OntarioMD Digital Health Virtual Symposium – Keynote: “Rise Up: Hope for Technology in Practice”

 

I had the privilege of being asked to speak at the OntarioMD Digital Health Virtual Conference last week.  There were almost 600 people in attendance, but if you could not be, here is a written transcript of my talk and the questions that followed!   Longer than my usual posts, but hopefully worth the long read!

Keynote: “Rise Up: Hope for Technology in Practice”

INTRODUCTION (Host):  Dr. Chandi Chandrasena

For over 35 years, Darren has been dedicated to comprehensive care. He has been a family physician and an influential leader in the health system, transformational in digital health across Canada. His work spans provincial and national initiatives with significant contributions to health IT, physician leadership, primary care reform, and quality improvement.

He has advised and contributed to numerous commissions and agencies, including the Ontario Ministry of Health, OntarioMD, Canada Health Infoway, Ontario Health, and the Ontario and Canadian Medical Associations. He is the former Chair of the Cancer Quality Council of Ontario and currently serves as the Chair of the Board for the Canadian Partnership Against Cancer.

In addition to clinical leadership, Darren has held executive roles in telehealth and now provides strategic guidance through KPMG Healthcare Consulting. His expertise lies in modernizing primary care, advancing health data analytics, applying AI in clinical settings, and scaling digital technology adoption. He continues to practice part-time in urgent care in Collingwood, Ontario.

A passionate educator with faculty appointments in the Department of Family and Community Medicine and the Institute of Health Policy, Management and Evaluation at the University of Toronto. You may also remember him as my predecessor — the former Chief Medical Officer at OntarioMD. Please join me in welcoming Dr. Darren Larsen. He is going to speak about Rise Up: Hope for Technology in Practice. Welcome, Darren.

KEYNOTE:  Dr. Darren Larsen

Thanks, Chandi. Really excited to be here and to be part of this fun day. OntarioMD is close to my heart — this is where I started my journey in digital health. My journey is not over yet, but it has been really rewarding to see OntarioMD take hold of so much of the new change happening in digital health across this province, and by proxy across Canada, through your EMR program.

I want to welcome everybody to this session today. I know there are over 500 people in this virtual room, which is amazing — I believe the biggest virtual conference ever for OntarioMD, which is fantastic. So welcome to all of you.

I am really here today to inspire you and bring a call to action about what you might be able to do within your own practice, but with a more strategic view around what is happening across the globe. We have 45 minutes together.

DISCLOSURES

As we start, I want to disclose where I come from. Chandi has already noted some of my work, but currently: I work through the Department of Family Medicine at the University of Toronto, as well as the Institute for Health Policy Management and Evaluation at the Dalla Lana School of Public Health at U of T. I chair the federal cancer agency, the Canadian Partnership Against Cancer. I do work as a Medical Advisor at KPMG Healthcare Consulting and do some small consulting work for startups and groups that matter to me — like the Centre for Effective Practice, Nia Health, and an Edmonton-based biomarker testing company.

I was offered an honorarium for today’s talk, but I have chosen to donate that to MSF, one of my favourite charities. Thanks very much to OntarioMD for allowing me to do that.

SHAMELESS PODCAST & BLOG PROMOTION

A little shameless promotion: this 45-minute session is not enough time to cover everything I want to share today. About a month ago, Chandi and I had the opportunity to record a podcast in a series hosted by Anthony de Valentino — a previous executive director of a large family health team, now running a digital media company called Rhythmic. We are part of a series called Breaking Silos, where we explore how we break down silos in healthcare. This podcast, which you can find on YouTube, goes deeper into many of the concepts I want to discuss today. Chandi and I had the chance to debate the pros and cons of our various views. If you get a chance, look it up on YouTube.

I also have a blog called Ideas for Healthcare where I go much deeper into these concepts in writing. It is just my name, dot com if you want to visit later.

TODAY’S AGENDA

We are going to cover several strategic areas today. I want to look more globally at how we are using technology and AI beyond our practices. I want to talk specifically about what government needs to do to lead with courage, not caution. I want to think about the institutions around us — our medical associations, our colleges, our universities — where we are training new medical students who will become doctors completely trained in the digital age. And I want to speak to technology companies about what we need from them as clinicians.

