Enhancing Diagnostic Accuracy with Art Papier of VisualDx
It’s said that 10-20% of all medical diagnoses are wrong. That’s a bit scary, but it’s not surprising. Medical diagnostics is inherently complex, the medical literature is always expanding, and doctors, like all humans, grapple with availability bias – they diagnose diseases they know instead of diseases they don’t. Diagnostic accuracy is clearly a problem.
Today’s guest is Art Papier, MD. Art is the co-founder and CEO of VisualDx. According to Art, there are common diseases and rare diseases, and there are common presentations and rare presentations. The bulk of diagnostic errors are made on common diseases with rare presentations. The goal of VisualDx, a clinical decision support system, is to “augment the brain” of physicians and help them make better diagnoses at the point of care.
On this episode, you’ll learn:
- The importance of being crystal clear on why you exist as a company. What problem are you trying to solve?
- Why and how most diagnosis errors occur with common diseases.
- Why physicians need a guidance system (and why that’s not Google).
- Why diagnosis is NOT “big data”.
- What’s the ETTO principle – Efficiency Thoroughness Trade-Off and how clinical decision support systems can help.
- Why it’s so critical for startups and entrepreneurs to have doctors on their team.
- Why Art favors pull technologies over push technologies when serving doctors. And the problems you have to overcome as a result of using them.
- Why machine learning (ML) is a great way to augment the brain, but will likely never be the full solution.
- How machine learning can be effectively applied to very good data (i.e. 20 years worth).
- Why labeling and terminology are so critical to any successful machine learning application.
- Why general artificial intelligence is a fool’s game and how to stand out with niche and nuance.
- How VisualDx is sharing it’s 20 years of imaging with consumers and combining it with machine learning to help them better understand common skin conditions with Aysa.
- How Apple CoreML allows Aysa to apply machine learning algorithms to consumer images without ever moving the picture off of the consumer’s device.
For full show notes and links: https://thehcbiz.com/116-art-papier-visualdx-diagnostic-accuracy/