Philippe GERWILL

Challenges of Using generative AI in medicine: 1. Data Quality and Availability: High-quality, large-scale datasets are essential for training generative models, but such data can be scarce, incomplete, or biased in the medical field. 2. Privacy and Security: Handling sensitive patient data requires stringent privacy measures. Ensuring that generative AI models comply with regulations like HIPAA is a significant challenge. 3. Regulatory Compliance: Generative AI applications in medicine must adhere to strict regulatory standards set by authorities such as the FDA, which can slow down development and deployment. 4. Bias and Fairness: Generative models can inadvertently learn and propagate biases in the training data, leading to unfair or inaccurate outcomes. Interpretability and Transparency: Medical professionals must understand how AI models make decisions. Many generative models, like deep learning networks, are often considered "black boxes." 5. Validation and Testing: It is crucial but challenging to ensure that generative AI models are thoroughly validated and tested in diverse, real-world scenarios. Integration with Existing Systems: Integrating generative AI solutions with existing healthcare IT systems (like EHRs) can be complex and costly. 6. Ethical Concerns: Using generative AI in medicine raises various ethical issues, including the potential for misuse, the impact on doctor-patient relationships, and the implications of AI-generated content. 7. User Trust and Acceptance: Building trust among healthcare providers and patients in AI-generated solutions is essential for widespread adoption. 8. Cost: Developing, implementing, and maintaining generative AI solutions can be expensive, particularly for smaller healthcare providers. 9. Scalability: Scaling generative AI solutions across different healthcare settings and geographies can be challenging due to variability in healthcare practices and infrastructure. 10. Responsibility and Accountability: Determining who is responsible and accountable for AI-generated decisions or errors is a complex issue. 11. Training and Expertise: Healthcare professionals need proper training to effectively use and understand generative AI tools. 12. Real-Time Processing: Generative AI models often require substantial computational resources, hindering real-time clinical applications. 13. Generalizability: It is crucial to ensure that generative AI models generalize well to different patient populations and medical conditions. 14. Data Annotation and Labeling: Annotating medical data for training generative models is labor-intensive and requires expert knowledge. 15. Patient Consent: Obtaining informed consent from patients for using their data in training generative AI models can be complicated.

8

Get the Digital Health Transformation, Implementing Artificial Intelligence and Machine Learning in Patient Health Care Systems on https://lnkd.in/dqnzCqZN Discover the future of healthcare with Digital Health Transformation, Implementing AI and Machine Learning in Patient Care, a groundbreaking guide for anyone passionate about the intersection of technology and health. This expertly crafted book jumps into the practical, ethical, and strategic essentials of using artificial intelligence and machine learning to transform patient care from diagnosis and treatment to personalized health experiences. In this comprehensive guide, you will learn not only the "how" but also the "why" behind AI’s integration into health systems, revealing its massive potential to improve outcomes, enhance efficiency, and reduce costs. Whether you're a healthcare professional, data scientist, health informatics specialist, or an entrepreneur implementing digital health solutions, this ebook is designed to equip you with the knowledge, tools, and strategies to succeed in the age of digital transformation. Explore engaging chapters that cover  The fundamentals of AI and ML in healthcare  Case studies showcasing real-world applications  The intricacies of electronic medical records (EMR) and the National Data Repository  Hands-on frameworks for AI implementation  Ethical considerations and patient privacy in the digital health age Packed with actionable insights, strategic frameworks, and expert perspectives, Digital Health Transformation is more than a book it’s your blueprint to leading the future of AI-driven healthcare. Get ready to harness the power of AI for better, smarter, and more inclusive patient care. Perfect for both seasoned professionals and newcomers eager to make an impact, this ebook will transform the way you understand and leverage digital health. Start your journey into the future of healthcare today! Buy now!

