AI and Healthcare How good or bad is healthcare data?—with Mika Newton
Apr 23, 2025
Delve into the messy world of healthcare data, where outdated and biased information can lead to poor medical decisions. Discover the concept of digital twins and the risks of using flawed data without consent. Explore the ethical complexities of data ownership and the importance of transparency in healthcare. Learn how empowering patients with data control can lead to better outcomes, while addressing the urgent need for clean, well-governed data in precision medicine.
AI Snips
Chapters
Transcript
Episode notes
Healthcare Data Quality Problem
- Healthcare data quality is generally poor and continuously decays over time.
- Data decay causes outdated and inaccurate AI recommendations in clinical settings.
What Causes Data Decay?
- Data decay includes changes in patient health, personal info, and administrative updates.
- These shifts cause data models to become stale and unreliable without continuous updates.
Four Types of Data Decay
- Four data decay types: temporal, structured, contextual, and semantic decay.
- Each type disrupts data accuracy and applicability in healthcare AI models.