I attended the two-day MedCity Converge conference in Philadelphia last week. The event was jam-packed with knowledge and networking from some of the health industry’s best and brightest. As someone who spends every waking hour thinking about wearable and health data, I was thrilled to hear so many diverse panels and points of view on the topic of patient generated data. At Strap, we've been working on scalable analytics for wearables since early 2014, so seeing the industry start to embrace the technology en masse is exciting and validating.
Anil Jain kicked off the day Tuesday with an informative keynote on the latest from IBM Watson Health. In 15 short months, Watson Health has grown to an impressive initiative through acquisitions and internal development. Watson’s cognitive computing platform is transforming the industry’s ability to process large amounts of data in real time, and brings with it the potential to save patients, providers, and payers significant amounts of time and money. Watson's value prop is strong, especially considering the fact that medical data will double every 73 days by 2020, and today we’re sitting on over 150 Exabytes of it.
Besides the keynote, most of the sessions I attended were directly about wearables and patient generated data. I was happy to see that plenty of the other sessions had a lot to say on the topic as well. I’ve done my best to condense my key takeaways from the conference below.
Patient generated data needs to be accurate and actionable
The technology behind consumer fitness devices is improving, but the fact that devices such as Fitbit may provide inaccurate readings that cannot be relied upon for clinical decision making has been well documented.
Additionally, many patient generated data applications still rely on user input. In certain cases, such as weight tracking, users may make mistakes or input inaccurate values. These potential hurdles can be overcome with proper data context and layers of analytics tools; however, analytics and alert tools need to be smart enough to detect these potential anomalies.
As we've known for some time, providers do not want step counts and raw data points on their EHR display. The more actionable and insightful the data, the more likely a provider will embrace using the data in a clinical setting. Data views that are built for specific conditions are a must; there is no "one size fits all" dashboard for patient generated data.
Often, patients are forgotten in this conversation, but above all else, the data needs to be actionable for the humans who are generating it. One key to patient-centric experiences could be providing the right incentives, such as monetary rewards or helpful content. Keeping patients engaged is a critical component of any mobile health strategy.
Patient generated data needs to fit into existing provider workflows
One of the biggest barriers to wearable data being used in a clinical setting is the amount of time potentially required for providers to review data. Providers do want access to additional context, but they are reimbursed primarily for their time. So far, it is unclear how much time it would actually take to use patient generated data in existing provider workflows.
Without access to an insightful dashboard of a patient's data, the time required to comb through activity, sleep, weight, and other data could exceed the typical time required to perform basic evaluations. But with a clean and insightful interface, providers could include the review of patient generated data under established CPT/HCPCS codes, such as 99204 for a comprehensive initial patient review.
Beyond the primary care setting, care managers will be the primary point for analyzing and receiving notifications about anomalies in data generated by high risk patients. Care management workflows are now well established and becoming fully integrated into the existing health care ecosystem, so bringing insightful data from wearables and apps to care managers is an important step to improve outcomes and prevent readmissions.
Insurance and Pharmaceutical industries are ahead of providers with the use of consumer wearable data
One of my favorite sessions was Digital Health’s Arrival in Pharma. Michael Doherty of Roche and Foundation Medicine talked about the use of wearable data for clinical trials, specifically around hemophilia and COPD, where sleep and activity data are critical components of data driven studies. We also heard form Pfizer, Lilly, and Bayer about the use of patient generated data in clinical trials.
In An Evolution That’s Leading Somewhere: The Next Generation of Wearables, we heard from Craig Hankins of UnitedHealth and Jim Mault of Qualcomm Life. Earlier this year, the two companies collaborated to release a new wearable device which rewards policy holders for hitting certain activity goals. The program goes beyond traditional step count goals, and looks at raw data throughout the day to reward and incentivize members based on minute-by-minute data changes. As Mault pointed out, the most impressive point of this program is that underwriters at UnitedHealth created new policies from the ground up that consider appropriate incentive levels based on member generated data.
It seems that while both patients and providers want to embrace data sharing in the clinical environment, the interoperability layer needs just a bit more time to solidify. Joe Kim from Lilly pointed out, “its hard to move data around when the rails aren’t the same width.”
Luckily the "health data railroad" is progressing nicely, and we’re probably just a couple of years away from being able to easily unlock our wearable data to any provider in the US. Platforms like Redox are making it easy for software vendors to interoperate with any EHR. BurstIQ is doing some tremendous work with blockchain technology that could unlock its potential for health care. At Strap, we're working to make mobile health data insightful and actionable for providers, payers, pharmaceutical, and wellness companies. There are countless others working day and night to make their dent in healthcare, and the next few years should be very interesting as the industry begins to fully embrace patient generated data.