Andrew Petrosoniak

Worth a look to see how HHS Innovation through CREATEHealth.ai is shaping the ethical use of AI in clinical practice, congrats Jeremy Petch, PhD!

33

How can we ensure that health technology assessment (HTA) promotes equity in healthcare, and why is it important to integrate equity considerations into the decision-making process of HTAs? These are pivotal questions addressed in a recent report, "Equity-Focused Health Technology Assessment at CADTH," by Renata Axler, PhD, Amil Reddy, CCIP, Farah Husein and Nicole Mittmann at Canada's Drug Agency. https://lnkd.in/gR5zuRmY The report emphasizes the importance of embedding equity into HTA processes to address the persistent and growing health inequities in Canada. It highlights how socioeconomic status, race, ethnicity, geography, and disability can impact health outcomes and calls for systemic changes to avoid exacerbating these disparities. The authors advocate for several specific strategies to integrate equity into HTAs. These include using tools like the Equity Checklist for Health Technology Assessment (ECHTA) and the PROGRESS-Plus framework to identify and address inequities from the scoping phase through to implementation. They also recommend inclusive literature searches that capture diverse and often underrepresented populations and economic evaluations that consider the impacts on different subgroups. By ensuring patient engagement processes reflect the voices of marginalized groups, and by incorporating equity considerations into all stages of HTA, CDA aims to create more equitable health systems. This approach is not only ethically important but also valuable for making informed, fair, and effective healthcare decisions that benefit all Canadians. #HealthEquity #HTA #HealthcareInnovation #PublicHealth #CDA #EquityInHealthcare #PatientEngagement #SocioeconomicDisparities #Inclusion #Diversity #HealthcareAccess #HealthTechnology #EvidenceBasedMedicine #HealthPolicy #SocialDeterminantsOfHealth

39

1 Comment

Emerging evidence from St. Michael’s Hospital in Toronto demonstrates a real-world use case of the potential of Artificial Intelligence to transform healthcare delivery. Their AI system, Chartwatch, has reduced unexpected hospital deaths by 26%, showcasing how AI can complement clinical decision-making and improve patient outcomes. Working in the background alongside clinical teams, the tool monitors changes in patients’ medical records and makes dynamic predictions every hour about whether that patient is improving or is likely to deteriorate in the future. The path forward lies in strategically and responsibly deploying AI solutions that integrate these insights into the workflow of our complex healthcare systems to drive impactful, data-driven, timely and proactive interventions that improve health outcomes and reduce complications across all populations. Sezin Palmer Aloha McBride Ziv Yaar Kim (Ramko) Dalla Torre Kristin Valente #HealthEquity #AIInHealthcare #DataDrivenCare https://lnkd.in/ez-zQ3KR

35

5 Comments

During my keynote at University of Ottawa last week I discussed how as our healthcare system becomes more complex, simulation will be more important than ever. Why? "Hoping" or "guessing" that a new process, system or space will work is increasingly difficult. The accuracy in predicting how a new process will perform in a complex environment is low. When complexity is high, run a simulation instead of guessing. This allows the team to expose the unknown unknowns before all the resources are allocated. Iteration is far easier and the upfront investment in time will pay dividends by gaining a greater understanding of how the new way will function.

48

1 Comment

On the expert panel on healthcare and AI at the University of British Columbia's Roger W. Gale Symposium (https://lnkd.in/grzjsvd6), my knowledgeable colleagues shared some penetrating insights: Charlene Ronquillo pointed out that when the view of the role of a nurse is improperly reduced to only tasks, it becomes more likely to be a candidate role replaced by AI. This brings up the importance of clearly understanding the full scope of roles within healthcare when thinking of holistic applications of AI to care.   Julie Robillard shared insights from her research with robots and an example of where AI becomes more approachable than human doctors: at the Children’s hospital where she conducts her research, in some instances the companionable robots exude an approachability where kids may, at times, share more about their symptoms and experience with the robots than their human care team.  Susan (Feng) Lu brought up the difficulty in parsing out the source of bias in AI. More specifically when it’s bias in the training data set and/or the algorithm. Additionally, revisions to the algorithm for the purpose of targeted reduction in bias can be difficult as well depending on the algorithm and identification of the source of bias. When asked if AI will mean there is no longer a provider shortage in Healthcare. I shared that I ask myself in the context of market equilibrium questions, are we optimizing for there to be enough doctors for how sick we are or are we optimizing for being healthy? Reframing our hopes and fears of AI in healthcare around how adopting AI can make us healthier under our direction and that of our care team. While we won’t be able to stop or slow AI adoption in healthcare, we can drive for democratization of our own data and be knowledgeable consumers in the drivers seat of our own healthcare data and the application of how our data is powering many of the AI models going to market today. We can benefit greatly from what AI can do for us as a collaborator under our direction. More of my comments here(https://lnkd.in/g86XWSVy).

