Quai.MD’s Post

We met with Nassib Chamoun, Founder & CEO of Health Data Analytics Institute, to get his take on #AI in healthcare. He hit on two important points: 🔹The realistic role of Large Language Models 🔹AI literacy of the healthcare workforce Check it out. ⬇️ #ViVE2024 #healthcare #healthIT #healthtech #HITsm

And that’s a wrap for the AI Symposium……. The ACDM: Association for Clinical Data Management would like to thank everyone who has joined us for the 3rd Annual Symposium on Artificial Intelligence in Clinical Trials. Whether you joined us as a delegate, or if you were speaking or sponsoring, thank you for making the Symposium a fantastic event. We would like to wish everyone a very safe journey home, and we hope to see you in October for the 2nd Annual ACDM Symposium on Risk-Based Quality Management. ACDM Board: Robert King, Sverre Bengtsson, Jo Marshall, Nina Reyes, Eva Alder, Richard Davies, Nicola Götz, Anita Kratchmarov, Ashley Howard, Amelie Spieser #clinicalresearch #clinicaltrials #clinicaldatamanagement #clinicaloperations #edc #ACDM #AI #artificialintelligence

NEW PAPER out in @BMJ TRIPOD+AI: reporting standards for #artificialintelligence studies developing or validating prediction models in healthcare powered by #machinelearning methods (inc. regression) VIEW PAPER -> https://lnkd.in/etCkc_Gp   #OpenAccess #ArtificialIntelligence #AIstandards #OpenScience #fairnessinAI #PPI #AI #machinelearning #AItransparency #responsibleAI #ethicalAI #AI #regulation #transparency #governance University of Oxford #Transparency is one of the six core principles underpinning the World Health Organization guidance on #ethics & #governance of #ArtificialIntelligence for health. TRIPOD+AI has therefore been developed to provide a framework and set of reporting standards (https://lnkd.in/edMKBtHx) - The ability to understand and critically evaluate the quality of #machinelearning prediction models & gauge their value in a particular setting or for a particular use case is predicated on complete and #transparent reporting…but numerous studies have shown the reporting of #machinelearning based prediction models is poor and incomplete, eg - https://lnkd.in/eK-CgqDz https://lnkd.in/eccnvyXj https://lnkd.in/ez8HV5Y9 https://lnkd.in/e8bniCBm   - Transparent reporting of a #machinelearning prediction model can expose flaws in the design, data collection, or study conduct that, if the model was used, could potentially cause harm to patients or exacerbate inequalities in healthcare provision   - Only when #machinelearning studies are completely & transparently reported can they be trusted, have value to inform #regulatory approval, be included in clinical guidelines & influence health policy -> improve patient outcomes   - Building on the original TRIPOD Statement we developed TRIPOD+AI following the EQUATOR Network guidance for developing reporting guidelines involving an international multi-stakeholder consensus process with 200+ experts and patient & public contributors - TRIPOD+AI has operationalised #fairness values by embedding them throughout the checklist by including reporting recommendations in the Background, Methods, Results, & Discussion sections of a study report.   - With input from our HDR UK #PPI partners/coauthors TRIPOD+AI includes an item on patient & public involvement to prompt authors to provide details on any #patientpublicinvolvement during the design, conduct, reporting (& interpretation), & dissemination of the study   - TRIPOD+AI prominently features #openscience practices with recommendations on #protocols, #registration, #datasharing and #codesharing to promote #transparency, #reproducibility, & collaboration between researchers   - TRIPOD+AI aims to assist #authors in the complete reporting of their study and help eg #peerreviewers, #editors, #policymakers, #regulators, #users, & #patients understand the data, methods, findings & conclusions of #artificialintelligence driven research

Really looking forward to this conversation with Ngan MacDonald and Noland Joiner next week! There are a wide range of both engineering and legal issues on the table related to AI in and for health applications. As usual, I'll be focused on the use cases and the data - the input and output on either side of the models themselves. Join us! Registration is free. https://lnkd.in/g66z-BdM