How AI will affect medical device business models.
April 2025 , Sam Moreland
Healthcare organizations face skyrocketing costs, complicated delivery hurdles, and growing demand from an aging population. But there’s a transformative trend underway: artificial intelligence and data-driven solutions that can fundamentally change both the way healthcare is delivered and outcomes for patients.
Making medical devices is costly for many reasons; R&D, manufacturing, regulatory, distribution, maintenance etc. Making a profit can be difficult due to the comparatively small install base (compared to consumer markets). This leads to 10 or 100+ times markups on the development costs for the consumer.
These high costs have huge disincentives for healthcare systems to adopt technology on mass. Due to increased cost pressures, the number of units they will buy will likely be lower, the will have longer procurement processes, and applications of those devices could be rationed. There are also serious issues with damage and theft in all healthcare areas and combined with the high costs of medical devices, it disincentivizes medical usage.
Where does data fit in?
There are three levels of medical data products:
Informational: Displaying or providing access to data in order to help a doctor diagnose or treat a patient. It has no capability to dictate care.
Diagnostic: The product can make decisions about the course of a patient's care.
Life supporting: The use of the product is currently keeping the patient alive, or is supporting this effort.
Currently data product maturity in healthcare is very low. We are still mostly at the level of displaying and aggregating data (informational). There are some workflow optimization tools such as medical scribes. But in terms of maturity, the level of data use is predominantly informational. There are 2 main reasons for this, the level of regulatory burden and company risk is lower.
However the FDA has recently released their framework for AI/ML regulatory adoption and improvement (which I have written about here). We are starting to see more diagnostic models, however these are fundamentally used in radiology but other specialties are fast catching up. Data based life supporting tools are generally in the form of alarms or PID systems on ventilators. These are very basic functions and can have huge issues such as alarm fatigue. There is now a huge opportunity to disrupt with a clear framework to work from!
Why should medical device companies care?
The short answer is people pay more for knowledge than they do for tools. What a doctor bills in an hour will outstrip the cost of an ECG monitor. The cost of an MRI is equivalent to the yearly salary of a senior clinician. Clinical decision making is a much higher value proposition.
Demand is also going to skyrocket. By 2040, about one in five Americans will be age 65 and we are currently in a shortage of physicians which is only going to get worse. Clinical decision making AI systems will and must take up the slack.
There is huge amounts of good and money to be made in being able to provide data products that start to take on more of the higher level functions of clinicians (diagnostics and life supprt). Companies will come along an offer integrated solutions, you can already see this with Phillips healthsuite and GE Edison. Medical device companies are perfectly situated to take advantage of the shift to AI!
Healthcare data.
Healthcare data mainly comes from sensors such as ECG, MRI, spectroscopy etc. Most of these systems are currently closed source and proprietary, partly due to regulation and partly due to competition. Until now there has been a consensus on the hardware model of monetization among companies.
But with the AI-DSF framework and the rise of wearables, the barrier to entry for new medical device manufacturers is lower than ever. New companies will enter the market with good devices and they will monetize in a different way, with data products.
Medical device companies are in a unique position. They have the experience of gathering data for clinical trials, the regulatory structures in place to support clearance and an already existing customer base. The next stage is to build out a data product (AI, data engineering and product) function to capitalise on the shift.
Companies will start to see the hardware as the cost of doing business and monetize through data products. There are many industries where the hardware model has been replaced with a service/platform model. In this model, harware is sold at cost or below cost with renumeration coming from application sources. Cloud computing, videogame consoles, low cost phones etc. Some medical products do this through the sale of disposables. Monetizing data over hardware allows greater utilisation, more data and happier customers.
Why is data so important?
AI that is good enough to start helping with diagnostics or life supporting activities need a lot of data, and medical device companies are perfectly poised to get this! There are no open datasets that can be leveraged or scraped from the web. It is medical device companies that can collect the data and create a moat.
Medical device companies must evolve from solely selling hardware to offering data-centric solutions. By leveraging AI and integrating robust data strategies, these companies can unlock sustainable revenue streams, accelerate clinical decision-making, and improve patient outcomes. If you’re ready to discuss how to integrate advanced AI into your current medical device portfolio or build out an AI development strategy, contact us for a consultation.