35

2 Comments

"Connecting Care across BC with the First-Of-Its-Kind Procurement of a Digital Referrals and Orders Solution" Look forward to presenting in one week at e-Health Conference and Tradeshow with Elise De Jager MA, BScN & Nancy Nguyen "Usability Studies played a crucial role in the first-of-its kind joint provincial procurement between PHSA, PHC’s New St. Paul's Hospital Project, and Doctors of BC. This team led a request for proposal (RFP) for a digital referrals and orders solution with aims of improving patient journeys and connecting providers." Provincial Health Services Authority Doctors of BC Providence Health Care #ehealth2024 #digitalhealth #digitaltransformation

17

During the pandemic some policy decisions were made on flawed studies- flawed in design, analysis or interpretation of the findings We wrote to NEJM saying the analysis on the Boston school mask study was flawed - we argued (in 185 words) that confounders, especially higher exposure to infection could explain the findings https://lnkd.in/gCC3wYQ9 Reanalysis of the data over a longer period confirms this with the authors stating: “methodology inappropriate…including prior immunity, temporal variation in rates of infection, & changes in testing” https://lnkd.in/gTRSZXbK It’s important that journals rectify wrongs The conclusions from the original paper request that masking Black children will help improve education outcomes. In contrast, masks in already disadvantaged children has the potential to make learning outcomes worse. Policies should not be made on flawed analysis, particularly when there is the potential to do harm. For these reasons, the original paper should be retracted. Journals need to rectify wrongs and ensure ongoing trust in peer review and the scientific process.

14

Learn about CHARTWatch, the AI early warning system that reduces the risk of unexpected death in hospitalized patients. Find out how it... https://ow.ly/n2mH50TFEji #AIInHealthcare #CHARTWatchStudy #EarlyWarningSystemAI #HospitalPatientSafety

2

Opal Patient-In-The-Loop data platform Impact Report is out: Patient access to their own #health #data in near real time. For many of us, the COVID pandemic brought many hardships but also revealed unexpected opportunities to rise to the challenge and contribute to create impact in any way we could. For me, this started with the Quebec government’s Fonds d’accélération des collaborations en santé (#FACS) program, powered by CQDM. It was an honour to serve as #digital #health expert on the FACS selection committee in 2020-2021 and to recommend the Quebec SmartCare Consortium (QSCC) Project for funding, together with committee members Brian Underdown, Dr, Michèle de Guise MD, Bernard Massie and Benjamin Haibe-Kains. Thank you Daniel Hétu (Lumira Ventures) for this chance. Since, as Chair of the Research Oversight Committee (ROC) along with fellow ROC members Shehroz Khan, Mark David Lim, Ph.D. PMP and Katie Roposa, we saw the complex execution of such a large scale endeavour. CQDM's stellar staff deftly managed *all* FACS projects and ROC: Diane Gosselin, Véronique Dugas, Ph.D., MBA, Fanny Guimont-Desrochers, PhD, Rebeca Baca-Diaz. Above all, it has been a true privilege to get to know the QSSC and Opal teams, their visionary leader John Kildea, Silvana Loffredo's dedication, Susie Judd's expertise, their exemplary #patient #partnership at the core of all design, "getting stuff done". Now. This impact report was prepared at the request of the ROC of the FACS, which funded Opal’s development and operations over the last 3 years via QSSC. The Opal Health Informatics Group l #OHIG (OHIG) continues to design, develop and operate Opal at RI-MUHC | Research Institute of the MUHC | #rimuhc, and is releasing it under an #OpenSource license. Patient Voices: 7,000 users ~1,500 patients answered surveys, participated in interviews and focus groups, or provided unsolicited feedback. FIVE common themes emerged: 1. Patients love Opal : Patients overwhelmingly feel that it has improved their experience at Centre universitaire de santé McGill | McGill University Health Centre, would highly recommend it to other patients, and want it to continue. 2. Real-time test results are essential: reduces patient anxiety, brings tangible value to both patients and the healthcare system. 3. Patients want all their data in Opal: Patients love that they can easily access some of their data in Opal, are keen to access more. 4. Opal can engage patients further: Patients willing to report symptoms, engage more using Opal. 5. Reading clinical notes is empowering: Patients who read their clinical notes in Opal report improved understanding, increased trust in their clinicians, more involvement in their care. See data for yourselves: dispels many myths about patients' desire to access and ability to use their own health data. #DigitalHealth #ClinicalNotes #LabResults #PREMs #PatientExperience #HealthData #Cancer #Pediatric #IBS #Nephrology #HealthInnovation

21

4 Comments

AI offers tremendous potential but also serious risks. Here's one application that is saving lives - reducing patient mortality by 20% at St. Michael's Hospital in Toronto: "CHARTwatch is a revolutionary early warning system that monitors patients to predict deterioration within 48 hours. It analyzes over 100 variables hourly, categorizing patients as low, medium, or high risk. If a patient is high risk, CHARTwatch alerts the medical team for prompt intervention, reducing ICU admissions and death rates. Implemented at St. Michael’s Hospital, it has shown a remarkable 20% decrease in patient mortality." https://lnkd.in/eScYrmKS