16

AGI for Workplaces | CodeNinja Health in Germany | Building Emergency medicine Ai stack When it comes to emergency medicine, doctors and frontline workers have a limited time to attend, diagnose, treat, and manage patients. In cases of serious emergencies, diagnoses often rely on rapid visual analysis that must be performed within 300 to 500 seconds. This analysis leverages frontline health workers experiences to identify patterns, which enables rapid diagnosis. However, there are two major issues. First, accurately diagnosing and deciding on a treatment often takes more time than available, and in emergencies, timing is crucial—missing the window can mean losing a life. Second, an incorrect diagnosis leads to ineffective treatments, potentially resulting in loss of life. CodeNinja is working with some of the finest doctors in Germany to develop an AI Emergency stack designed to assist doctors and frontline health workers in making effective diagnoses under high pressure. We believe this technology can save thousands of lives every day and can be retrofitted in Emergency vehicles, existing Emergency apparatus in hospitals and mobile units. We are excited and grateful that CodeNinja is working on some of the most impactful global technologies with some of the finest minds out there, pushing the state of the art forward. ( Dr. Shahid is the founder of APPS Europe, Global chair of APPS, heads Emergency Department in Manheim and founder of Golden Hours ). CodeNinja Inc. #Goldenhours

38

🌍 The Potential of Data-Driven Innovation in Healthcare Spending As healthcare spending continues to rise, with medicines accounting for 55% of total public spending and big hospital complexes driving up prices, a transition to more sustainable methods is critical. The concept of "budget-conscious innovation," which is commonly used in other industries to improve technology at a low cost, is becoming increasingly important in healthcare. This approach not only seeks to reduce spending, but also to preserve or improve outcomes through technology breakthroughs made by entrepreneurs. 🔗 Gottlieb and Makower (2013) found that while the initial costs of technological innovation may be considerable, the long-term benefits far outweigh these costs, resolving both regulatory and perceptual obstacles along the way. Read the study https://lnkd.in/dhUQVwaP How will data science impact healthcare financing in the near future? High income countries are more aware about its importance, however the need is more pressing in LMICs like Egypt with our high population. Cost reduction involves utilising data analytics to find inefficiencies and high-cost regions. Predictive Models: Using artificial intelligence to predict patient care needs and manage resources more effectively. Personalised Treatments: Using big data to create cost-effective medicines over time. Let's talk about how data-driven innovations may improve healthcare equity and efficiency. What are your opinions on integrating more data-driven initiatives into healthcare systems? #HealthcareInnovation #DataScience #HealthTech #Cost-Effective Innovation

7

3 Kommentare

In healthcare, context and patient input for example are data points that often end up being captured in free text fields, turning into what is called "unstructured data" and being then hard to difficult to interpret and exploit. Natural language processing (NLP) is transforming healthcare by converting unstructured data from EHRs into actionable insights, enhancing patient care beyond traditional methods. This short read explores the possibilities offered by this approach. This is innovation in action! #innovation #healthcare #nlp #ehr #structureddata #unstructureddata #data #patientoutcomes #digitalhealth https://lnkd.in/e7uZ-XdJ

2

1 Kommentar

Trip to Artificially Intelligent Physician: Anatomy of the Clinical Moment I am often asked what happens in digitally enabled consulting rooms between physician, patient and Digital Creatures during a Clinical Moment; which is the point and time when accurate diagnosis is made based on interpretation of biological samples; make critical decisions or actions. Using Roter's Interaction Analysis System (RIAS); Biomedical Clinical Moment is about 8 minutes of 17.4 minutes of patient–physician consultation time. In this artificially intelligent pathology workflow, machine learning algorithms and deep learning computational models; streamline and enhance pathology diagnostic process; and even suggest possible diagnoses. In this workflow, AI systems assist pathologists in analyzing digital images to identify patterns and detect abnormalities like cancerous cells or other disease markers; because AI tools are trained to recognize specific features in tissue samples, such as shape, size and distribution of cells. The Learning Rule or algorithm is able to assist because it has learned patterns in data, like characteristics of human tissue. This is how it works: - 1.  Data Collection: AI models are trained to identify specific tissues on large datasets of labelled tissue images, such as histopathology slides, 2.  Feature Extraction: During training, these tissue features are extracted from tissue images like shape, color, size, texture and pattern of cells. 3.  Pattern Recognition: The learning rule adjusts the parameters of the model (such as weights in a neural network) by repeatedly comparing its predictions to the actual labelled data, using a process called backpropagation. Errors are minimized over time through an optimization algorithm (like stochastic gradient descent), allowing the model to learn distinguishing patterns in tissue images that indicate normal or pathological states. 4. Classification: Once the model has learned these patterns, it can classify new tissue images, identifying abnormalities such as cancerous cells, inflammation, or necrosis, with high accuracy. AI models can detect subtle differences in tissues that might be missed by human eyes, improving diagnostic precision. In Digital Transformation, this AI System can be called a Digital Organism or Creature. It is a Digital Spirit or phantom that helps in identifying patterns, automating repetitive tasks, and reducing human error, and thus improving diagnostic accuracy and consistency. When Learning Rule is supervised, AI System is close to human pathologist, but when its unsupervised, it goes “free-style” in all patient information called “Dark Data”; to detect unknown patterns in tissues, such as clustering similar cell types together without prior labels. When it transfers learning, it allows models trained on large datasets to adapt quickly to new but related tasks, such as identifying new forms of tissue abnormalities. In 8 minutes or less; Clinical Moment is Done

