Wound dressing with sensor monitors healing process

Is the wound healing or is it infected? Physicians must change the bandage to find out because wounds are covered with non-transparent dressings. What happens under the wound dressing is a "black box". Armin Haas and Professor Kai-Uwe Zirk want to change that. COMPAMED.de asked them about their approach.

Mr. Haas, since 2020 code'n'ground AG has joined forces with the Private University of Applied Sciences for Business and Technology (phwt) to "elevate primary wound healing to a new digital level using sensor technology and artificial intelligence", according to your press release. The project has been completed for the time being. Did you reach your goal?

Armin Haas:  Our goal within the scope of the project was to produce evidence that sensor technology on the skin can measure parameters such as moisture and temperature and provide information about the wound healing process thanks to connected AI. We managed to accomplish this goal. This is how the project tasks were distributed: phwt, headed by Professor Zirk, was in charge of the technical part pertaining to the wound dressing. Code’n’ground used its expertise as a software company to set up the AI architecture. We were responsible for the AI portion and facilitated the data transfer from the wound dressing.

Professor Zirk, how did you approach the project?

Prof. Kai-Uwe Zirk: The overarching project goal was to detect a disruption in wound healing without the need to remove the bandage. As Mr. Haas just mentioned, our part was to develop a corresponding sensor structure for capturing the measured parameters, which can be easily integrated into a wound dressing. In the end, we incorporated conductive structures into a plastic film, which facilitate simultaneous measurement of multiple parameters. The plastic film is applied to the wound dressing to enable continuous temperature and humidity measurement. An increase in temperature may be a sign of inflammation. A heavy increase in moisture – due to wound drainage, for example – may also indicate impaired wound healing.

You incorporated a wire into the wound dressing. What is it made of and is it applicable to both processes?

Zirk: We use the same sensor structure for both measurements. It currently has an ultrathin copper layer, with a thickness of just a few micrometers. The arrangement of the sensor structures is pivotal for us to capture both parameters simultaneously. If the temperature changes, the conductivity changes as well. That's one parameter. If we want to track the moisture level, we examine the environment of the structure. This is where we can detect a change in electrical capacitance. The structure is incorporated in a specific arrangement between two plastic films. The film is elastic, stretchable, and kind to the skin. It is applied to the wound dressing via thermal connection. The dressing is placed in sterile packaging, allowing nursing staff to open the packaging as usual, apply the wound dressing and subsequently cover it with bandaging.

Does the wound dressing collect and transfer data continuously?

Haas: It sends data continuously, which is subsequently always processed via AI. The dressing has a storage medium that stores and transmits the data. An app then reads the data.

That means nursing staff must use an app to get a wound assessment?

Haas: Yes, for technical reasons. We decided against continuous transmission and chose to store the data. However, we also want to make it possible for patients to transmit data. An app is ideal for this, and it is an important part of our AI architecture. Right now, if a patient has a surgical wound, the hospital has three measuring points for said wound, namely, each time the wound dressing is changed.

This is followed by a wound assessment to determine how the wound heals. The wound is then bandaged again. The next time you get information about the wound healing process is when the dressing is changed again. In contrast, the smart wound dressing not only records data at three points but also measures temperature and moisture throughout the entire wound healing process. That means we generate up to 30,000 measuring points, allowing us to monitor the healing process. To do this, we collect parameters pertaining to the correlation between wound healing and duration as well as temperature and moisture. There should be an alarm if the parameters no longer match to alert the nursing staff to check the wound! This makes the sensor in the wound dressing the first source of data.

We also want to involve the patients. The proposed app likewise lends itself to this purpose as we want to create more measuring points by having the patients help drive the data transmission. They can document the wound healing process via the app by indicating pain or feelings of tightness, for example. One could subsequently correlate the other data, meaning moisture and temperature. All this was part of our modeling, though not yet part of the actual implementation.

Another source of information would be feedback from healthcare professionals. When a wound has become infected, you need to know what treatment you should administer. Do you just change the bandage, or should you administer antibiotics? We want to record these interventions when there is a disruption to determine the most successful treatment in the future. This outlines the AI and the architecture we developed in this setting. We use different sources to draw conclusions and ultimately support the medical team.

The wound dressing has an attached data logger, which contains the collected data. How big is it in size?

Zirk: The size will hinge on the quantity in the future. With a smaller number of pierces, it will probably remain somewhat bigger in size as the development effort is immense. If one can expect a larger quantity, the data device can become tiny in size because you can choose an entirely different technology. About 1.5 centimeters in diameter and maybe five or six millimeters in height is realistic if we have large quantities.

Is the wound dressing industry ready?

Zirk: Right now, we have available demonstrators. In other words, we have so far proven the function of the wound dressing in the lab via tests on technical models. The next step would be to use the wound dressings on humans and let AI process the collected data. We are presently searching for partners who are interested in this system and would like to join and support our efforts.