4

2 Comments

👉🏼 ChatENT: Augmented Large Language Model for Expert Knowledge Retrieval in Otolaryngology-Head and Neck Surgery 🤓 Cai Long 👇🏻 https://lnkd.in/e7_3F-hi 🔍 Focus on data insights: - ChatENT outperformed ChatGPT4.0 in both the Canadian Royal College OHNS sample examination questions challenge and the US board practice questions challenge. - A significant error reduction of 58.4% and 26.0% was observed, respectively. - ChatENT demonstrated enhanced performance in analyzing and interpreting OHNS information, generating fewer hallucinations and showing greater consistency. 💡 Main outcomes and implications: - ChatENT is the first specialty-specific knowledge retrieval artificial intelligence in the medical field utilizing the latest LLM. - Promising applications in medical education, patient education, and clinical decision support are highlighted. - Overcoming limitations of existing LLMs, ChatENT signals a future of more precise, safe, and user-friendly AI applications in OHNS and other medical fields. 📚 Field significance: - Artificial Intelligence - Medical Education - Clinical Decision Support 🗄️: [#Otolaryngology #ArtificialIntelligence #MedicalEducation #ClinicalDecisionSupport]

1

I attended a webinar last week sponsored by Medtech Canada on the topic of Canada’s new proposed Bill C-72, An Act respecting the #interoperability of #healthinformationtechnology and to prohibit #datablocking by health information technology vendors, otherwise know as "Connected Care for Canadians Act", which has yet to be passed into law. It is a laudable endeavor and long overdue, however I am concerned about their definition of health information technology that is currently defined as; “includes hardware, software, integrated technologies, intellectual property and upgrades that are designed for creating, maintaining, accessing, using or exchanging electronic health information or that support such activities.‍” Does this definition of “healthcare technology” include physiological/hemodynamic monitors used in the emergency department, #operatingroom and #criticalcare areas, simple vital signs monitors, #ventilators, #infusionpumps, dialysis machines, whose essential performance and intended purpose does not include what one would normally consider to be a health information technology device, but that usually does produce some sort of data output? The newly published critical care ventilator standard from ISO/TC121, ISO/IEC 60601-2-12, does require a “functional connection that permits data transmission from the ventilator to an electronic health record.” There is a wonderful discussion on the requirements in that standard. These devices not only produce and send discrete data, such as a blood pressure or a pulse rate at a single point of time, they are also capable of sending continuous high fidelity waveform data. Is this type of continuous data included in this bill? Although the definition of healthcare technology includes “hardware, software…that are designed for creating…electronic health information,” do the above-mentioned devices and other similar devices fit these criteria? I certainly hope they do, but I am somewhat skeptical whether they will be interpreted in this manner. It is my hope that there is a clarification in the final version of this bill, that specifically states that it includes all diagnostic and therapeutic medical devices, and software as a medical device. I also hope that other countries, if not done so already, pass similar laws. #interoperability, #healthinformatics, #medicaldevices, #HealthCanada

7

1 Comment

Excellent, and thought-provoking, article on #healthcare in #Canada by Andrew Coyne. -> I do, however, think that patient cost-sharing (with protections for vulnerable populations) has an important role to play - yes, even within a universal healthcare framework. -> Our 2022 study found that [a]mong 28 OECD countries with universal health-care, Canada is only one of 6 that does not use any form of cost-sharing for core medical services, which can be used to incentivize more efficient use of scarce health resources and potentially reduce wait times. -> Notably, 8 of the 10 top-performing high-income universal health-care countries have some form of a cost-sharing arrangement - including Australia, the Netherlands, and Switzerland. https://lnkd.in/gv_5yCxq https://lnkd.in/gUNVdBEh

6

1 Comment

Excited to share my first experience in public health research! Led by Joshua Tan Kuan, the team focused on predicting future healthcare utilization by exploring the performance of various machine learning models and oversampling techniques, leveraging extensive healthcare data. We also examined model explainability, which provided deeper insights into this important area. And this is just the beginning—there’s more to come as we incorporate more data and advanced analytics techniques!

22

Such impactful topics were discussed at the Data Effects #Vancouver including AI in medical diagnostics, data interoperability, and the transformative potential of innovation in healthcare. For example, just like how my dog, Milo, can intuitively sense when something is wrong, AI can now predict health issues before they become critical. It's fascinating to see parallels between natural intuition and technology. Thank you CityAge for such a great event! Alexandra T. Greenhill, MD Fraser Health Authority David Castle Federica Di Palma Martin Wright Danielle Gifford Anita Palepu Victoria Lee Miro Cernetig Harvey Hawes Sam Chung Ray Simkus #AI #QuantumComputing #Innovation #HealthTech #DataScience #HealthcareInnovation #TechForGood #FutureOfHealth #SmartHealthcare #DigitalHealth #AIHealthcare #QuantumTech #TechInnovation #ArtificialIntelligence #HealthData #NextGenTech #MedTech #TechConference #Vancouver2024 #InnovativeSolutions #HealthcareAI Canadian Innovation Network #CanadianInnovationNetwork #CityAge Sapius AI Myant

81

6 Comments

Canadian physicians are embracing digital health tools. This is a key part of reducing admin burden, letting physicians focus on what matters most: providing high-quality care. It’s important we prioritize the work to ensure these tools truly enhance care delivery. We can maximize their potential through better health system integration.

15

1 Comment