An AI Can't Ask Questions of the Future - Data Trust in Clinical Research

Henrik Nakskov | An AI Can't Ask Questions of the Future - Data Trust in Clinical Research | Pharma Prescribed Henrik Nakskov is a Danish life sciences specialist with more than 20 years in clinical drug development — a medical chemist by training who became one of the field's more original thinkers on data integrity. He spent a decade at Novo Nordisk in haematology and coagulation, held senior roles at NNIT, and led biometrics at Symphogen. That experience now runs through three companies: Sims Consult (his life sciences consultancy), Novyra Life Sciences (where he's CEO and Clinical Information Management Lead), and Safe Implant Technology (where he leads the science). Henrik cares more about where data comes from than almost anyone. In this episode he explains how he borrowed from Second World War supply-chain theory and Toyota's just-in-time management to rebuild the way trial data is collected and cleaned — taking last-patient-out to database release from a normal 8–12 weeks down to one week, and in some studies to two days. He tells the story of trauma trials collecting 1,102 data points in three hours under conditions of war and car accidents, and of mapping the full chain of a stroke patient's care — onset, admission, CT scan, treatment — so precisely the FDA used it to teach what data intelligence actually means. Then the conversation turns to AI, where Henrik is refreshingly precise. He runs Grok and ChatGPT against each other to hunt for the "black holes," builds explicit frameworks and guardrails so the models can't wander, and makes a sharp philosophical point: an AI can only work from references it has already learned — it cannot ask a genuinely new question of the future, so the more you use it, the more surveillance your safety monitoring needs to compensate for the unknowns you've introduced. Coaching over policing, trust over box-ticking, and a continuous-monitoring vision for where quality is heading. A masterclass for anyone who works with clinical data. 🎧 LISTEN TO THE AUDIO Spotify: https://open.spotify.com/episode/40WRQLAD3Ed6ibFV9Fm7wZ?si=QRbYluxTT-2E6O3jeGRWiA  Apple Podcasts: https://podcasts.apple.com/us/podcast/an-ai-cant-ask-questions-of-the-future-data/id1829715970?i=1000792188722  Podbean: https://pharmaprescribed.podbean.com Youtube: https://youtu.be/9ZbJ9PXOJBM  Full transcript and episode notes: https://pharmaprescribed.com/episodes ⏱️ CHAPTERS 0:00 Intro 0:35 Meet Henrik Nakskov 6:00 Faster trials, better quality 12:00 From Novo to consulting 16:00 Building Sims Consult 25:00 Culture and change management 34:00 Trust, standards and due diligence 39:00 Coaching, not policing 41:00 AI in clinical quality 42:00 Frameworks and guardrails 49:00 Why old-school approaches fail 51:00 The black-box risk problem 57:00 A continuous-monitoring vision 1:00 Quickfire and golden rule 👤 ABOUT HENRIK NAKSKOV Medical chemist and clinical data specialist with 20+ years in drug development. Former Project Data Manager at Novo Nordisk (haematology and coagulation), senior roles at NNIT, and biometrics lead at Symphogen. Founder of Sims Consult; CEO and Clinical Information Management Lead at Novyra Life Sciences; CEO and science lead at Safe Implant Technology. Specialist in data integrity, clinical informatics, quality management systems and the governed use of AI in regulated research. Occasional contributor to regulatory guidance. 💼 Search "Henrik Nakskov" on LinkedIn 🎙️ ABOUT PHARMA PRESCRIBED Pharma Prescribed is the podcast where leaders, innovators and hidden voices in healthcare open up. No soundbites, no spin. Just raw insight, one prescription at a time. Hosted by biometrics consultant Adam Walker. 🌐 https://pharmaprescribed.com ▶️ https://www.youtube.com/@pharmaprescribedpodcast 💼 https://www.linkedin.com/in/adamwalkerbiometrics 🏢 https://uk.linkedin.com/company/pharma-prescribed-podcast #HenrikNakskov #DataIntegrity #ClinicalDataManagement #ClinicalTrials #AIinPharma #Biometrics #PharmaPrescribed #ClinicalInformatics #DataIntelligence #QMS #DrugDevelopment