As a family doctor working in urgent care in a mid-sized Ontario town, I have very specific needs. So do those in inner-city urban practices, suburban practices, or large academic family health teams. One of the advantages of my career history is I have worked in many of these environments over time — rural Alberta, an academic family health team in downtown Toronto, a suburban practice in the GTA, and now in Collingwood, Ontario.

We will spend a significant portion of time discussing what we as clinicians can and should do to help drive this transformation in our practices and make our work lives better. Some of this is practical. I won’t name many specific companies — I think it is important that you do your own due diligence and find the right fit for you. You have help now: OntarioMD is helping, there is meaningful progress on a vendor of record for scribes, and we will see more tools roll out. But I want to give you a framework today — with actionable items — for how to think about technology from a practical practice perspective.

THE DISCONNECT BETWEEN AI’S POTENTIAL AND OUR REALITY

I want to paint a picture of the disconnect we are living in right now with technology and AI. In the real world, AI can write poetry, compose music, and generate realistic video. It passes medical licensing exams. It analyzes massive, complex data sets and finds signals hidden in the noise. AI can do remarkable things.

And yet in our practices, I still cannot find my patient’s chart if they saw another doctor outside my network. We still have deeply disconnected data silos. We are largely delivering clinical care the same way we always have — which is effective and relatively efficient — but essentially unchanged from two decades ago, despite massive shifts in the technology landscape. When we moved robustly to EMRs, we largely digitized analogue processes. We did not transform them.

Healthcare has always lagged behind industry in adopting technology, for good reasons. We are trained to be sceptics. The stakes in individual patient care are high. We think carefully about the pros and cons of any new practice change. That caution is appropriate. But I am now wondering whether the pace of technology change has become so fast that we have lost the ability to keep up.

IT’S TIME TO BE BRAVE

This is the core message I have for you today: it is time to be brave. We do not want to be reckless, but we do want to be brave enough to try new things, take calculated risks, and ask ourselves — what does this technology actually help us with? Or,what might make things more difficult in our practice?

Chandi and I debated this in our podcast. We keep adding layers and layers of technology onto the care we deliver, but when is someone going to take something away? If we keep layering, we are creating more complexity. These technologies need to replace and improve upon what came before — not simply add on. We need to make sure our workflows are genuinely improving, and that our time spent is becoming more efficient rather than just adding more clicks.

We spent 20 years digitizing healthcare with our EMRs — prescriptions went electronic, labs became accessible online. This was necessary work. But the next chapter, I believe, is about humanizing the digital space: reducing the cognitive load on clinicians, reconnecting us with patients, freeing up time to care. The humanization piece is about making technology work for us, rather than us working for the technology.

A PERSONAL STORY: WHEN THE SYSTEM FAILS

Let me share a real-life example. This past December, my daughter unfortunately jumped and landed on her heel, sustaining a trimalleolar fracture requiring significant hardware and surgery to repair. She ended up in the emergency department of a large downtown teaching hospital — six hours in emerg. Things went as smoothly as they could: seen by the resident, x-rays done, CT scan done, fracture reduced, placed in a cast, and told, “We’re going to call you. You need surgery within two weeks.”

Three to four months later, we had still not received a call. Seven weeks after the initial visit, we had received two bills for the cast. No call, no follow-up, nothing.

Here is the thing: this is a young adult with two physician parents. We could activate our network and make sure she got the right care. But what about our patients who do not have that privilege? Who do not know how to navigate the system, or who do not have the ability to access it easily?

We need to be paying attention to those patients. And much of the burden of that advocacy falls on clinicians. The cognitive load and moral injury of constantly checking, chasing referrals, calling providers, begging for specialist appointments — this is what drives exhaustion and burnout in primary care. We just need things to work. And the question I keep asking the system is: is anyone really paying attention to this?  Could a smarter technology have made a difference?

