next

Moving from the focus on policy and digital health regulation, with 400 speakers and over 3500 attendees, HLTH Europe, focused on healthcare innovation, will take place in Amsterdam between 17-20 June 2024.

If you’re still thinking of going, use the CODE HE24_TJASAZ to get a 150 EUR discount on your ticket!

Let’s connect on Linkedin!

I have a five-year-old, and I’m glad I have a son, not a daughter.

Many parents I speak with share a common belief: if they have daughters, they are convinced that their girls will face greater challenges in life than they would if they were boys. As it turns out, this can translate in health as well.

Migraines are three to four times more common in women than in men. Women have historically been diagnosed with depression twice as much as men, and have a higher risk of dementia. Approximately 80% of all patients diagnosed with autoimmune diseases are women.

We give birth, rehabilitate, go through a midlife crisis, and end up in menopause. Non-stop resilience training, with the constant pressure to be good mothers, exemplar and ambitious employees, and attentive lovers. Yes, women live longer than men but spend more years in poor health. According to Eurostat, in Europe, women spend on average 20 years with disability, 6 years more than men.

How did we get here? For decades, science has neglected the female body’s impact on health. At some point, it became clear that pregnancy carries specific health risks, so women of reproductive age were historically excluded from clinical trials. 4% of biopharma R&D budgets are focused on women’s health. Four percent. In many aspects, we are in the dark about where impacts on health come from. “We still don't know why female hormones play such a major role in many disease areas, not only in migraine. We have the same unspoken issue and unsolved mystery in terms of prevalence, disease progression, and diagnostic treatment response in Alzheimer's disease. We have it in multiple sclerosis,” says Dr Antonella Santuccione Chadha, pro-bono CEO of Women's Brain Project (WBP), a Swiss-based non-profit organization that wants to find scientific answers to these questions, and translate them to political measures that need to be taken to facilitate change across the planet.

Changing the narrative and the consequences of women’s health is far from easy. According to Gabor Mate and research, the reason why most patients with autoimmune diseases are women, is attributed to societally conditioned characteristics such as women often being compulsively concerned about the need for others, they believe they are responsible for how other feel and often suppress healthy anger.

On the societal level, we are making some progress in awareness about the need for a better gender balance and diversity in leadership positions. The problem isn’t just the gender ratio on management levels; it’s how we define value in society. The majority of nurses and professional caregivers are women. The demand for these jobs is growing due to the aging population, yet these professions remain poorly compensated and unattractive. Economics and health go hand in hand, and often women pay the price. “I find it very cruel how women as caregivers are treated. From when we are little girls, we're seen more as caregivers, and are socialized to do all the caregiving. In general, women tend to live slightly longer than men, and when they need to care for themselves, there's no one left to care for them. Consequently, they need to rely on institutionalized care and paid care. So it's almost, oh, thank you for the years you cared for others, but now that it's your time to care, you’re ill, you don't have the financial safety net, and you don't actually have a partner. This is not just about brain health, physical health, and emotional health; it's also about financial health as well. We as women have to be very careful in terms of our financial health and our pensions,“ says Anna De, Head of Stakeholder Engagement at WBP.

Share

Let’s Make Babies

I consider myself extremely lucky: I live in a country that gives women a year of paid maternity leave and the right to an abortion (limited to 10 weeks). I’m horrified by the direction of women’s health policies in Poland, where women can die because doctors fear treating women with severe complications in pregnancy that could be solved by abortion. Or the US, where progress is going backward. Since 2022, in many US States, women lost the right to an abortion and they gained zero in support if they do have a child. In 2022, the United States remains the only OECD country without a federal provision for paid maternity leave, and even today paid maternity leave is not universally available to self-employed, domestic, part-time and casual workers in some other OECD countries.¹ At the same time, if women do want to have a child and can’t have one naturally, medical health is far from accessible to everyone. In Slovenia, we can get a few IVF cycles covered by the healthcare system. In the US, women will mostly get charged for it - The average cost for a complete IVF cycle in the US is about $12,000 plus medications.

