Could Digital Health Make Most Medical Devices Obsolete?

Posted in Mobile Health by Brian Buntz on February 18, 2014

Don't say you weren't warned about the coming mobile health tidal wave.

A recent keynote address at MD&M West suggested that digital health technology could end up displacing half of traditional medical devices. That figure might even be an underestimate, says regulatory expert George Samaras of Samaras & Associates Inc. (Pueblo, CO). In an email, which we have posted below, Samaras stressed the importance of first distinguishing between sales numbers or the number of types of medical devices in determining the scope of potential upheaval:

I think in terms of sales numbers there is no question that medical devices that do the following will be greatly sought after:

  • Simplify healthcare (including personal healthcare and monitoring of family health).
  • Cater to instant “gratification” (think à la polaroid) and provide effective feedback.
  • Are not cost prohibitive (once the code is created, the marginal cost of additional copies is negligible and therefore can be very profitable).
  • Increase the throughput of healthcare providers (making them and their organizations more profitable)

Of course, they still need to be safe, effective, efficient, and satisfying to use. Especially, satisfying—not frustrating—to use!

However, I think that most medical devices will not be “pure software.”

Google glucose-reading contact lens
Google is developing a contact lens to measure glucose levels in diabetics. And Apple appears to be hiring medical device professionals to help with development of the iWatch.

Here is my reasoning for the “underestimate” suggestion:

  • The principal value added by most Class 2 and 3 medical devices, which are not mechanical things like structural prosthetics (e.g., knee replacement) or physical accessories (e.g., laser hand piece), is increasingly contained in the algorithms.
  • The critical electro-mechanical infrastructure for algorithms, which used to be quite dear, has now become nearly a commodity with the proliferation of embedded systems for consumer use. There just never was sufficient volume for this to happen with just military and industrial use. The exact same thing is happening with sensors and sensor fusion techniques that permit extracting greater value from the sensed signals.
  • The software tools, while still quite dear, for creating high reliability software systems are also coming down in price (sometimes to the price of almost free—check out and soon will be within reach of sole practitioners like myself.
  • The ability to inexpensively create complex electro-mechanical parts (including 3-D printed circuit boards) using such techniques such as additive manufacturing, and within the decade self-organizing nanoscopic entities, will shift the real intellectual property values to algorithmic creations and innovations (including those algorithms driving the printers and nanobots). Math, physics, and high-quality software engineering will gain high perceived value!
  • The forthcoming crop of healthcare professionals (physicians, nurses, pharmacists, etc.) grew up tech-savvy prior to choosing their profession–they will expect sophisticated HIT and will be greatly dissatisfied when they do not have it or have to struggle with it (e.g., HF/UX issues).
  • Economic and competitive pressures in healthcare delivery, combined with the parallel new crop of tech-savvy managers and an emphasis on collective purchasing, will put downward pressure on medical device prices—meaning, at least from my perspective, that the cost of goods sold needs to decrease – which brings us back to the earlier point that the marginal cost of duplicating software is very low and the fixed cost, if distributed over a very large number of copies, can be made quite low as well.

There are other factors as well, but I think these really jump out at me.

I realize that there is a great deal of current enthusiasm about mining “big data”, but I am hopeful that is just another passing fad, like so many prior management fads. Mining huge volumes of data from not necessarily representative populations without controlling bias or understanding effect size, will have little validity (not internal, nor construct, and almost undeniably not external) given the inherent sourcing problems. It doesn’t mean it is useless; it just means that it will not be as ubiquitous as some are suggesting.

The problem we have now (that absolutely will have to change before this can all be realized) is that the medical device industry does not have the experience, expertise, and toolsets to create high-reliability, high security, software systems. The industry must learn what the military, aviation, automotive, and other industrial sectors learned long ago. Until that changes—and I think it will slowly, possibly even through attrition—I would not yet bet all my money on the huge proliferation of medical devices that were mostly software. It will offer some interesting investment opportunities…

Samaras is a biomedical scientist and engineer at Samaras & Associates Inc. (Pueblo, CO). Trained as an electrical engineer, he has doctorates in physiology and industrial engineering and is a licensed professional engineer, board-certified human factors engineer, and an ASQ-certified quality engineer. He has a number of biomedical patents and publications in physiology and engineering (hardware, software, human factors, and quality). He has worked at the FDA/CDRH as a reviewer and manager, was a medical school and engineering graduate school professor, and founded an engineering firm that he ran for a decade.