Tag Archives: machine learning

Medial EarlySign: machine learning for population health (podcast)

Medial EarlySign analyzes standard EHR data to identify individuals at high risk for disease. The company’s first solution, ColonFlag uses longitudinal blood test data to identify patients who are at high risk for colorectal cancer.

I spoke recently with Medial executive Tomer Amit, who filled me in on the company’s approach and explained why the company has been named a Cool Vendor in AI by Gartner.

  • (0:15) What unmet need are you serving?
  • (1:05) You talk about using data that’s already available. What kind of data?
  • (3:02) When you mention “historical data” are you talking about longitudinal data for an individual patient or aggregated data for a population?
  • (4:18) Why is colorectal cancer an initial focus for the company, with your ColonFlag solution?
  • (5:13) Does ColonFlag replace colonoscopy or encourage someone to get one if they have an indicator that they are at greater risk?
  • (6:38) I see how it could help an individual. Would it actually help at the population level?
  • (7:45) You started in Israel and the EU, which have strong longitudinal medical records. Can the approach be applied in the US where that’s not the case?
  • (10:41) You have run your tests in different places around the world. Does the model differ by population or is there a universal algorithm?
  • (12:10) How do you protect your intellectual property? Once you are out there, are there just rules of thumb people can use instead of working with you?
  • (13:14) What traction have you gained with customers or partners? What industry recognition have you received?
  • (14:45) Are there other domains you are investigating beyond colon cancer? Other data beyond blood tests?
  • (15:58) What’s your 5-10 year vision of what’s possible and what Medial’s role will be?

By healthcare business consultant David E. Williams, president of Health Business Group.