I wanted to have a child before I turned 30, because I didn’t want to be an “old” mother and I knew pregnancies get complicated as we age. I had no clue how much more complicated. The Centers for Disease Control and Prevention (CDC) reports that 1 in 5 heterosexual females ages 15 to 49, with no prior births, are unable to conceive on their own after a year of trying, and about 1 in 4 (26%) women in this group have difficulty getting pregnant or carrying a pregnancy to term. Additionally, in one third of cases the cause of infertility lies with men.

“Men experience age-related fertility decline as well. It's not as sudden and perhaps disruptive as it is for women who do have declining egg quality at a certain age. But infertility affects men too, but no one's telling them,” says Leslie Shrock, author, entrepreneur, and angel investor working at the convergence of health and technology. In about 25 percent of infertility investigations, men are never even examined.

Subscribe now

Menopause: Not a Disease, Definitely a Market Opportunity

First of all: menstruation is not a disease and neither is menopause. However, both physiological processes can have a negative impact on the quality of life of women. When we’re in pain, treatment can be helpful. Yet in menopause, women often just need to “deal with” discomfort on their own. Many menopause symptoms impact wellbeing: mood, sleep, hot flashes, bone density, mental health, are cause by hormonal changes.

Menopause research has gone through hoops: hormone replacement therapy with estrogen was approved and used between 1942-1975, and as detailed by a New York Times research, halted after published research showed menopausal women who took estrogen had a significantly increased risk of endometrial cancer. Other treatments have been prescribed since, with another scientific setback in 2002, when an 8 year long clinical trial for a combination of estrogen and progestin, a synthetic form of progesterone — had been stopped prematurely.

WHO considers that social, psychological and physical health support during the menopausal transition and after menopause should be an integral part of health care. In 2021, women aged 50 and over accounted for 26% of all women and girls globally. This also makes menopause a market opportunity - Grand view research estimates the global menopause market size was at USD 16.93 billion in 2023 and is expected to grow at a compound annual growth rate of 5.37 % from 2024 to 2030.

The Startup World is Mobilizing

In 2024, with the rise of advocacy around women’s health, investments are picking up as well. While digital health funding has been falling for the last few years, the percentage of investments in femtech inside digital health investments grew from 7,6% in 2020 to 13,26% in 2022, according to Techcrunch.

Is this progress? According to investor Marija Butković, not big enough. “I've been in this space now 10 years as a woman in tech and the last seven years as a female founder in tech myself. I've spoken probably to thousands of founders over that course of time, and I'm under the impression that situation is worse than ever. I see all these initiatives and conferences and office hours and mentorship programs and I still don't think that we are any closer to the real solution. Not the same amount of money is going to female founders. Nothing almost has changed. This brings me to the conclusion that whatever we are doing now is not enough. Maybe it's not wrong, but it's simply not enough,” she says.

Numbers backing this sentiment are quite depressive. According to Harvard Business Review, women-led startups received just 2.3% of VC Funding in 2020. And here’s the real emotional rollercoaster: The good news is the number of female VCs is rising. Female VCs are twice as likely to fund women-led businesses, especially when they address women-related issues, which men can have a harder time identifying with. However, research published in Organization Science which examined 2,136 startups tracked by the Crunchbase database found that female-founded firms that received funding from female VCs had more difficulty getting funding in the future. In fact, women-founded businesses that received funding from female VCs were two times less likely to raise additional funding than women founders who were funded by male VCs. Researchers concluded that “when women receive support from other women, observers may implicitly believe that the relationship was motivated by considerations other than merit, which leads to a discount in perceptions of competence.”

Hence experts in Women’s Brain Projects and investors such as Marija Butković emphasize governments need to become more engaged in the conversation in policymaking and financial support to solve these issues.

What To Do Next?

1. Education

The concept of women's health extends beyond pregnancy care, breast and ovarian cancer, and fertility. It also includes issues like gender-based violence and aspects of sexual health and pleasure.

