James Faghmous of Mount Sinai Arnhold Institute for Global Health Speaks

Feb 24, 2016    

On Wednesday, James Faghmous, Founder and CTO of the Arnhold Institute for Global Health at Mount Sinai’s Icahn School of Medicine, came to speak at Metis on “Reimagining Global Health”.

Best Data Science Resource

Python. It’s great not just for data science, but general all purpose problem solving

James approaches data science from a pragmatic problem solving perspective. He mentioned that data science has traditionally be great at exposing problems, but less so at solving problems.

For example, today healthcare is characterized by these 4 traits:

  • Episodic: We don’t visit or even think about healthcare if we’re healthy.
  • Centralized: both geographical where you have to go to big cities for better health care, and expert-based where you referred to specialists.
  • Reactive: Treatment and healthcare emphasis only when there’s a problem
  • Disease focused

Instead, James and Mount Sinai are looking to move to being:

  • Continuous: Frequent access to healthcare, and for doctors to keep track of patients
  • Distributed: Better access to
  • Preventive
  • Multi-dimensional: physically, psychologically, hollistically..

“High-Need Patients” are only 10% of US population but they makeup for 50% of the costs in our health care system. If we can use health records more effectively, and in the proper context, we can cut the cost down. Mount Sinai isn’t the only hospital that wants to keep patients out.

Consider a hypothetical patient: Margaret in Harlem. On paper, she has no medical insurance with a poor credit score. She would get almost no healthcare except for critical. But if she has access to a community health worker, we can find out that she’s widowed, late on rent, has not had a doctor visit in years, and fell on ice a few months ago. These are all preventible and addressable issues if they were known beforehand!

So James’ team has an initiative to partner and send community health workers to gather psycho-social information. Then combining this personal data with large scale and clinical data, we can predict the patient’s state to better address their needs.

PERSONAL DATA      LARGE SCALE     CLINICAL DATA      PREDICT Y
-------------      -----------     -------------      ----------
-Social Support +  -Pollution   +  -Health records  = Patient State
-Housing           -Climate        -Preventive Care
-Employment        -Neighborhood

Even in rural areas, community health workers are equipped with cell phones to collect and have access to a data repository to provide even basic healthcare.

It’s really awesome to see people using data science for good.

# metis # data-science # speaker
-->