AI AS AN ACCELERANT, NOT A DESTINATION

There is a lot of hype around AI right now. AI companies are building large models, producing interesting products, and generating significant investment — but not yet a lot of return. That said, there is true promise.

In medical practice, our first real venture into this space is with ambient scribes. And for those of us who have been using scribes for a while, this is pretty much a no-brainer. I believe this should have been last year’s conversation. There is genuinely no physician office that should be without a scribe today. The benefit is real and immediate.

But when I look back at my EMR, I realize I don’t actually need a lot of artificial intelligence right now. I just need actual intelligence. A lot of what I do every day in clinic is repetitive and automated. I want software that knows my patterns, automates repetitive work, and anticipates what might happen next. That is where the artificial intelligence comes in — but I want to start with the intelligence first.

THE RESIDENT IN THE ROOM: A TEACHING MOMENT

I have real-world experience teaching residents and students at Women’s College Hospital, where I was faculty for 14 years until this past year. I remember watching a new resident sitting in front of a patient, conducting an interview. In 15 minutes with the patient, she spent — I timed this — 13.5 minutes looking at the computer and approximately 90 seconds looking at the patient.

When she came out, I asked, “How did that feel?” She said she felt nervous, unfocused, and uncertain. Meanwhile, looking at the camera in the encounter room, I could see the patient rolling their eyes, on their phone, feeling completely unlistened to.

At that point, I said to the resident: just turn your screen off. Turn it away. Don’t look at it. Lean in, talk to your patient, truly listen, ask informed questions, and come back. She finished the next interview in 12 minutes instead of 15 — faster. She came away with an excellent differential and really useful information. And at the end, she said to me: “This actually felt great.”

This is what we are going to get with ambient scribes. When we are not typing away at a screen, we can truly lean in, listen carefully, pay attention, and analyze what is being said. The technology then creates the document — a structured, working product that feeds multiple downstream workflows. We focus on the patient. The scribe captures the visit.

THE ROLE OF GOVERNMENT: LEAD WITH COURAGE

Governments across Canada are sitting on a historic opportunity. Billions of dollars are flowing into healthcare right now — including significant funding just announced for EMR upgrades and a common data platform for primary care. But public frustration is at an all-time high. We have technology that could be used today, but we do not yet have the regulations and legislation to make it safe in every case, or to set the right boundaries and limits.

The question to government right now is not, “Is transformation possible?” It is, “Do you have the courage to lead it, and can you move fast enough?”

In my consulting work, I have these conversations with governments across various provinces. It comes down to a few things. We have been funding pilots forever. We need to get out of this pattern of small initiatives evaluated over three years, producing a report that sits on a shelf — and is often already outdated by the time it is published because technology has moved on. We have to switch to rapid evaluation cycles: 60 to 90 days, real-world criteria for scaling, clear thinking about rollout and change management, and ensuring clinicians actually adopt what is being funded.

If we learned anything from the PrescribeIT story, it is that significant public funding can be invested in a product over a long period of time with minimal clinician adoption. That is a problem for government, and it is a problem for us as clinicians trying to do the right thing.

We now have a federal minister, Evan Solomon, responsible for thinking about AI, digital solutions and innovation. Some provinces have a local equivalent.  We have interoperability legislation coming (Bill S5).  My point is simple: every month that we delay technological practice improvement is another month we are failing to address physician burnout.

INTEROPERABILITY: THE LONG-STANDING GAP

We have been talking about data interoperability at OntarioMD, Infoway, and elsewhere for decades. The delay, in my view, is relatively simple to diagnose: there are no consequences for vendors who do not play well with others, and no incentives for them to do this work.

And even as clinicians, when we talk about patient data flowing — flowing to where? We do not yet have a place where data must move to improve care. It does not follow a patient to hospital. It does not follow them into long-term care. Real-time data flow remains largely a pipe dream. We need to fix this.

Certification must match interoperability as a core requirement. We need standards, not just checkbox compliance. When a vendor says they use HL7-FHIR, I want to know how they are using it and what clinical barriers it removes. We need functional certification — not just technical. If we adopt this approach, the technical barriers will disappear relatively quickly. The workflow barriers may persist, and those are the ones that actually matter.