In developing countries, explanations for health problems or general human biology can be heavily influenced by traditional beliefs. Some communities still believe that during their period, women are cleansing of evil spirits or that if twins are born, that is a bad omen.

We need to upgrade the knowledge about these aspects early on in the lives of new generations, everything from mental health care and its impact on physical health to safe sex, and more. This goes to all cultures and markets.

Awareness is the starting point of change. If you are interested in Fertility, you might take a look at Leslie Shrock’s books: Bumpin’: The Modern Guide to Pregnancy mixes the latest clinical research with practical advice for working families, or Fertility Rules, published in June 2023, addresses male and female fertility. If you’re wondering about sex and pleasure, Emily Nagoski’s Come As You Are: Revised and Updated: The Surprising New Science That Will Transform Your Sex Life, might give you a better understanding of sexual health.

2. Invest in Women’s Health Knowledge and Science

At the moment, health education startups are paving the way. Mariatheresa Samson Kadushi started health education company Mobile Afya in Tanzania, after having worked with homeless children where she realised that helping them is just treating the symptom, the solution is to raise awareness about contraception and safe sex. Similarly, Shamala Hinrichsen is working primarily with indigenous communities in East Malaysia through the app Hanai. In 2023 they co-founded Sheher - a multilingual mobile app that delivers accurate medical information, shares women's lived experiences, and facilitates digital access to medical care. But these companies face infrastructure challenges and technology affordability issues, not to forget the before-mentioned fundraising challenges of female founders leading these ventures.

I think we live in an exciting and hopeful era. But hope alone is not enough - for it to be justified, it requires action, from everyone. As said by Marija Butković: “Every man on this planet has one or more women within their immediate circle that they care about. You have a wife, you have a partner, you have a daughter, you have a mother, right? And that's why they should care about women's health as well.“

3. There’s plenty of things you can do already this second:

1. Support the Women’s Brain Project in establishing the first worldwide sex and gender precision medicine research institute.

I want to donate!

2. Get educated by listening to the episodes embedded in this article or by browsing through women’s health topics and episodes on the Health Podcast Network, or specialised channels such at HIT Like A Girl Community.

3. Spread the knowledge and share this article.

Share

Let’s connect on Linkedin!

Let’s connect on Linkedin!

2023 is drawing to a close, and as we look back, here’s what was covered over the year:

A lot of AI.

The natural language processing series was a perfect introduction to understanding the difference in how AI is used for symptom checking, voice recognition, data structuring, and where generative AI fits in the picture.

For an even deeper dive into understanding the origins of data for healthcare innovation, how companies in the US consolidate healthcare data, and how it's reused for new findings and improved patient and clinical experiences, you can hear from leaders at Komodo Health, EPIC Cosmos, Seqster, and other key players who are connecting the dots among health data silos in the US. For example, in 2019, Epic launched Cosmos, a special program for data mining of patient records data gathered in Epic systems. Today, Epic Cosmos contains over 227 million patient records.

Komodo Health is currently tracking individual encounters with the healthcare system for over 330 million patients.

We started the year with AI, and also finished it with a special live podcast UPSIDE DOWN: How AI Changed Healthcare in 2023. The future seems to be bright, with speakers expecting a positive impact on the healthcare documentation burden, health equity, healthcare personalization, and a boost in value-based care.

Watch the full discussion or tune in the podcast.

Healthcare Digitalization Strategies in Europe, Healthcare Innovation in APAC, LATAM and Africa

Faces of digital health explores the diversity of healthcare systems and healthcare innovation worldwide. We travelled across several continents in 2023 (eyeing Australia and India for 2024!).

Europe is on a roll with healthcare digitalization

Representatives of the Dutch, German and Slovenian Health Ministries shared their strategies and investments in healthcare digitalization. One big change is expected by the ministries: hospitals must begin considering IT as an integral component of healthcare provision. Consequently, the costs associated with digital services and tools will need to be incorporated into the price of healthcare service provision.