2

1 Kommentar

I‘m proud to join the panel talking about digital twins, the paradigmshift in health data management through cloud and the efforts we need to take to improve the health outcome. The White paper to the paradigmshift, including the building block description you can find here: https://lnkd.in/eXS3UAAT Better data, better insight, better outcome! #Cloud #AWS #digitaltwin #healthdata

40

The growing use of #artificialintelligence in the 21st century is influencing a societal and economic shift towards increased automation, data-driven #decision-making, and the integration of #AI systems into various economic sectors and areas of life, impacting job markets, #healthcare, government, #industry, and #education.

1

😱 How would you feel if you like an AI doctor more than a human one? In a study published last year (https://lnkd.in/ecUrzyRZ ) on 195 randomly drawn patient questions from a social media forum, the chatbot responses were preferred over physician responses and rated significantly higher for both quality and empathy. I am not surprised about this and personally it does not scare me at all. I believe AI should complement human healthcare providers, offering additional support and resources. Human doctors bring unique qualities like intuition, creativity, and the ability to adapt to complex situations that AI might struggle with. Since I moved house in 2021 and registered to a new GP, I never met them in person but only had phone calls or email exchange. Now I wonder if they were real or just an advanced AI bot..... #AI #innovation #GenAI #digitalhealthcare https://lnkd.in/ecUrzyRZ

6

Fascinating use cases here for GenAI in healthcare, WITH companies that are actually executing them today. A great read; thanks for the reshare Archimydes

21

Some personal observations on OpenAI's recently released #ChatCPT4o, frankly speaking, was an astonishing demonstration of the man-machine interface using voice. #1 The Web is moving from a #Search paradigm to an #Answer paradigm, with the unintended consequence of massive disruption to the #Search advertising business models. Google will have competition from not just OpenAI but others, including Apple, Anthropic, Mistral AI, Microsoft and Meta #2 Apple's #Siri is under severe competitive pressure from #ChatGPT4o; it would not surprise me if Apple adopted the underlying model to augment #Siri in #ios18 and future operating systems. While Apple has the compute capability, I don't believe they have a #LLM that can compete with #ChatGPT4o #3 #Alexa by Amazon will need a radical change to architecture to support #GenerativeAI or its deadman walking. #4 Google #Gemini and related #GenAI models from Google have just fallen off the back of the #Peloton; how fast can they rejoin after introducing the #Answer paradigm? #5 Which Web Browser will integrate #ChatGPT4o the fastest to fulfill the #Answer based paradigm? #6 Microsoft needs to reevaluate the #GuardRails on #CoPilot and #Bing; both have reasoning limitations that limit their usefulness. #7 If #ChatGPT4o is an indicator, #ChatGPT5 will have real-time search capabilities, easily configured #Public and #Private Learning Models, and more Model-on-Model capabilities, in other words, using more discrete domain specific fit-for-purpose models that are ensembling in real time. #8 #Compute optimization is happening in terms of token efficiency rather than just tokens per second. This translates to a significantly lower delivery cost per #Answer. The question now begs, when will the compute cost of an #Answer approach the cost of a #Search ? #9 The demand for #AI silicon, whether #Cloud or #Edge based, will be significant, along with a doubling of energy requirements every 12 months for the foreseeable future, as the economic and productivity gains are too significant to be ignored. #10 The right question will now be more important than the knowledge that supports it. #Knowledge is now abundant, good questions not so much. #11 Knowledge source and integrity of that source will now become paramount. Original knowledge source identification, provenance and quality will drive the value of any given #LLM. #12 How many knowledge sources will begin blocking and #paywalling #MachineLearning due to commercial copyright infringement concerns? #13 The Signficant Six #SSIX Alphabet Inc., Amazon, Apple, Meta, Microsoft and NVIDIA are in a game of financial, technical and human capital leapfrog in #AI. Expect mega aquistions of novel innovators to protect positions and extend advantages. In closing, #PandorasBox is open. The future belongs to those who ask the right questions and harness the power of intelligent answers. #AI #ML #LLM #GenAI

