Artificial Intelligence is not a feature being added to analog devices but more of the expansion that offers the possibility of a single-use disposable to capture the economics that previously belonged to the capital equipment, the imaging suite, and the operator’s accumulated skill.
Analog device companies sold instruments and rented physician talent to produce the outcome. AI-enabled intelligent disposables will be selling the reproducibility of the outcome itself. As we move towards payment based on episode-based reimbursement and reproducibility is the only thing that pays in that environment.
This is the framework of the value provided versus value captured dynamic that I have been writing about. The coordinating layer is where the profit migrates, and the defensive move for a device company is to pull the coordinating layer inside the disposable before a health system or a platform vendor builds it above you.
There are a few mechanisms that will drive this migration. And they are all economics at their base.
Skill compression is the actual product. As an example, we are seeing a deluge of PE thrombectomy products coming to the market. This is a fantastic example of a judgment-dependent procedure. Clot burden, clot character, how much blood you are willing to lose, when to stop.
Today the market is rate-limited by a few hundred proceduralists who are genuinely good at this. When we add intelligence inside the catheter it expands the addressable operator pool to general interventional cardiology and vascular surgery, and eventually to community hospitals that currently transfer these patients out. That is market creation through deskilling and no longer market share capture as was the case with Inari and Penumbra who grew market by training champions. This next generation grows by not needing them.
The question that is the correct question to ask is “who is going to pay for this AI?”
Payment is where AI stops being a cost and becomes an asset. The payer does not buy the algorithm, they buy variance avoidance. Fewer reinterventions, fewer ICU days after intermediate-high risk PE, less contrast and radiation, shorter length of stay, fewer transfers.
Under CJR-X logic and the direction of episode payment, a device that itself captures what happened during the procedure, versus a human writing it down as a summary captured afterwards generates its own longitudinal evidence stream and can therefore underwrite a value-based contract.
An analog device cannot do that because it produces no data about itself. Asymmetry is where the market is moving.
There will be a clinician uptake question. If AI accelerates skill acquisition and everyone wins, the question to be navigated is who will manage the clinician conflict? Proceduralists who built reputation and referral volume on being the hard-case operator may lose status when the hard case becomes routine. Those are the champions who will most likely drive early adoption and write the first papers and present on the podiums.
What may occur is that the skill compression does not remove the expert but instead moves the expert up a layer into case selection, program design, and managing the operators below them.
Practicing medicine may move from the hands of a gifted practitioner to a digital practicing strategist who is above the case load and architecting the digital landscape.
Skill has always been the most valuable and the most unevenly distributed asset in medicine. Whether you survive a pulmonary embolism may have depended, to a degree on who happened to be on call. If these companies are right, the best hands in the world become the default hands, everywhere, at once.
Read full article: https://www.linkedin.com/pulse/ai-pullings-analog-devices-future-joe-mullings-eswxe/?trackingId=ZD2n9oZ8GeU4k5FBponrJg%3D%3D
