Digital Health Technologies and the Tricky Questions They Pose
The problems facing modern health care systems in industrialized nations are considerable — especially in the United States, which now devotes nearly 20 percent of its GDP to medical treatment. In the past nearly six decades, U.S. health care costs have risen nearly triple the rate of the consumer price index. While no other nation pays as much for health care, many other countries — from Japan to Germany — must contend with aging populations and spiraling costs.
Digital health technologies, an umbrella term encompassing everything from IoT-enabled medical devices to AI to health care apps to genomics to wearables, has long promised to help improve health care outcomes and cut costs by driving efficiency. But it is tricky to harness these disparate technologies to drive dramatic change and align all incentives for all stakeholders across the health care industry.
Tricorder, Where Art Thou?
To date, however, the rollout of such technologies is largely confined to isolated areas, said Chris Zant, a principal in Deloitte Digital’s life sciences division in an interview at Dreamforce. “The tricorder isn’t there yet,” he said, referring to the mythical “Star Trek” device that could help diagnose a range of diseases. “There’s nothing out there is just going to diagnose anything that is wrong with you.”
Digital health technologies are beginning to transform health care for a number of committed patients and caregivers. Some patients with atrial fibrillation — irregular heartbeat — now have the option of monitoring their condition themselves, sending worrisome waveforms to specialists for further analysis. And while digital health enthusiasts have long mused that mainstream consumer mobile devices would become medical devices, that possibility appears more likely with the debut of the Apple Watch 4, which has an integrated FDA-cleared ECG monitor.
If a mainstream electronics company were to debut a smartwatch with integrated blood pressure monitoring, the impact on health care could be considerable. More than 100 million people in the United States alone have hypertension. A device that could regularly assess blood pressure without repeatedly constricting the arm could have a good shot at winning mass appeal.
“But otherwise, I think you’re looking at key intersections — key disease states and the invested patients in those disease states,” Zant said. Consumer-facing digital health technologies “need to be affordable” and not pose undue regulatory hurdles.
Which Health Dashboard to Use?
The idea of a health dashboard, which gives consumers a real-time overview of their health and a warning if something is amiss, has long been an attractive idea. But building such a dashboard is by no means straightforward. For one thing, there is a tendency for many digital health technologies to have individual apps and companies with electronic medical record (EMR) software to resist full integration with competitors’ offerings.
But there are two paradigms vying to help pull the pieces together. “If you kind of look at it from a vendor perspective, there is the Apple paradigm that says: ‘I own it. It becomes my health wallet,’” Zant said, referring to the Apple HealthKit app, which accepts data from an array of sources. “Anything that can talk to your iPhone can permission [to send] the data into that HealthKit, and that becomes a single point of truth.” The potential for the app to integrate with EMRs could make the goal of a patient dashboard a reality for Apple fans. But weaving in EMR data still a challenge. Every EMR is unique and integrating them together is challenging, even with standards like FHIR and Argonaut. “[Those standards] are there, but they don’t really get used. So it ends up being a ground war on EMRs,” Zant said. But the paradigm of integrating EMR data, sensor data and IoT data on a smart device is promising. A patient could have the proverbial single source of truth for health care data while permissioning certain data streams for caregivers. The other paradigm for integrating health care data is more of a Google Android approach, where the data resides in the cloud, which can integrate with EMR data on the back end. “I think it might be one of those or maybe both of those two paradigms [that will win] — whatever people are comfortable with,” Zant said.
How Should We Deal with EMRs?
Electronic medical records were supposed to help transform health care. On the surface, it is relatively straightforward to see the potential for replacing paper-based record keeping with a digital version. But the fact of the matter is: many doctors have not liked EMR software. In the United States, the health care industry has spent billions on making EMRs flexible and transferable, as Becker’s Hospital Review writes. There are signs things are improving. A Healthcare IT News survey said the majority of doctors were satisfied with their EMR vendor.
But EMR vendors don’t have a lot of flexibility when building their systems. “There are key functions you’re just not going to get around: the billing and claims administration functions along with the true patient record keeping,” Zant said. “The EMR vendors are making some efforts [regarding] usability, the analysis of data and seeing your patient as sort of a single view.”
Another challenge for many doctors has been the explosion of health care apps. As the majority of physicians have either an iPhone of an iPad, many vendors have responded by developing apps dedicated to their products. “But the challenge is [doctors] say: ‘Well, I don’t want one mobile app from pharma company A, another from pharma company B and another one from med device company C. That’s more confusing than just looking in my EMR,’” Zant said. “A physician doesn’t want to use 32 apps — one for every disease state they treat.”
Companies like Apple and Google are hinting they may streamline this dynamic. “Apple has been talking about their care-at-home paradigm and Google’s been talking about integration with the doctors’ offices,” Zant said.
Paging the AI Doctor? Or AI Paging the Human Doctor?
Ever since IBM Watson trumped human Jeopardy champions in 2011, the promises of AI in health care have been great. But at present, the most promising applications of artificial intelligence continue to be on well-defined tasks.
There is room for AI adoption to expand to a number of specialized use cases. For example, a 2017 study presented at the ASCO oncology conference focused on electronic patient reported outcomes that found patients who logged into an online site to answer a series of questions had better outcomes than those who did not. The patients had increased survivability just because they were inputting a couple of checks per day and because a nurse practitioner was looking at the patient information and providing feedback as needed. “I think we could put a paradigm in place where we take that nurse out of the loop and put a machine learning or artificial intelligence algorithm there to, say, flag every patient that you know, diverges from the norm by more than 5 percent,” Zant said. “The app could say: ‘Hey, Bill, you’ve answered these five questions this way today. It’s time to alert your doctor.’”
As artificial intelligence becomes more mainstream in health care, it also raises the question of, first, threatening doctors by replacing a number of tasks they routinely perform while, second, providing them with new tasks to do. “What if that artificial intelligence engine is running 24×7, sort of dropping notifications for Dr. Smith including in the middle of the night?” Zant asks. “And the other question [for a doctor] is: ‘What am I going to do with all this data? I’ve got three, maybe four minutes per patient. And now you’re creating this, this stream that is going to take four minutes every hour to look at it, and you start to lose patient time,’” Zant said. “We need to refine that interface to figure out what is the right flow to assist the physician but not overwhelm them so that we’re not losing more patient face-to-face time.”