France wants to become the leader in digital health. Learning from the challenges encountered in DTx reimbursement in Germany, France has designed a distinct approval and reimbursement process for DTx companies. This approach aims to facilitate greater success for market players by providing clearer and more navigable rules.

Subscribe now

Great Connectivity and Opportunities in the APAC and more Challenges with digitalization in LATAM

Vietnam, Thailand, and South Korea all pride themselves on a large proportion of the population aged between 15-64 years old. As mentioned by Beth Ann-Lopez, CEO of Docosan, there’s a lot of highly educated talent in the region, making growing a company here a good opportunity.

The Chinese healthcare market is experiencing significant momentum from government policies, industry advancements, and big tech involvement. However, entering the Chinese digital health market as a foreign entity can be challenging due to strict data privacy laws and regulatory authorities. Also, if you think you have a unique service, get ready to be copied. Read more.

In LATAM, plenty still needs to be done for national digitalization progress: with the fast technological advancement, healthcare can sometimes get caught up in legislation challenges. For startups, many collaborations happened during the pandemic, but have hit regulatory walls after.
Mexico has a predominantly public healthcare system, with workforce shortages. In situations like these, the government is often focusing on closing the gap between people accessing basic care, rather than investing in tech infrastructure, says Rafael Lopez, CEO of Diagnostikare.

The challenge in Argentina is the regional fragmentation, and region-based rules, which make scaling across the country challenging.

Africa

Africa is sometimes mentioned as a continent with great potential for startups developing more affordable medical devices. In the West, clinical experts complain that these innovations are below the clinical standard. But in an environment where something is better than nothing, these solutions have a great impact. African countries are facing healthcare scarcity on every corner, however, with the aid of foreign investments, innovation is progressively addressing these challenges, one step at a time.

Subscribe now

⭐ Top episodes in H2 2023 ⭐

Podcast downloads build up over time. This gives episodes earlier in the year an advantage, hence I’m listing top 5 episodes based on downloads in H2 2023:

1. 🏆 Bringing ePrescribing To The Next Level: To Patients (FDB (First Databank, Inc.) Vela, Lathe Bigler)

2. 👏 What Are Investors Betting On In Generative AI in Healthcare? (Justin Norden, MD, MBA, MPhil, GSR Ventures)

3. 👏 Europe: How are Slovenia, Germany and the Netherlands Envisioning Future Healthcare Digitalization? (Bianca Rouwenhorst, Sebastian Zilch, Alenka Kolar)

4. 👏 How Can Digital Tools Aid Long-Term Chronic Care Management? (Sean Duffy, Omada Health)

5. 👏 How Can AI Help Predict Patient Drug Response? (Rafael Rosengarten, Genialis)

Which Digital Health Events Will You Attend in 2024?

We’re entering 2024 with many digital health gatherings to look forward to. The events industry is in full bloom. Here is a list of events happening in H1 2024 in Europe:

February

• European Digital Health Excellence Forum: 14 and 15 February, Rotterdam, Netherlands

April

• DMEA: 9 to 11 April, Berlin, Germany

• BOOM Summit: 10 and 11 April, Basel, Switzerland

May

• ICT&health World Conference 2024: 14 May 2024

• Vitalis: 13-16 May, Gothenburg, Sweden

• Radical Health Festival: 22-23 May, Helsinki, Finland

• HIMSS Europe: 29 - 31 May, Rome, Italy

June

Bits & Pretzerls Healthtech: 5-6 June, Munich, Germany

HLTH Europe: 17-20 June, Amsterdam, Netherlands

Share

5 reflections on healthcare in the US and Europe …

HLTH is recognized as the United States' number one health innovation event. In 2024, HLTH is coming to Europe. In a special Faces of Digital Health podcast episode, filmed in Ljubljana, Slovenia, Jody Tropeano, Head of Content at HLTH US, and Henry Stoneley, Head of UK and Netherlands, reflected on healthcare in the US and Europe, key trends, and challenges. They also shared their top tips when attending large conferences:

1. Prepare, prepare, prepare. Not when you arrive to the event but a week or more before.

2. Figure out your goals and objectives. If you want to learn, analyze the agenda and mark the sessions you want to go to. If you want to network, set-up meetings in advance.

