Nature Portfolio’s Post

𝐀𝐥𝐦𝐨𝐬𝐭 𝐡𝐚𝐥𝐟 𝐨𝐟 𝐅𝐃𝐀-𝐚𝐩𝐩𝐫𝐨𝐯𝐞𝐝 𝐦𝐞𝐝𝐢𝐜𝐚𝐥 𝐝𝐞𝐯𝐢𝐜𝐞𝐬 𝐮𝐬𝐢𝐧𝐠 𝐀𝐈 𝐡𝐚𝐯𝐞 𝐧𝐨𝐭 𝐛𝐞𝐞𝐧 𝐭𝐫𝐚𝐢𝐧𝐞𝐝 𝐰𝐢𝐭𝐡 𝐫𝐞𝐚𝐥 𝐩𝐚𝐭𝐢𝐞𝐧𝐭 𝐝𝐚𝐭𝐚 It strikes me that this is a potentially serious patient safety issue. As one of the key authors (Chouffani El Fassi) of the new paper that reviewed clinical validation data for 500+ medical AI devices and reported on this states: "Although AI device manufacturers boast of the credibility of their technology with FDA authorization, clearance does not mean that the devices have been properly evaluated for clinical effectiveness using real patient data" Clinical validation can use up to three different methods by which researchers and device manufacturers validate the accuracy of their technologies: 1. Retrospective validation - feeding the #AI model image data from the past 2. Prospective validation - uses real-time data from patients 3. Randomised controlled trials - isolating the therapeutic effect of the AI-based solution If phantom or synthetic data was used then it calls into question how that was generated and indeed whether that was appropriate data with which to train/ test the AI solution. There are some serious questions raised here that the FDA needs to answer. A cautionary tale for the #MHRA indeed. #clinicalvalidation #innovation #digitalhealth #data #regulator #patientsafety Dr Mustafa Ghafouri Patient Safety Learning Keith Grimes Dr Annabelle Painter Caroline Chibelushi Dr. Keith Tsui Haris Shuaib 🚀 Manish Patel Mark Ratnarajah Hisham Haq Erwann Le Lannou Kanwal Bhatia Dr Avi Mehra Joe Fennell Nelson Jordan Dr Andrew Jones Dr Cos Fantis Dr Martin Slingsby Janine Jolly The paper published in Nature Medicine can be found here: https://lnkd.in/eEjhs6Pm With thanks to the The Medical Futurist for raising this in his newsletter.

Enhancing Clinical Trial Efficiency with Real-Time Data Adaptive trial methodologies powered by predictive precision medicine allow for real-time data analysis and continuous monitoring. This dynamic approach ensures patient safety, accelerates development timelines, and optimizes resource use. Learn how precision medicine is paving the way for a new era in drug development!  #AI #ClinicalResearch #InnovationInHealth

Patient care, medical research, and clinical trial enrollment could all significantly improve due to AI's incorporation into electronic health records (EHRs). This short communication highlights how AI may improve the recruitment process regarding speed, accuracy, and overall cancer clinical trial efficiency. AI can automate this procedure by utilizing machine learning algorithms, identifying potential trial participants quickly and precisely. - National Center for Biotechnology Information (NCBI) AI is quickly changing healthcare in every way. AI is being used to improve clinical trial outcomes and speed up the development of life-saving medications because of its capacity to increase productivity, improve patient experiences, and uncover vital insights. Check top 7 application of AI in clinical trials in this blog :https://lnkd.in/dGfnRcS5 Ensure Regulatory Compliance in Clinical Trials with Our AI-Driven Risk Management Solutions : https://lnkd.in/dmmx_ExR #AIinHealthcare #ClinicalTrials #DigitalHealth #PatientOutcomes #AIForGood #Healthcare #Innovation #MedTech #MachineLearning #RiskManagement #AIinMedicine #ClinicalResearch #RegTech #HealthTech #EHR #NextGenInvent #AIInClinicalTrials

