watson_health_illustration_17IBM has announced the launch of Watson Health with early partners Apple, Johnson & Johnson, and Medtronic feeding data into this advanced analytics platform, paving the way for analysis of this data on an incredible scale.

IBM’s interest in applying Watson to healthcare is nothing new, going to Watson’s Jeopardy victory over four years ago. We discussed then our perspective on Watson’s applications and limitations in healthcare, a perspective that generated a lot of discussion. Since then, IBM has applied Watson in a number of areas, like in oncology where, in partnership with Sloan-Kettering, Watson is helping guide chemotherapy decisions.

Watson Health represents a dramatic scaling up of the use of this advanced analytics platform in healthcare. According to IBM,

The IBM Watson Health Cloud brings together…clinical, research, and social health information together in a de-identified, HIPAA-enabled data repository. And with Watson Health Insights services, that data can be mined for deeper insights through cognitive and advanced analytics.

Through these partnerships, Watson will get to dive into an incredibly wide range of data. First, Apple will offer up its HealthKit data for analysis. With the recent launch of ResearchKit, users can now consent to the use of their data in healthcare research. And with the remarkable success of the first round of ResearchKit apps in terms of recruitment, the dataset available for analysis will be enormous – perfect for Watson-driven analysis.

Medtronic is focusing on a more specific use case. According to the New York Times, “the company plans to use the Watson software to spot patients trending toward trouble and automatically adjust insulin doses and send alerts to care providers and the patients themselves.” As artificial pancreas systems, combining continuous glucose sensors and insulin pumps, get closer and closer to clinical use, it’s not hard to imagine how valuable a platform like this could be. And this effort is likely just the start. Medtronic’s implantable devices, like its pacemakers and defibrillators, generate incredible amounts of data – data that is linked to, among other things, claims data.

In addition, IBM announced the acquisition of Explorys, a startup that has compiled an enormous dataset consisting of health data on 50 million people. And its partnership with Johnson & Johnson will reportedly focus on using data from multiple sources – from claims & EHR data to data captured from apps & wearables – to improve outcomes in patients getting prosthetic joints.

Ultimately, the insights and tools generated by Watson’s application to all of this data will have to be judged by the medical community, just like we critically assess randomized clinical trials, case control studies, or retrospective data analysis. And the design of the analysis done and platforms built on Watson will have to account for various limitations and challenges, from privacy to data quality. But the opportunities here are incredible.