3. Make a plan and stick to it. Serendipity is a part of the experience, don’t count on it as your sole strategy.

Many conferences offer event apps that help with networking and planning. Make sure to check them out.

Watch the full interview on Youtube:

Conference organisers need to know the key trends and dynamics in the industry to be able to build an appealing program that attracts the right audience. In the interview with Jody and Henry we discussed how they see they key topics and trends in the US and European Healthcare.

Here are a few takeaways.

1. DTx are “out of fashion”; it will take time for generative AI to be used for clinical use cases

Digital therapeutics have had and exciting few years, but that excitement subsided significantly in 2023, at least in the US, one of the challenges being a lack of a reimbursement system to fund these solutions “I find digital therapeutics fascinating. But as it often happens in healthcare, we get this new idea, everyone is excited how it may change the industry, and then you realize after a certain number of years, how hard the US healthcare system is, a you have to go back to the drawing board,” said Jody Tropeano.

At the moment, the key topics in the US remain consumerism, what are players such as Amazon doing in healthcare and what their strategy is moving forward, the boom of Medicare Advantage in the US, and generative AI, with the focus on how will it make the work of clinicians easier for documentation management.

2. It Will Take a Decade For Value-Based Care to Have a Broad Impact

Despite its potential to improve healthcare outcomes and costs, scaling value-based care remains challenging. Value-based care use cases are seen a lot in the specialty care space, for example in oncology, MSK and gastroenterology, but he U.S. healthcare system, predominantly operating on a fee-for-service model, has shown slow progress toward widespread adoption of value-based care, indicating a significant transformation still needed.

3. Generative AI's Emerging Role and Cautious Optimism

Generative AI is seen as a promising development in healthcare, with potential benefits for clinical efficiency and workforce support. However, its application is in early stages, with only a fraction of digital health startups utilizing generative AI. The anticipation surrounds its future impact, especially in improving healthcare delivery and provider satisfaction.

4. Healthcare Workforce Dynamics and the Need for Sustainable Models

There's a growing concern about the workforce shortage in healthcare globally, with Europe facing additional complexities due to workforce mobility within the EU. This situation underscores the urgency for reconfiguring healthcare delivery to create sustainable models, considering the impending workforce shortages and the varying healthcare systems across different regions.

5. The Dilemma of Catering to the "Worried Well"

There's a recognition that the healthcare model should shift from being reactive (treating illnesses as they occur) to proactive (focusing on wellness and prevention). This shift might inadvertently cause a surge in healthcare resource consumption by those who are more health-conscious and affluent, have the buying power, and are primary customers for healthcare startups.

Investments often target solutions that cater to profitable markets, typically the "worried well." The expectation is that successful innovations in these areas might eventually trickle down to broader populations.

All the best in 2024! Stay healthy and stay tuned!

Let’s connect on Linkedin!

Each year, the focus of excitement in digital health innovation shifts to a different area. In 2016/2017 it was mobile apps. I even recorded a documentary about the hopes and hypes around digital health, to figure out how big of an impact can we expect from mobile health apps. Then it was AI. In 2017/2018 a plethora of blockchain startups popped up, promising transparency, compensation with tokens for medical data sharing, and bigger safety of healthcare data among other things. When the pandemic struck, it seemed healthcare would forever shift from in-person consultations to a predominately telemedicine-centric model.

In 2023, it seems that all we hear about is generative AI. However, if we look at the data and reports, several things become apparent.

We’re in the early stages of the use of generative AI in healthcare.

Let’s rewind for a year.

The broader public became aware of the power of generative AI and what it looks like with the launch of ChatGPT in November 2022. Just one week following its widespread release to the general public, ChatGPT achieved a milestone by attracting 1 million users. After a month, it had 100 million users, and in November 2023 the number of users was 180 million.

