Medical innovation requires patience due to strict regulations

A smart camera that continuously monitors heart patients, software that analyses vital signs and alerts doctors in time: innovations like these promise to make healthcare safer and more efficient. Yet it often takes years before such technologies reach patients. Strict European regulations, extensive clinical trials, and ethical requirements make the journey from idea to hospital bed long and complex. 

Researchers at the Catharina Hospital in Eindhoven know all about the long journey. Together with Eindhoven University of Technology (TU/e), Eindhoven MedTech Innovation Centre (e/MTIC), and Philips, they are developing a smart camera that uses AI to detect patient vital signs. The project, called Advance ForSee, has been in development for five years.  IO+ was present at the camera's launch, where researchers also discussed the lengthy and complex process of medical product development.

From idea to market

Modern medical devices are increasingly being combined with smart software that goes beyond traditional technology alone. The European Medical Device Regulation (MDR) contains rules that software must comply with if it is used as a medical device. For example, monitoring patients, providing information about their health, or supporting treatment decisions. 

Before the equipment is allowed on the market, it must meet strict requirements. How strict the requirements are depends on how significant the information collected is and what decisions are based on it: the greater the impact on health, the stricter the requirements and supervisory regimes. 

Manufacturers must demonstrate that the software works reliably, safely, and effectively for its intended medical purpose — whether that be home monitoring or clinical predictions — and that the system complies with European legislation before it can be used in healthcare.

The introduction of these strict rules in 2021 has raised the admission requirements and clinical evidence burdens, leading to longer development cycles. The average duration of development, clinical evaluation, and market authorisation of software as medical devices takes between three and seven years. 

Phase 1: the idea

The researchers participating in the Advance ForSee project talked about how the process has been going so far. They started with the desire to monitor heart patients who are only seen every six to ten hours, because their mortality rate is still too high; namely, 40% of patients in that nursing ward die from complications. What happens between check-ups determines whether patients' health improves or deteriorates, according to cardiologist and professor Lukas Dekker.

Several steps must be taken to realize this idea. First, it must be determined what the doctors want to measure and how.

Phase 2: funding

The entities working on Advance ForSee must obtain their money through funding. There is no time frame regarding how long it will take to find funding. Supporting the creation of the cameras and software in Eindhoven is ZonMW, a funding organisation for innovation and research in healthcare.

Phase 3: creation

Once the initial hurdles are overcome, the product, in this case the software, must be created. For Advance ForSee, it must be ensured that it is safe and that the information stays within the hospital walls. Additionally the software and AI need to be able to handle and analyse large continuous data streams. The software must also be trained in such a way that it can analyse patients regardless of their sleeping position or facial features, for example. “This requires 20 years of training data,” according to Dekker. Such data can be acquired from imitated situations or existing data. 

Phase 4: trials, trials, trials

Clinical trials must also be conducted to gather data and, where necessary, modify the software and cameras. At the moment, five years after the project was started, the smart cameras are being tested in clinical trials. “Hundreds of patients have already been monitored,” Dekker says. More clinical trials are still required. 

Phase 5: research papers 

Then, when the software and cameras do what the researchers need them to, research papers must be written and peer-reviewed. It is unknown how many research papers are necessary or how long they will take to write and be peer-reviewed.

Phase 6: getting approval to produce 

When there is enough evidence, it must be approved by entities that deal with ethics, EU law, medical law, and EU AI laws, for example. Getting such approvals can take many months.   

Filming patients and storing their data also raises ethical and privacy concerns. A way of saving time is to involve patients and nursing staff in the development from the beginning, as the Advance ForSee project does. This way, the ethics committee can immediately see that privacy, ethics, and the needs of both patients and nursing staff have been taken into account. Involving patients and nursing staff at an early stage also minimises the chance that the product will have to be modified for not meeting user standards.

Phase 7: producing

Then, a manufacturer needs to be found to produce the product. In the case of Advance ForSee, it is the cameras that need to be created as the software is being programmed by TU/e. “Philips will most likely produce them,” according to Arthur Bouwman, anaesthesiologist at the Catharina Hospital and professor. 

Adaptation for each patient group

The Advance ForSee camera and software are currently designed for patients who have undergone heart surgery. “The model must be adapted for each patient group,” says Bouwman. Adaptations require the developers to redemonstrate that it meets all requirements. 

The process takes more time, but also offers more opportunities; “For the time being, it is only aimed at heart patients, but later on, the intention is to also use it in the recovery from other procedures,” says Dekker.