MEM Cast Episode 147: Choosing and Interpreting Diagnostic Tests
7 snips
Jan 27, 2023 
Dr. Nicola Cooper, an acute medical consultant and associate professor, dives into the tricky world of diagnostic tests. She discusses why even experienced healthcare professionals struggle with concepts like sensitivity and specificity. A compelling case study of a smoker illustrates the potential pitfalls of misinterpreting test results. Cooper emphasizes the importance of pre-test probability and statistical constructs in making sound medical decisions, all while unraveling the complexities of defining 'normal' in diagnostics.
AI Snips
Chapters
Books
Transcript
Episode notes
Tests Provide Probabilities, Not Certainties
- Diagnostic tests can give probabilities, not certainties, because how "normal" is defined is statistical and can vary.
- Factors like biological variation and test operating characteristics impact the accuracy of test results.
Consider Critical Difference Value
- Understand the critical difference value, which is how much a lab test result can vary before it reflects a true change.
- Don't overinterpret minor changes; for example, cholesterol can vary by 17% naturally without clinical significance.
Judgment in Uncertainty with Bayes' Theorem
- Doctors must make judgments under uncertainty, interpreting new test information based on initial clinical probability.
- Bayes' theorem underlies how pre-test probability and test accuracy combine to update disease likelihood.