Many of the digital health companies have been using generative AI well before 2023. GSR Ventures and Maverick Capital prepared a nice overview of generative AI healthcare companies, putting those that launched their products after 2020 in red brackets:

In essence: many players have been here for a while.

According to Galen Growth report on generative AI in healthcare, only 2% of digital health ventures that use AI, also use generative AI. However, 72% of digital health startups working in biotech use generative AI.

Beyond generative AI, “old” AI is on the rise, with ups and downs in the growth of new FDA approved devices. The FDA database of AI/machine learning-enabled algorithms had almost 700 entries in October 2023. The majority of approved algorithms are used in radiology. And by majority, I don’t mean a bit more than 50% but actually 87% of all approved AI/ML devices in 2022 and 79% in 2023. The year-over-year increase of FDA approved AI/ML-enabled devices was at 39% in 2020, fell to 14% and 15% in 2021 and 2022, and is predicted to be 30% in 2023.

Subscribe now

All these algorithms aim at increasing efficiency. For example, in June the company Ezra got a 510(k) clearance for their algorithm, which makes Full Body MRI, which previously took one hour, now available as a 30-minute MRI scan. (BTW 510(k) clearance doesn’t refer to completely novel devices but is a clearance for devices that are equivalent to a device already placed into one of the three classification categories on the market).

GenAI Brough Hope to Healthcare

Regardless of only 2% of digital health ventures using generative AI, 2023 was the year of generative AI even in healthcare. It brought hope that we are close to getting the tech right in healthcare instead of causing EHR-related clinician burnout and PTSD from overwhelming administrative tasks in collecting healthcare data. Doctors gave a warm welcome to AI scribes. These generative AI-powered assistants are capable of recording, writing, and structuring medical notes in real-time during patient visits, signifying a major step forward in healthcare technology. At the moment, that’s about it, as far as making doctors’ life easier goes.

Beyond Generative AI: The Flashy and the Ugly Side of AI in 2023

Among the flashy news of AI-powered digital health innovations was definitely the launch of Forward Health CarePods - self-sufficient pods deployed in malls, gyms, and offices in the US. The pods offer members willing to pay a 99$ monthly membership fee to get a full body scan, heart health, thyroid testing, blood pressure, weight management, diabetes screening, COVID-19 test, HIV screening, and kidney, and liver health. The results of the tests are seen by a doctor in the background, who can then write a prescription if needed.

Among the Black Mirror type of AI-related revelations in 2023 was the news that UnitedHealths pushed employees to adhere to the recommendations of a flawed algorithm, leading to numerous patients being unjustly denied rehabilitation care they needed. It was an example of the greatest fear of AI coming true.

AI algorithms are usually trained on historic data and can be prone to errors when applied to real-world data. A good step towards mitigating errors would be to use the AI outputs only as a guiding principle, where a human with proper expertise can still decide differently to an algorithmic recommendation. However, as reported by STAT, UnitedHealth employees were directed to keep patients rehab stays within 1% of the rehabilitation days projected by the algorithm. Technically speaking the algorithm did what it was supposed to do: it “optimized” care - if we looked at optimization through the insurance cost reduction and profit increase lens. Patients paid the price and now UnitedHealth faces class action lawsuit. Which brings up the natural question: how do we regulate AI so that regulation doesn’t stifle innovation but also doesn’t harm patients?

Share

Regulation TL/DR 2023: work in progress

The pioneering legislative piece - the EU AI Act was published 2 years ago. The premise is simple: Categorize AI application into four categories, the highest level of regulation for those algorithms that could be used for discrimination such as the Chinese social scoring system. But 2021 feels like a distant memory and with the availability of ChatGPT a whole new spectrum of questions arose around AI, and how would it be possible to regulate it. Unlike in machine learning, generative AI by definition generates something new with each prompt, which makes it very difficult to compare outputs and their accuracy. Hence the discussions about the right approach are still open.

