Atul Butte: Precision medicine makes doctors nervous

The noted physician, software engineer and open science champion said that the hardest part of big data is knowing what questions to ask and finding people capable of figuring that out.
By Tom Sullivan
01:11 PM
Precision medicine

SAN FRANCISCO — The University of California at San Francisco is embarking on an ambitious project to track 15 million patients, map them, predict what will happen in 90 days and what could occur in one year to establish what Atul Butte, MD, described as the new definition of an accountable care organization.

Analytics, big data and ultimately precision medicine will figure prominently into that future state, just not likely overnight.

“Precision medicine makes doctors nervous because if we’re moving into an era of precision medicine that means, by nature, what came before was not precise,” Butte, who is director of UCSF’s Institute for Computational Health Science, said here at the Big Data and Healthcare Analytics Forum.

[Also: Hospitals rank pop health, value-based care, patient experience as top strategic drivers of precision medicine]

Take ICD-9, for instance. The classification system delineated lung cancer as left lung and right lung, Butte said, and he joked that ICD-9 was the definition of imprecise medicine.

“ICD-10, there a lot of codes, all of a sudden we got precise,” Butte said. “But we have a long way to go.”

One area of great progress that Butte pointed to is publicly available data sets of health and medical information, breast cancer specifically. Anyone can access 60,000 samples of data about breast cancer right now, including 2,400 research teams that have conducted experiments and shared the data.

That manner of democratization makes finding these data sets as easy as downloading a song from iTunes, he said, such that even high school kids can access it and create new technologies, such as a California student who programmed an artificial brain one year and wrote an algorithm for diagnosing leukemia the following year.

“When I think of public big data like that, it's retroactive crowdsourcing,” Butte said. “If a high school kid can do that, every scientist is going to have to be able to, as well.”

[Special Report: How real-time big data analytics strategies are improving care quality and efficiency]

And the healthcare industry needs these innovations because — despite all the money in Silicon Valley and other innovation hotspots — there simply are not enough companies to create all the drugs we’re going to need for precision medicine.

Butte predicted that the next wave of big data will be open clinical trials information because if it's de-identified then doctors and researchers can learn from even failed clinical trials.

“Having a lot of data can change the world of medicine the way it has in other fields,” Butte added. But it won’t be easy and the hardest part is finding technologically–savvy and understand data science. “The hardest is knowing what to do with all the data and knowing what questions to ask. There aren’t enough question-askers.”


This article is part of our reporting on the Big Data and Healthcare Analytics Forum, taking place this week in San Francisco. Other stories include: Best practices for data visualization and Wowrack partners with Security First to create cloud security services.  


Twitter: SullyHIT
Email the writer: tom.sullivan@himssmedia.com

Like Healthcare IT News on Facebook and LinkedIn

Want to get more stories like this one? Get daily news updates from Healthcare IT News.
Your subscription has been saved.
Something went wrong. Please try again.