22

11 Kommentare

Raindrops are a great metaphor for the #digitaltransformation of Healthcare. What do I mean? The promises of the digital health revolution are many. Decreased workload, improved diagnostics, improved efficiency, greater access, enabling technologies etc. Furthermore, their wide adoption seems imminent! Not only does it make sense, but in every other industry, those same technologies are in place and delivering on their promise. In fact, the digial revolution should hit healthcare like a tsunami and allow us to start over on some version of the medical bay on the Starship Enterprise. It is a cruel experiment: your car drives you to work but when you get there, you are faxing patient referrals. A patient can book an international flight but when they try to book a clinic appointment, the task is insanely complicated. But not everywhere. At Kaiser Permanenet in NCAL, it is faster to book an online appointment than to do it on the phone (I did it). Some patients have the benefit of a sci-fi avatar driven all knowing virtual assistant to help them recover from surgery. A couple of posts back, I talked about the speed at which smaller hospital groups are able to adopt new technology efficiently. Unburdened by legacy tech stacks and outdated leadership models, they can actually identify pain points for which there are mature digital solutions and implement them. Done. Each of these smaller groups, or digitally enabled private practices, represents a rain drop of technological adoption. Over time, adjacent raindrops will coalesce, competition will lead to curiosity, curiosity will lead to familiarity, familiarity to trust, and trust to broad adoption. In the 1990s, many proclaimed that sequencing the human genome would usher in a wave of cancer cures within a decade. Nearly 40 years later, we are starting to see some raindrops. Today, it's AI. We hoped that a computer science tsunami would wipe the healthcare slate clean, but instead, we are getting raindrops of success and people are either surprised or disappointed. We should not be. For each raindrop, in that location, the job is done. Eventually, others join in. The network effect grows. And it grows quickly. So: go ahead, be a raindrop. Lead the pack. Forge the way where it makes sense. You won’t be alone for long. What do you think? Is this a good visual to explain how innovation spreads in healthcare and why it pays to be an innovator? Mohan Nair Logan Plaster Daniel Kraft, MD John Nosta Professor Shafi Ahmed Fabrizio Billi Gil Bashe Mark Froimson MD, MBA Ivan Tornos Todd S. Eury Douglas Fairbanks Peter Birch Ali Hashemi Saul Marquez Michael Havig, MD Rafael J. Grossmann, MD, MSHS, FACS Claudio Díaz Ledezma Daniel Couchman Kendall MPH Sanjeev Agrawal