In the US, the National Institute of Standards and Technology (NIST) published a Voluntary AI Risk Management Framework, which can act as a guideline to the industry. Then in October, the Biden Administration released and signed an Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. In healthcare, it requires Health and Human Services (HHS), in consultation with the Secretary of Defense and the Secretary of Veterans Affairs, to establish an “HHS AI Task Force” by January 28, 2024. The HHS AI Task Force will have a year to develop a regulatory action plan for predictive and generative AI-enabled technologies in health care.

We’ll check-in in 2025.

Subscribe now

What do the Experts Say

Now that you have the overview, here’s how experts from the industry commented the current state of AI in healthcare.

In a live panel discussion podcast recording in November I spoke with:

• Shweta Maniar, Global Director, Life Sciences Solutions and Strategy for Google Cloud,

• Rachel Dunscombe, CEO of openEHR International Foundation,

• Kira Radinsky is CEO and CTO of Diagnostic Robotics,

• Harvey Castro, MD, MBA., Physician, Speaker, Strategic Advisor for ChatGPT & Healthcare.

You can watch the video recording:

Where Can We See AI in Practice?

… For Personal Use

Medical students are using ChatGPT for education purposes - to create summaries or turn text into mp3 files they can listen on their commutes.

… For Client Management

Diagnostic Robotics uses generative AI to suggest care managers how to respond to a specific patient when they’re trying to convince them in an emphatic way to take preventive care measures. Instead of completely replacing people with bots, which many users still don’t like, people talking to patients can get additional ideas about how else to address individual’s concerns.

… For Precision Medicine

Google's platform is being used by various organizations to create solutions for personalized medicine. For instance, Google introduced the medical imaging suite, an AI-assisted diagnostic tool used by Hologic for cervical cancer diagnosis and by Hackensack Meridian Health for detecting metastasis in prostate cancer patients. In the pharmaceutical sector, Google has launched AI-powered solutions to accelerate drug discovery, including suites for early research, target identification, and multi-omics, integrating AI with other advanced technologies. This focus on precision medicine considers individual variability in genetics, environment, and lifestyle to support personalized patient care, aiming to improve diagnosis, prevention, and treatment plans. Shweta Maniar hopes these technologies will also contribute to greater equity in healthcare.

Promising Times For Value Based Care

Healthcare systems globally are slowly maturing in digitalization. At the same time we are watching mounting worries about escalating healthcare expenses. Consequently, we have no choice but to find inventive strategies to preserve health and manage costs effectively. This is precisely where value-based care comes in, and where AI will pay a crucial role, said Kira Radinsky, CEO and CTO of Diagnostic Robotics. “Value-based care is a move towards paying clinicians based on patient outcomes, in contrast to the fee-for-service model. In other words, if patients are healthy, you make more money. This requires a lot of AI.” Primary care physician have a few thousand assigned patients and it’s impossible to continuously monitor all of them. When patients leave the doctor’s office, doctors get completely disconnected from what is happening with patients. “Here comes AI. How can I allow you to identify patients who are deteriorating so you can start interventions before they become sick,” said Kira Radinsky. To build population health predictive models, Diagnostic Robotics obtained more than 60 billion historical claims, and almost 100 million medical records, to identify patterns of care management for different patients. “After 15 years of only documenting data, I think this is the first two years when you started seeing a shift where physicians will only be incentivized if the patients are healthy and not only based on documentation and services they provide. In other words, the environment is ripe for an AI intervention.”

Leave a comment

Diagnostic Robotics is based in Israel and operates in the US, Israel, Brazil and South Africa. While their American healthcare provider customers still have headaches due to their healthcare data decentralization, and challenges in combining patient data from different sources, countries such as Israel or even South Africa, have centralized data centers with a lot of data and a lot of opportunities to build models on top of that data, she also said.

The Data Question in Generative AI

Good data is the basis for good AI. However, challenges with healthcare data are far from solved. Rachel Dunscombe emphasized we still have a lot of work to do in unifying how we record data. “We've got a data legacy problem, a technical debt that we've built in where we have far too many data models. Blood pressure, for instance, in the UK is recorded in hundreds of different formats. So from my perspective, we need the foundations to be right. Part of this is not to rush, and sure we have the right data foundations,” she said.

