Coming Soon: Doctors As Data Analysts

A healthcare system that is fully data-driven is not that far off, says Kaiser’s John Mattison.

Uber and Airbnb let you rate your driver or your stay. But many users don’t know that your driver or your host can rate you, too. Imagine if you couldn’t get an Uber driver to take you because your rating was too low.

These mutual rating systems provide an example of reciprocal transparency, and it’s coming to health care, predicts John Mattison, chief medical information officer at Kaiser Permanente. Public ratings for your doctor or nurse? Mattison says it is just one of the ways data, especially big data, will change health care. In a speech at the Big Data Innovation Summit in Boston, he ran rapid-fire through a set of the changes data will wreak on health care.

Change #1: A new data environment.

We’re comfortable with living our lives in the natural environment — but Mattison envisions a new environment is on its way for us to live in. He described something called the “plecosystem,” which is a multiplatform ecosystem for our personal data. As he explained it, people put information into the cloud, use smartphones, create social media accounts and profiles — and they increasingly have a “quantified self” through wearable data gatherers like Fitbit and Jawbone. All of these are also platforms that could be brought together to form our personal plecosystem. He also recounted the idea of the “exposome,” which is a collection of all records of our lifetime exposures — educational, artistic, media, social, health, etc. — to create a full record of both nature and nurture, driven by emerging, dense sensor networks.

Data on who we are in combination with what we’re exposed to in our lifetimes can be a pretty powerful foundation for predictive analytics. In medicine, genetic data is often used for risk assessment for chronic disease, but with exposure data layered on top, it reaches a level of precision that is currently unachievable. Mattison says data from the plecosystem will become part of how health care happens, especially as the price of sequencing genomes continues to fall below today’s $1,000 level. “The cost of sequencing the genome is declining faster than Moore’s Law,” Mattison said. That means we are two years from being able to use sequencing cost effectively to tell us who should have scans for things like prostate and breast cancer, and who shouldn’t bother.

By 2020, data gathered “in silico,” as he put it, will be 10 times more plentiful than traditional medical data like lab tests and medical imaging.

Change #2: Virtual humanity.

Mattison also expects that our plecosystem data will be communicated via avatars — virtual representations of the self. “Avatars are going to be the ultimate communication channel and sales and marketing channel,” he says. “People will produce data, companies and institutions will collect that data, and there is no way around it. What will become important is knowing who, besides ourselves, can access our avatars, and what they do with them.”

Imagine having your routine physical without ever leaving your home. Or having your avatar deliver the full set of flu symptoms you’re experiencing to your doctor’s office so you don’t have to get out of bed (since getting rest is probably what you’ll be advised to do anyway!), while your doctor can make a house call without ever leaving the clinic. It’s mind-bending stuff.

Change #3: Corporate data governance makeovers.

While all this sounds like a huge change from the patient–physician relationship standpoint, Mattison says companies have their work cut out for them if they’re to get their data governance house in serious order to deal with what’s coming.

He offered three principles for governance in the plecosystem world:

  • Maximize the completeness and integrity of your metadata — the data that describes the data you keep.
  • Maximize the automation of your metadata tagging.
  • Make sure there is a chief data officer and local data concierges. These are the people who know how the data was collected, and know things about it that might not have been documented.

Governance will be necessary to help people within organizations use the data effectively without becoming overwhelmed by the sheer quantity of data that will be available to them. While doctors won’t want patients streaming data about themselves into their medical files, or entering their own data, Mattison’s view is, “that’s complete and utter nonsense” — it will have to happen, he says. What healthcare organizations, and others, need to do is come up with ways to filter the data, analyze it, and know when something needs to be called out for attention.

Sounds complicated? It is. So here’s John Mattison’s reading list for our Big Data future:

2 Comments On: Coming Soon: Doctors As Data Analysts

  • ANGELA I LIN | November 2, 2014

    Thank you for the blog.
    The volume of the health data is enormous. The doctors know that, and would be quick to acknowledge the importance of the role ” local data concierges”, in both gathering and dispensing data.

    The challenge remains to deliver the care to the individuals who actually need it.

  • Job Too | November 15, 2014

    Doctors in making diagnosis are indeed data analysts as well as end users. The logical progression is that soon we should be able to send data on our illness to a database equally suited to prescribe a treatment solution.

Add a comment