MHN 2016: Duke's Dr. Bloomfield on Apple vs Google, autism study update, what's missing from HealthKit

By Jonah Comstock
Share

We wrote last week that, at MobiHealthNews 2016, Duke Director of Mobile Strategy Dr. Ricky Bloomfield brought breaking news from WWDC that Apple would soon add HealthKit support for the HL7 Continuity of Care Document to iOS 10.

But that tidbit, exciting as it was, was only a small part of Bloomfield’s talk, in which he shared some early anecdotes and triumphs from Duke’s early forays into both HealthKit and Apple ResearchKit. First though, he spoke about why he thinks Apple has outshined its competitors in this regard.

“The reason Apple and Google come to the top is because in order for this to be meaningful, it has to be on a device that you’re carrying with you,” Bloomfield said. “And of course, most people have an iOS or an Android device. A third party platform is never going to get traction because the way these devices and operating systems are now designed, the framework for healthcare has to be built into the framework of the operating system. And the reason that has to happen is primarily one of security: Only the OS creator can manage that security of exchanging data between different applications in a way that is beneficial to the consumer.”

Between Apple and Google, Apple has seen more adoption for HealthKit than Google has for Google Fit, its comparable offering. Bloomfield offered a theory as to why that might be.

“Apple standardized a set of data elements so everyone could use those in a consistent way. And it’s interesting that we talk about open standards but there can be closed standards too. This is technically a closed standard, but because of the fact that it was made available for anyone to use free of charge, it became a way you could transfer high-quality information, which made it usable,” Bloomfield said. “Whereas on the Google side, they didn’t have the same level of granularity for their data elements and actually kept it much more open in terms of being able to define your own data elements. But of course any time you can define your own data elements and you leave it much more open, that optionality, as we call it, makes it much harder to standardize and to share data.”

Bloomfield said that the effect of HealthKit was obvious from the very first patient, an elderly man with high blood pressure who found himself getting lightheaded during a vacation in France. He was able to send blood pressure readings from a connected device to his doctor back in California. She then adjusted his medication and his symptoms abated.

“This is a very simple example, but if you think about what happened here, this was someone who was traveling abroad, who didn’t speak the local language, who wasn’t familiar with the healthcare system there, and who really didn’t want to take all this time out of his vacation and the expense to go to an urgent care in France to get this taken care of,” he said. “And he was able to take care of it meaningfully very quickly through this technology. … This is just the tip of the iceberg in terms of what these technologies will be able to do.”

Bloomfield said that the missing piece for HealthKit is still data analysis. There’s still no engine powerful enough to really enable this technology to scale to an entire patient population, he said.

“One of the biggest gaps right now, is that the layer of analytics that is required to take this data and do something meaningful with it before it gets to the provider,” he said. “Right now that doesn’t exist. … What we’re really talking about and what we really need and don’t yet have, are more advanced analytics, so you can display not just a graph of the raw data to a provider but you can display the risk that this data will result in for that patient, and the risk for heart disease, stroke, cancer, pretty much everything we want to know.”

He also shared an update on Autism and Beyond, Duke’s ResearchKit study designed to increase knowledge about how autism manifests in children, which could eventually lead to faster diagnosis. They’ve now had 2,000 sign-ups, 5,000 videos uploaded, and 6,000 questionnaires contributed. They’ll publish their first insights in a few weeks.