THE ROLE OF ASSOCIATIONS, COLLEGES, AND UNIVERSITIES

Medical associations, regulatory colleges, and universities have a unique role in the Canadian healthcare system. They are trusted by clinicians in a way that governments and vendors are not. With that position comes great responsibility.

Medical associations need to expand their advocacy to include workflow advocacy — not just conversations about how we are paid. Ask “When new technology is being funded and rolled out, is it actually changing the way we work? Is it making things better for clinicians, clinics, and patients?” It is not enough to throw money at a problem and expect it to change. We need to think hard about how change actually takes effect — including how we support clinicians through the learning curve.

Regulatory colleges have a lot to do in shaping standards. For many years, even during practice assessments, colleges would demand PDF printouts from EMRs. When I was at OntarioMD, we made the case that colleges should instead learn how to conduct assessments using the EMR itself. This is evolving, but the pace needs to pick up. Standards for clinicians also need to be permissive — not restrictive — while still maintaining appropriate guardrails.

At the university level, there is an opportunity to convene research, academia, and practice to create new conversations and begin training the healthcare system for what is coming — not just what currently exists. This is an interesting moment: this year’s entering medical school class is the first that will have had their entire medical education in the era of AI. They are using GPT, Claude, and Gemini for decision-making, research, and learning — throughout the entirety of their training. That will change how doctors are made, and we need to prepare our teaching environments for that reality.

We need to teach digital literacy and fluency as a core competency — not an elective. The goal is to ensure that students do not graduate thinking that staring at a screen and doing two hours of charting at 9 PM every night is normal. That has to stop being normal.

WHAT WE NEED FROM TECHNOLOGY COMPANIES

I have been using an EMR for nearly two decades. It was transformational — no greater single change to my practice. But the product I bought in 2007 is essentially the same product I am using today. Very little has changed. Perhaps it has moved from a server under my staircase to the cloud — but after two decades of technological advancement, our clinical systems have largely not kept up. We need to expect more from our technology companies.

I am worried that in some cases, they are solving the wrong problems. We are good at recording data and building dashboards — but those dashboards are largely built for executives and health system leadership, not for clinicians delivering care. We are building interfaces that prioritize data capture over care delivery.

What I actually need as a clinician seeing patients is software that anticipates my next need, learns my patterns, and surfaces the right information at the right moment — creating fewer clicks, not more features. Workflow is the product. Small inefficiencies multiplied by 40 patients a day become a massive inefficiency across a week, a year, a career.

To our technology vendors: before you ship a feature, put it in front of real clinicians in real clinics with real patients. If it does not make their day demonstrably better, go back to the drawing board. And if you are going to add something, take something away. One less click. One fewer page to open. Please, please, please…. make things simpler.

WHAT WE AS CLINICIANS NEED TO DO

We cannot just sit around and wait. We need to be active in this space and stop waiting for permission or for someone else to fix the problem.

We need to think about what we need in our practices, what our workflows look like, and where our pain points are. And if we do not have time to work through that ourselves, ask the people working alongside us. Our office administrators, referral clerks, and secretaries know every pain point in the practice. Many times, we just have not given them the opportunity to tell us.

We are not powerless. We may be tired. There is a significant cognitive load and immense burden in primary care these days. But we have credibility, expertise, and collective influence. This means: show up where decisions are being made. Adopt what works and say so publicly. Physician innovators like Dr. Keith Thompson are using technology in practice, analyzing it, and publicly sharing what works and what does not — that helps vendors improve and helps the rest of us figure out where to invest our energy and time. Speak honestly about what is broken. That is a form of power.

ADOPT WHAT WORKS — STARTING WITH SCRIBES

We resisted scribes for a long time — concerns about privacy, security, workflow disruption. Through OntarioMD’s work on the vendor of record and others, much of that has been addressed. I have now been using a scribe for over 14 months. I have pretty much stopped looking at my computer when a patient is in front of me. I look at the patient. The quality of my encounters has improved dramatically.

