More or Less: Behind the Stats The digital ‘robots’ unlocking medical data
7 snips
Feb 17, 2024 Ben Goldacre, Director of the Bennett Institute for Applied Data Science at Oxford, explains the use of OpenSAFELY to unlock medical data. They discuss the balance between research benefits and privacy concerns with large medical datasets, pseudonomization techniques, and the importance of accessing health records for analysis, including insights from COVID-19 research efforts.
AI Snips
Chapters
Transcript
Episode notes
De-anonymizing Data
- Tim Harford and Ben Goldacre discuss how easily patient data can be de-anonymized.
- They use Tony Blair's medical history as an example, demonstrating how public information can identify individuals within datasets.
Medical Data Dilemma
- Medical data sets offer valuable insights but pose significant privacy risks.
- The COVID-19 pandemic highlighted the urgent need for researchers to access these records while protecting patient confidentiality.
OpenSAFELY System
- Ben Goldacre's OpenSAFELY system lets researchers analyze sensitive medical data without direct access.
- Researchers write code on dummy data, and an automated system runs it against real records, returning only the results.