New Health-LLM model demonstrates the potential to align with or surpass human performance in various medical diagnostic tasks. Leveraging large language models, Health-LLM brings advanced capabilities to medical diagnostics, combining vast medical knowledge databases with patient data to generate accurate predictions. This model indicates a significant move towards LLM-centric diagnostics, where AI can assist clinicians in making more informed decisions, potentially increasing the efficiency and accuracy of patient care across various medical tasks. The promising results of Health-LLM's accuracy and F1 scores suggest a future where AI can support the complex domain of healthcare diagnostics, complementing human expertise.

The EU AI Act has the potential to significantly influence organisations as they navigate compliance while bringing innovation to market and striving for optimal patient outcomes. Interesting perspective from MedTech Europe. #EUAIact #artificialIntelligence #MedtechEurope #medicaldevices #AIregulation #digitalhealthcare

🚨 Senate releases healthcare AI policy roadmap with $32 billion funding proposal. AI in healthcare is getting a boost with a new Senate roadmap. Senators push for funding AI R&D projects, especially in biotechnology and medicine. They emphasize patient safety, transparency, and innovation, urging clear regulations and rigorous testing. Read more on the bipartisan efforts shaping AI policy in healthcare. Story by MobiHealthNews: https://hubs.li/Q02xYs-h0 #AInews #AIinHealthcare #medicalAI

I am looking forward to speaking on Wednesday, June 26th at 10:15 am ET at #IQVIA's Innovation in RWE and Medical Affairs, an IQVIA Virtual Conference.    Join my session with Patrick Hopkinson and IQVIA's George F. Nikolaidis as we explore recent developments and the implementation of JCA in practical terms. With just 6 months to go until oncology and ATMPs will be eligible for JCA, we will discuss the remining uncertainties with scoping, timelines, and evidence requirements and how industry can prepare for all eventualities. #JCA #EUHTA #HTA #HTA regulation

Very excited to be presenting at the IQVIA virtual conference Wednesday to discuss implementation of EU JCA. With just 6 months to go it is critical that companies address the challenges of JCA submission development but also critically embedding JCA into their organizational processes. Join our session with Edel Falla and George F. Nikolaidis as we discuss the key methodological and process redesign challenges that need to be considered to ensure optimal preparation for January 2025. #JCA #EUHTA #HTA #HTA regulation

https://buff.ly/3YeD9VK A recent article highlights the issue of FDA-approved AI devices lacking essential validation data, posing a significant challenge in the realm of clinical evidence for AI medical tools. The insufficiency of requisite validation data in numerous FDA-authorized AI devices has been brought to light, raising concerns about their efficacy and reliability in medical settings. This gap underscores the importance of stringent validation processes to ensure the accuracy and safety of AI technology in healthcare. #FDA #AIValidation

43% percent of the 521 artificial intelligence (AI) health devices authorized by FDA between 1995 and 2022 lack publicly available clinical validation showing they were tested using real patient data, a new study found in Nature last week. The article emphasizes the importance of clinical validation for AI health tools authorized by the FDA to ensure safety and effectiveness. Challenges are related to the lack of standardized language for defining validation methods and the confusion surrounding different types of clinical validation studies. The study conducted a thorough analysis to define clinical validation methods and classify 521 FDA medical AI device authorizations between 1995 and 2022 based on validation methods. The need for increased transparency and publication of clinical validation data is highlighted to enhance trust and credibility in medical AI devices. Analyzing 521 FDA medical AI device authorizations, the study reveals gaps in clinical validation data, underscoring the need for increased transparency. To build trust and credibility in medical AI devices, it's imperative for regulators, industry, and academia to collaborate on prioritizing prospective studies and sharing more validation data. #AIinHealthcare #ClinicalValidation #FDAauthorization #HealthTech #PatientSafety" https://lnkd.in/eJ2S-fW5