A couple of recent articles got me thinking some more about the anticipated impact of “consumer health devices.” The first was in the Wall Street Journal and reported on the annual consumer electronic show (CES) in Las Vegas. The other was a review in Nature Medicine by Eric Topol on the current state of artificial intelligence in medicine.
The Journal piece made the point that CES had more health-related offerings than ever before. Sure, the show still featured the latest home entertainment offerings (how about an 98” 8K OLED TV?), but now also showcased everything from a watch that can check your blood pressure (see below), to a sensor and associated app for assessing the quality of sperm in a specimen collected in the comfort of one’s own home (use your imagination). You can get a rundown on “all the new cool gadgets” here.
Topol’s piece was a comprehensive (216 references!) look at the capabilities and applications of artificial intelligence in medicine. In particular, he assessed the impact of “deep neural networks” on clinicians, on health care delivery systems and on patients. These networks are a particular kind of machine learning that generate sophisticated pattern recognition algorithms based on an annotated set of images. The review is worth the read, but I would say the punch line is that there is incredible potential, but not (yet) much evidence, of the transformative power of artificial intelligence in medicine. That said, the inexorable growth and availability of large bodies of data in digital format (necessary to train algorithms), ever more rapid and less expensive computing power, and nearly unlimited data storage capacity in “the cloud” portend a much more extensive role for this technology.
So we are at an interesting moment in time. There are a lot of companies trying to commercialize all kinds of new devices and services based on new sensor technology and data interpretation dependent on artificial intelligence; early adopters (consumers) are buying and using them, often without medical direction; and traditional healthcare providers are ill-prepared to incorporate this technology into their practice, or to respond to patient-generated data and clinical insights.
I think this all raises a lot of important questions. Here are just a few:
- How good is “good enough” when evaluating whether a particular device or technology should be used clinically?
- What is the role of physicians in interpreting patient generated data or machine generated diagnoses?
- What are the appropriate standards and enforcement mechanisms to assure data security and privacy?
- Which of these new devices or services actually improve health?
I will be discussing these and related issues in a session titled “Healthcare’s Digital Disruptors: Hope vs. Hype” at the upcoming South by Southwest Conference in Austin. If you’re at the conference, stop in and join the conversation!