In 2013, 153 billion GB of healthcare data was collected by devices. In 2020, that number will climb to a whopping 2.3 trillion GB.
But it’s not the data that’s important, it’s the analysis, says Chris Pratt who works with strategic initiatives at IBM Canada.
Speaking at the iTech Vancouver conference, Pratt talked about how the next wave in cloud and cognitive will focus on deriving value from data. Having the data is not the only step — it’s about mining it for insights.
How?
Get the TechPortfolio briefing
The month’s fintech and funding news, plus new tools for founders, in one short email.
Enter: IBM Watson.
From Jeopardy to the Hospital
When IBM Watson debuted in 2011 and won global attention for beating Brad Rutter and Ken Jennings on the game show Jeopardy, much of the hubbub focused on IBM Watson’s ability to accurately answer questions, quickly.
Health care was the first big job IBM picked for it. According to IEEE Spectrum, the day after Watson’s Jeopardy win, IBM announced that it would apply the same language skills to medicine. In 2015 it formed a dedicated Watson Health division, and by mid-2016 that division had bought four health data companies for a total of about $4 billion.
The appeal was easy to see. Most of a patient’s record isn’t neat numbers but written notes, discharge summaries and reports. IEEE Spectrum puts that narrative text at about 80 percent of a typical record, and it’s exactly the kind of material a system built to read language was supposed to handle.
At the time, IBM Watson had one API. Fast forward five years and it now has more than 50, Pratt says. Leveraging external APIs is a fundamental step now for any business that wants to extract value out of its data.
An API, or application programming interface, is a ready-made service that one piece of software can call from another. Instead of building language analysis or image recognition from scratch, a company can send its data to an API and get the results back, which is how a small team can put AI into its own product.
According to a 2016 survey by Accenture, 70% of corporate executives are planning to significantly increase their investments in AI-related technologies compared with two years ago, and 55% are planning to use machine learning and embedded AI solutions extensively.
Why Health Data Is the Big Prize
The healthcare space in particular is ripe with opportunity for data insight-mining. Pratt says the average person is likely to generate more than one million gigabytes of health-related data in a lifetime. That is the equivalent of 300 million books, according to IBM Watson Health.
Pratt says being able to analyze that data is critical to innovation. “Can we afford a health care system that is not augmented [with cognitive technologies]? The answer is no,” he said.
Want to get started with cognitive technologies? Here are the three key ingredients:
#ICYMI at #iTechVan: 3 things @IBM's @crispyrat says you need for #CognitiveComputing:
1⃣ Information
2⃣ Insight
3⃣ Infrastructure pic.twitter.com/R4s5GKunOb— Tech Ed Canada (@TechEdCA) October 6, 2016
What Happened Next
Pratt was right that health data would keep piling up, but Watson’s own record in health care turned out to be mixed. A 2019 IEEE Spectrum investigation, titled “How IBM Watson Overpromised and Underdelivered on AI Health Care,” described how messy real medical records proved for the system. One high-profile project at the University of Texas MD Anderson Cancer Center was cancelled after a 2016 audit found the centre had spent $62 million on it.
In January 2022, IBM agreed to sell the healthcare data and analytics assets of Watson Health to investment firm Francisco Partners, saying it was focusing on its hybrid cloud and AI platform strategy. Those assets now operate as a standalone company called Merative. The Watson name lives on in IBM’s business AI products, including watsonx, a generative AI and data platform IBM introduced in May 2023.
What It Means for Health Tech Builders
Pratt’s core point still holds: data on its own isn’t worth much until someone can analyze it and act on it. What the Watson story adds is that analysis has to fit the way doctors and hospitals actually work, and has to be proved with real patients and messy records, not just impressive demos. For health tech startups, the useful takeaway is to start with a specific clinical problem and clean evidence that you’ve solved it.
Frequently Asked Questions
How much healthcare data were devices expected to collect?
Devices collected 153 billion GB of healthcare data in 2013, and that figure was expected to reach 2.3 trillion GB in 2020. Chris Pratt of IBM Canada said the analysis of that data matters more than the data itself.
How many APIs did IBM Watson have by 2016?
Pratt said Watson had one API when it debuted on Jeopardy in 2011, and more than 50 five years later. An API is a ready-made service that one piece of software can call from another.
How much health data does one person generate?
According to IBM Watson Health, the average person is likely to generate more than one million gigabytes of health-related data in a lifetime, the equivalent of 300 million books.
How did Watson perform in health care after 2016?
Results were mixed. A 2019 IEEE Spectrum investigation said Watson had overpromised in health care, and a project at MD Anderson Cancer Center was cancelled after a 2016 audit found $62 million had been spent on it.
What happened to IBM Watson Health?
In January 2022, IBM agreed to sell Watson Health's healthcare data and analytics assets to Francisco Partners. Those assets now operate as a standalone company called Merative.