An article about the long-standing challenge of medical device interoperability concludes that this goal will not be realized until “basic systems engineering and risk management principles and practices” are fully embraced by all stakeholders in the healthcare community, with a full appreciation for the varied and often conflicting interests at play among manufacturers, regulators, and healthcare facilities.
The Association for the Advancement of Medical Instrumentation (AAMI) reports the authors warn that failure to take such an approach could open the door to “potentially catastrophic harm” to both patients and organizations. The commentary, titled “Confronting Systemic Challenges in Interoperable Medical Device Safety, Security & Usability,” will appear in the October issue of the Journal of Biomedical Informatics. An online version of the article has already been published.
In making their points, authors and human factors consultants Elizabeth Averill Samaras and George Michael Samaras quoted an AAMIBlog post written by AAMI President Mary Logan, who called for a “greater scientifically focused commitment to a systems approach” to address the complex challenges in healthcare technology. In this post, Logan also cautioned against implementing solutions developed “in our silos and comfort zones of expertise.”
A siloed approach is not likely to be successful as the safety and usability of interoperable medical devices lies at the intersection of healthcare delivery, regulations, and business practices. While all stakeholders want safe, reliable, and interconnected systems of medical devices, each “has its own overlapping requirements…, including points of vulnerability and … dissonance” that need to be taken into consideration, Samaras and Samaras wrote. For example, so-called “vendor-neutral interoperability” is usually attractive to healthcare delivery organizations but is seen by manufacturers as a threat to their business interests.
According to the authors, a balance needs to be struck whereby manufacturers are able to protect trade secrets and intellectual property while still providing the “requisite risk management and design information necessary to properly identify and mitigate potential hazards, as well as verify risk reduction efforts.” Such information is necessary because—when using a systems approach—the safety and usability of the entire connected system must be evaluated, in addition to that of each individual device, since “emergent properties and new risks arise at the interfaces” of system components.
Samaras and Samaras suggest an approach that would allow manufacturers to anonymously share their risk analyses so the entire system could be analyzed without disclosing any individual manufacturer’s confidential data. However, according to the authors, this “would not resolve all obstacles.”
Overall, the authors believe the “most responsible path forward” is “proceeding with caution, observing well-established safety, security, and usability principles and system validation, followed by (or in concert with) careful testing of new approaches (and re-testing for reproducible results).”
I can’t see the full article, so if the following is already addressed, then great. If not, it needs to be:
In an article I wrote for BIT in 2007 ago about the work of Henry Petroski, I asked the question as to “whether anyone has examined, let alone established, the validity of extrapolating 20th-centry medical technology, development, and REGULATORY PRACTICES (emphasis added)for the 21st-century healthcare system.” The immediate context of that sentence derived from a 2003 article in NEJM about the 2002 Beth Israel network failure where the author stated that “an extended computer network [was] designed to meet the requirements of a much simpler environment.”
The environment within which circa 2016 medical technology is operating includes regulations that derive from the Medical Device Amendments of 1976, which incorporate the notion of “substantial equivalence” in clearing devices for marketing via 510k’s. Does that still make sense 40 years later? My admittedly limited understanding of the concept leads me to believe its grounded in the concept of “intended use”. Can we compare the intended uses of today’s technologies with those of 40 years ago? Or even devices cleared on the basis of being substantially equivalent to a predicate device that was cleared on the basis of a being substantially equivalent to a predicate device … Does that raise up the possibility of a “telephone game” problem?
My guess is it doesn’t, but I’ve never seen the issue discussed. If it has been, I’d like to see it covered in BIT. If it hasn’t, when does it get considered? In 10 years? 20? When?
Just for the hell of it, I just tried to search for “FAA substantial equivalence” but got no hits. For a fact, I hope the flights I have booked for the coming weekend aren’t in planes that were cleared based on substantial equivalence with Wright Flyers (ironically enough, I’m going to Kitty Hawk).
Again, you can ignore the above if it’s covered in the article. I certainly hope it is. But if it’s not, when will it be? All I ask is that experts debate the question that I posed in 2007 and publish the arguments and any conclusions. No system solution for here and now is complete without considering this topic.
The question of the validity of regulatory practices reminded my of an exchange I once had with a smart FDAer who I knew and who I liked. I asked, via email, “Do the quality systems regulations result in quality systems that result in quality products?” His answer, in the spirit of my question, was “That’s our position and we are sticking to it.”