34

9 Kommentare

Title: The potential for artificial intelligence to transform healthcare: perspectives from international health leaders. We don’t need AGI, just well engineered ANI tools integrated with existing healthcare infrastructure, making it more personalized, precise, innovative, and critically cheaper and accessible to all. Those are real world success factors for solid business use cases. See… https://lnkd.in/eK8sEhEa Abstract: Artificial intelligence (AI) has the potential to transform care delivery by improving health outcomes, patient safety, and the affordability and accessibility of high-quality care. AI will be critical to building an infrastructure capable of caring for an increasingly aging population, utilizing an ever-increasing knowledge of disease and options for precision treatments, and combatting workforce shortages and burnout of medical professionals. However, we are not currently on track to create this future. This is in part because the health data needed to train, test, use, and surveil these tools are generally neither standardized nor accessible. There is also universal concern about the ability to monitor health AI tools for changes in performance as they are implemented in new places, used with diverse populations, and over time as health data may change. The Future of Health (FOH), an international community of senior health care leaders, collaborated with the Duke-Margolis Institute for Health Policy to conduct a literature review, expert convening, and consensus-building exercise around this topic. This commentary summarizes the four priority action areas and recommendations for health care organizations and policymakers across the globe that FOH members identified as important for fully realizing AI’s potential in health care: improving data quality to power AI, building infrastructure to encourage efficient and trustworthy development and evaluations, sharing data for better AI, and providing incentives to accelerate the progress and impact of AI.

2

Healthcare AI Use Cases. PPNCC - A framework to segment the healthcare space when looking for opportunities. If we segment the market into 4 big operational quadrants (Patient facing, Provider facing, Non-Clinical, and Clinical) and look at the big areas of opportunities in each area, it could be easier to wrap our minds around the space. In the video I mention about 40 different startups that are creating disruptive innovations in each of these quadrants. https://lnkd.in/ghuvQMBX

4

The book "How AI Can Democratize Healthcare" by Michael Ferro and Robin Farmanfarmaian explores the transformative role of #AI in reshaping #healthcare delivery. The book discusses various innovative technologies that are central to this transformation, including: #DigitalTherapeutics: This refers to the integration of digital systems with traditional therapeutic practices. It encompasses the use of apps and other digital tools that help patients manage their health conditions alongside prescribed medication, improving outcomes through personalized, accessible interventions. AI Voice Technology and Vocal #Biomarkers: This technology uses features of a patient's voice as diagnostic tools. It's more advanced than typical voice assistants because it analyzes vocal characteristics to monitor health, potentially identifying issues like respiratory conditions early through symptoms like changes in voice or cough frequency. Wearable Tech and Clinical Grade Sensors: These devices, which include FDA-approved EKG monitors and blood pressure sensors, facilitate continuous monitoring of vital signs. Such technologies enable early detection of potential health issues, which is crucial for prompt and effective medical intervention. Continuous Care: The authors envision a future where healthcare is omnipresent, leveraging AI to monitor health parameters around the clock. This continuous monitoring could lead to a proactive healthcare model that focuses on preventing illness before it starts, significantly shifting from the traditional reactive approaches. Democratizing Healthcare through AI: The book posits that AI can make high-quality healthcare #accessible and #affordable for all, regardless of geographic location. This is likened to how major tech companies have democratized information access. AI could similarly #democratize health services, making them available everywhere via smart devices. Impact on the Underdeveloped World: The authors discuss the potential of AI to significantly impact healthcare in underdeveloped regions, where access to medical facilities and professionals is limited. AI could bridge this gap, allowing for effective health monitoring and management using simple technologies like smartphones. Economic Aspects and Sustainability: The book also addresses the financial implications of integrating AI into healthcare. It explores who pays for these advancements and how sustainable models for funding AI and sensor technology in healthcare can be developed to ensure their broad and effective implementation. #digitalhealth #dtx

11

1 Kommentar

Google Health - Is this available to the public? Could someone use this to make a "Personalized Super Symptom Checker" based on their personal EHR and the AI help of Google Bard/Gemini (which was trained in medical concerns) - and maybe to offer this as a chatbot (Patient Listener) sorting through the patient's story, which often the doctor does not have time to listen to? If interested, I made this comment based on the concepts in my book - "Future Healthcare Today: How Technology is Revolutionizing Holistic Wellness” - https://lnkd.in/gzi7se9r Greg Corrado Joëlle Barral #MedGemini #AI #Healthcare #ElectronicHealthRecords #EHRs #diseaseprediction #medicalresearch #MedicineAI #HealthTech #AIResearch #ClinicalReasoning #HealthcareInnovation #AIHealthcare #DeepLearning #ClinicalApplications #MedicalInnovation #AIAssistance #LongContextReasoning #MultimodalReasoning #FutureofHealthcare

3