I know this because in my networked health environment, patients are surveyed after every visit, and I receive that immediate feedback. The correlation is clear: when I use my scribe, my patient feedback is consistently better. I feel more connected to care. The evidence around these tools is clear. We do not need another study. We need adoption.

When I use my scribe, I talk through the physical exam with my patients — which they really appreciate. Even with children, I narrate what I am finding in a way that involves them. I am teaching my scribe what to record for me. This dramatically changes everyone’s experience of the encounter, and people leave feeling truly heard.

As OntarioMD’s showcase data shows, scribes are saving up to two hours per day. I know this personally — I no longer chart at night. When I leave my clinic at 5:15, I am done. That is 10 hours a week. Over a career, that is years of life returned.

THE FUTURE: AUTOMATING THE INVISIBLE WORK

Scribes are the first step. The next phase is automating the invisible work — the hours spent charting, reading incoming documents, managing the queue. Scribes are already recording and documenting each visit in real time, in structured templates we can adapt. With a click, they generate referral letters. I am regularly creating patient education notes that I email through our patient portal. The product I use is about to enable prescription renewals, and the future includes looking around the edges — surfacing things I might otherwise miss.

The longer arc of this is AI-powered clinical decision support. We are starting to see this emerge in tools like UpToDate, Open Evidence, and Heidi. These integrated clinical decision support features read the chart in real time as I create my note, and will soon be able to surface relevant gaps — a screening exam that has not been done in two years, a medication interaction with something I just prescribed. We are shifting from addressing individual problems to looking after patients more fully.

OUR PATIENTS ARE ALREADY USING AI

Our patients are already using AI before they walk through our doors. They have searched their symptoms on ChatGPT, Gemini, or Perplexity. They have received a list of differentials, researched treatment options, found articles, watched YouTube videos. Some of this is accurate; some is not. They have come in with questions, sometimes with a firm — and occasionally wrong — conclusion they are ready to defend.

So how do I work with this? I now begin by asking: “Did you look anything up about your problem? What do you think is going on?” That opens a conversation where they relax and share. I validate what they found when it is accurate — and much of it is more accurate now than the old Google era of ten companies trying to sell you something before you got to the actual information. When something is off, I gently redirect: “AI doesn’t know your full history. You have context that it doesn’t. Let’s explore this together.”

I want to be the expert they return to — and I can be, if I help them interpret the output of their AI rather than dismissing it. Chandi and I were at a conference this week where it was noted that over 80% of people are now consulting some kind of AI-related product about their health concern before coming to see a physician. This is real. And I think this is an opportunity, not a threat. An AI-engaged patient is a more engaged patient — a better partner in their care.

WHAT AN AI-ASSISTED VISIT LOOKS LIKE

If fully implemented — and this is likely a one to three-year horizon — an AI phone agent or chatbot speaks with your patient before the visit to determine whether they need an appointment at all. It asks smart questions, consults the EMR, and helps decide: does this need to be scheduled? If so, which provider, what type of visit, what modality, and how much time is needed? All before the patient arrives.

Then there is the clinician readiness phase. The AI agent summarizes the chart for me — particularly useful in urgent care, where I am often seeing patients I do not know. Diagnoses, chart notes, labs, flagged risks, care gaps — all surfaced in advance. Instead of reviewing 20 pages of EMR, I am reading a one-page summary.

In the room, the ambient scribe collects information as we speak. It surfaces prompts for things I might have forgotten, flags relevant guidelines, and helps ensure I am doing the right thing for this patient. But AI is only making suggestions — I sign the notes, I am responsible for the care outcome. There is always a human in that loop.

After the visit, the wrap-up is also more efficient. Plain-language summaries for patients are generated automatically. Lab requisitions, diagnostic imaging orders, referrals, follow-up notes — all auto-generated. Staff see clear, prioritized work lists. And as we evolve, chronic disease monitoring through portal-based tools, inbox triage, and automated red-flag detection will make our in-between-visit work far more manageable.