Shweta Maniar echoed the need for caution, despite the excitement of the general public and the fact that it can be slow and time-consuming to do AI right. “We need to be slow and methodical. You want to focus on keeping the human in the loop as you're developing this technology and doing proof of concepts,” she said.

ChatGPT, Bard or Claude have become essential tools for many people, helping them draft messages, marketing ideas, strategies, and more. Developers use it to check their code. So what role do data standards play in the works where generative AI can structure data? At this point, especially as assistants. “Generative AI has got a role in being part of the deciding proposing, peer reviewing of data structures for the future. I think that is absolutely logical. It's just going to add power to us getting the data right,“ said Rachel Dunscombe.

Regulation: Food Labels for AI

Among clinicians, the impressions of generative AI are mixed. There’s early adopters and there are clinicians that say they will retire early if these technologies get introduced in their clinical practive, said Harvey Castro. Fear from AI comes from lack of understanding and transparency where AI is used and how. A potential solution could be to add marks sucks as food labels to AI products. “When a patient comes to the doctor, they need to know that AI is being used and then vice versa. When the doctor goes to work, they need to know that the hospital picked the AI and which AI and what's going on. With labels me as a doctor, I can know that this particular AI has this bias, has these issues, and that way we can address it,” said Harvey Castro, expressing hope that larger companies in the AI space will be cautious. Smaller players might not be. “I'm actually more worried about the one off companies that may not really fully understand these bylaws and use some of these APIs to start creating products that really aren't safe or that aren't even regulated by the FDA where they should be regulated by the FDA,” he elaborated.

Rachel Dunscombe also sees food labels a good approach forward. We also need a comprehensive approach to integrate AI into healthcare, driven by user-centered design, and looking at parallels in the analogue world. Who can one turn to, if an AI seems to be off? “We need to know how do you report an issue? How do challenge something? In a human place, you'd ask for the supervisor, the ward manager, whatever else. How do we get that equivalent?” she illustrated, emphasizing that the service design aspects need to be considered everything from the citizen to the clinician, because it's not just about the AI, it’s about reimagining healthcare.

Share

The Future is … Bright!

As 2023 draws to a close, what can we expect in the upcoming years?

1. Better interaction with the healthcare system

Shweta Maniar sees opportunities in healthcare optimisation, with better information flow between patients and healthcare systems. “Humans want greater interaction with their health care system, but it doesn't necessarily mean they want increased human interaction,” she said.

2. Equity

Rachel Dunscombe hopes we will be able to curb the spending curve in healthcare and use large language models to improve equity, to make things multi-language, and to make the advice applicable to people in a specific household or context.

3. Precision medicine

In life-sciences, Shweta Maniar expects further advancements in drug development.“We can have enhanced chatbot interactions for general health questions for first-line triage, but where I see things are going to be going is how this can support advanced research and development to support how researchers are going to identify are the future therapy candidates using different simulations and aggregated population data and using LLMs to mine large trails of research data,” she said.

3. Speed and efficiency

Kira Radinsky is looking at the potential from the clinicians’ side, expecting new AI capabilities to help clinicians make faster but still accurate calls on patient care. “My biggest belief is specifically in the preventive care space where there's a lot of buildup of relationship and obtaining data and then summarizing it to a physician to make a quick call. The more those tools can help make this efficient, the bigger the implications are going to be on the burnout of physicians.”

4. Personalized health

Harvey Castro has high hopes for a move forward to preventative medicine, with personalized approached on an individual level. He believes large language models will play a pivotal role at analyzing individual DNA profiles to identify specific health needs, which will lead to highly personalized medical treatment. This could result in a more targeted approach to medication, where, instead of a standard dose, a person might need a significantly lower dosage of a medication, or perhaps none at all, due to proactive lifestyle adjustments. The overarching vision is a healthcare future focused on preventing illness before it starts.

We have reasons to be optimistic.

Share

Leave a comment