IBM Watson Health AI gets access to full health data of 61m Italians

Intro

The race for AI is a race for data. The one that will win it will be the one with the most data. That’s true also when talking about health AI, where companies like IBM and Google are breaking away from other health AI companies, both big and small — because they simply have access to more health data used as training sets for algorithms that translate to faster, better go-to-market.

This post will review the world of health AI and health data trading through an example of a deal done on 2016 between IBM and the Italian government. The deal gave IBM access to vast amounts of health data of 61M Italians. It made IBM a substantial profit and showed how little organizations, like the Italian government, know about health data trading. Beyond just giving an example, I’ll show what this tells us about the growing problem of lack of access to health data for developing health AI and what can be done about it.

Part 1: Economies of scale in data trading, i.e., power = size X focus

In 2016 IBM received access to medical records of 61 million Italians, including demographic data; all medical conditions, diagnoses, and their treatment; emergency and other hospital visits, including dates and times; prescriptions and their costs; genomic data and information about any cancers; and much else besides.

No doubt, being IBM and investing over $6B in IBM Watson — IBM’s AI bet focused currently on healthcare solutions — helps you when bargaining data trading deals. If you are not, though, IBM and do not have so much money to spend, how do you get access to health data to develop a health AI?

The cost of data for to IBM:

IBM committed to invest $150M to open an IBM Watson Health research center in Milan. But the real economics of the deal should take into account other parts of the deal and especially the price of data, as the table below shows.

These figures may seem exaggerated and ridiculous to someone who does not know what data trading is and how it works, especially health-related data. As the CEO of q Data Marketplace, a company that enables organizations to commercially access or share health-related data between them, I know the figures and the dynamics of how data trading actually works.

Summing up the deal: IBM paid $150M but got back ~$610.7B (after we made ridiculous discounts for the cost of data). The bottom line: IBM made a profit on this deal of ~$610.5B, not including future revenues from research and solutions it would commerialize.

Had the Italian government gotten expert advice on health data trading it would have gotten a much higher upside from this deal than just dozens or even several hundreds of jobs for unemployed Italians (the unemployment rate in Italy stands at around 11%). It could have easily gotten many billions of $ for this data and still get the deal done.

The acute problem of access to health data

Access to health data, for example for developing health AI, is a widespread and acute, growing problem. It is widespread because any health AI company is constantly trying to get more data to train its algorithms to make them better not only compared to themselves but also compared to others. Access to health data is an acute, growing problem because there is an acute lack of access to health data globally which is growing with the increase in the amount of organizations developing health AI. Another factor making access to health data an acute, growing problem is the growth of health data itself, through more and new sources of data (think IoT, medical apps, smart medical devices, genomic sequencing, electronic health records etc.). Without access to this data the wonderful world of AI will not be possible.

Without solving the problem of widespread, large scale access to health data the future of AI might look very similar to that of the Search and eCommerce industries: monopolistic, with lack of competition and a continuous decrease in consumer value (or patient value in the case of healthcare). How does a future where a company like IBM or Google (or even Amazon if it makes the right moves) monopolizes the health AI industry look like?

How to create widespread, large scale access to health data?

If you are an health AI, health analytics, pharma or an insurance company you are asking yourself how do you solve the problem of access to data?

Here’s where a data marketplace platform, connecting data buyers and data providers comes into play. At q Data Marketplace we connect companies interested accessing data, like health AI, analytics, Pharma and insurance, to data providers with data about genomics, EHR/EMR, imaging and much more. The data marketplace model is a new one but it’s the only one that makes sense when talking about making data available in a widespread manner and on large scale. There are privacy and security issues, standardization and data quality issues, data pricing issues and many more, but we’ve solved them to deliver a one-stop-shop for data trading — from A to Z, with unique technology that surpasses current solutions (like API) while keeping to regulation (e.g., HIPAA, GDPR).

The more organizations will join the growing network of health data buyers and providers the more value will be created for them and eventually for patients. Keeping a silo approach, where each company and hospital lock and sit on their own data like a dragon on its gold, will not produce the leapfrog we want to see, especially when it comes to AI.

If data is the new oil, the critical resource and key to technological leapfrogging, then let it flood the land.

To read more about this case, enter this link .

Elad Leshem
CEO
q Data Marketplace
www.q-dx.com
International Direct: +972–54–94–84–000
Linkedin Profile