CALL TO ACTION

Our first step is simple: pick one tool. Start with a scribe. I always tell my peers — just use it with five patients and see what happens. Or commit to 30 days and take notes. How do you feel? When are you getting home? How is your interaction with patients? There will be friction, because any change brings friction. But friction means you are paying attention and thinking critically about the product you are using. That is a feature, not a bug.

Then share what you learn. Tell your partners. Talk to your patients about it. Post in your professional networks. Let others see what you are doing so they can learn alongside you. Be an honest broker with the system, with your vendor, and with your colleagues. Use your collective voice to say loudly and clearly what you truly need in practice.

This is the moment where we are changing practice again — in the most pivotal way in the last 20 years. Our patients deserve better. Our clinicians need this kind of technology. And the system cannot sustain itself unless we move and rise along with it.

Thank you all for listening. My kids often say to me: “Will you stop ranting?” And I will say: “But this is an exciting rant!” If you want to know more, visit my website or look up the podcast that Chandi and I did together. I am always happy to speak individually. My contact information is on this slide. Thanks so much for your attention.

Q&A:

Q: How do we move pilots forward and scale them up?

I no longer use the word “pilot.” I want a limited production release — a product that is good enough to put into real use right away, with an immediate conversation about how to scale it in short order. We have taken a very cautious, academic approach to evaluating products. There is a meaningful difference between a deep research-based evaluation and a rapid benefits assessment. We want to fail fast, fail forward: what works, keep it; what does not, change it; and ask your vendors and implementation partners for more in real time.

A lot of the foundational vetting — privacy, security, compliance — is being done by agencies like OntarioMD. That is their job. From that foundation, move quickly. Pick something. If it does not work, pick another. Sometimes a quick insight comes from simply sitting in your waiting room for 10 minutes and listening. What is obviously not working? Start there, analyze it fast, and ramp it up.

Q: How do we address the cost of all these tools and the integration challenge?

As physicians, we are not always great at calculating long-term return on investment — but we should be. There is a price to each of these tools in time, energy, and financial cost. If a tool genuinely improves your efficiency and allows you to see one more patient or go home 30 minutes earlier, that is a measurable return on investment. Ask yourself each week: did I feel any better? Am I happier? Did I go home earlier? Am I less tired? That itself is an ROI.

The good news is that price points are coming down as AI infrastructure becomes cheaper. And when staff time is freed up by automation, it does not always mean reducing headcount — it often means freeing up your team to do the work they have been unable to get to before. Sit back, define what you want from these products, and measure it simply. These do not need to be complicated analytics.

Q: How do physicians get their voices heard at decision-making tables?

My real question to physicians asking this is: are you actually being directed to act, or do you just feel like you are being told? Those are different things. If you feel unheard, the answer is to get into the conversation directly. If there is a committee at your Ontario Health Team, your hospital, or your primary care network — show up for one meeting. That might be all it takes. Have your voice heard, and then go back to your clinic and start a small change.

The three things I always teach in leadership: show up, speak up, and ante-up. That third one means do a bit of the work. This is not always net-new time — it is about feeling in control again. And if you genuinely are being told what to do, connect with the physicians who are at those tables and get your message through them.

Q: How should organizations guide patients on AI health tools?

We cannot evaluate every single tool out there — there are simply too many. What we need is to create clear boundaries and guidelines that physicians, patients, and the community can use to assess anything they encounter. There is a lot of hype out there — Instagram influencers promoting health apps the same way snake oil salesmen once promoted miracle elixirs. The same degree of thoughtful caution applies.

When a patient shows me an app, I ask to see it. In about eight out of ten cases, what they show me is actually quite interesting. The other two, I redirect. But the underlying principle is curiosity. Our patients spend 99.999% of their care journey with themselves, not with us. I need to know what they are doing and thinking between visits. I need to ask.

Our associations and agencies should produce high-level guidance and regulatory frameworks — like Bill S-5, which creates a strategic legal structure for provinces to build on. The more granular guidance can then be built out at the provincial, regional, practice, and individual level. Approach all of this with curiosity and a sense of possibility. If we can find something new to engage with — even something small — it can be genuinely energizing. Things have been stale for a while. We have been tired. Finding something new to grab onto is actually kind of fun.