Wireless intelligent systems
Helge Bjarup Dissing Sørensen, who is an associate professor at the Department of Health Technologies at DTU, has witnessed health technology in rapid development since he introduced a course in personal portable health technology at DTU in 2012.
”From the very first day of the course, I made a prediction to the students that health technology would move from the hospitals to the private sphere in the coming years.”
"We are in a transition phase, where ground-breaking health technologies are under development."
Associate Professor Helge Bjarup Dissing Sørensen
At the course, the students now develop wireless portable technology solutions with biomedical signal processing and artificial intelligence (AI), for example an alarm for epilepsy, an alarm for critical heart rhythms, or they use cameras to take pictures of psoriasis or birthmarks and design associated image analysis methods.
”We are in a transition phase, where ground-breaking health technologies are under development. In the ‘WARD’ project together with Rigshospitalet and Bispebjerg Hospital, we are trying to do automatic monitoring with knowledge-based methods, while the patient is at the hospital. We put wireless sensors on the body. Sensors that transmit data in real time 24/7 to a cloud computer, where we have allocated algorithms based on knowledge of medical doctors combined with engineering science. The algorithms are based on mathematical modelling and AI that can detect complications or other abnormal events in the patient. When complications or signs of complications are registered, an alarm is sent directly from the algorithms to nurses’ and medical doctors’ mobile units. We are currently upscaling this system so it can run at hospitals in all Danish Regions.”
Today, it is a clinical workflow that health staff does manual measurements from each patient three or four times per day, but this will soon change radically with the above mentioned health technology, Helge Bjarup Dissing Sørensen assesses.
”We will go from manual measurements and looking up patient data manually to automatic measurements and estimations of the patient’s well-being. I.e. checking for complications. Soon there will be around the clock monitoring, while the patients are admitted and later at home to keep an eye on the patient’s recovery. In this way, we obtain 24/7 monitoring of the patient’s health. Thus, we have modern health technology in the clinic and at home. In principle, you could imagine that the GP prescribes that a citizen takes home this monitoring equipment.”
Smaller, better, smarter
Today’s wireless sensors are smaller than a few years ago, and they will most likely continue getting smaller, better and more user-friendly in the future.
”Citizens prefer that monitoring equipment is invisible when they need to wear it – just like a hearing aid hidden inside the ear. Today, we use a wireless blood pressure monitor, among other sensors, which is based on a cuff that pumps. In this case, we are working with algorithms that provide equivalent estimates compared to other sensors such as so-called medical watches that are very pleasant to wear. In the future, sensors will be smaller and measure even further modalities, different signals, and it will open up for estimating additional early signs of diseases. This could be advanced stress parameters, critical combinatorial health parameters or respiratory abnormalities etc. We also experiment with wireless registration of mental conditions. Furthermore, we have, together with Rigshospitalet, in a previous project developed (and patented) a method for identifying early markers for Parkinson’s, allowing us to detect the disease several years before it has been fully manifested.
In the future, many of the body’s different types of signal will be detectable with sensors, and the technology is spreading out, but the user-friendliness is still a bit lacking in some sensors.“
Systems create a new foundation
”A likely scenario for the future is that the GP will have a number of wireless systems on the shelf, which people can take home for a period of time, and then these can record how the person is coping. In principle, in two ways. Online, i.e. we get an instant response, if there are any irregular health conditions, and offline, where the results will be pulled out of the device when it is returned to the GP or the clinic. These systems will discover if there are any abnormal events instantly or provide the medical doctor with a data foundation for further evaluation of the patient.”
Helge Bjarup Dissing Sørensen also believes that we soon will be able to find wireless systems at discount price (but regulatory approved) on the shelf at the pharmacy or in the supermarket.
“However, there is a major difference between the gadget-like so-called health devices, which are found in abundance, and the regulatory approved devices, which currently are very limited in numbers, but which we will see many more of in the future.”
Not a substitute for the medical doctor but a ground-breaking addition
”A Danish company has proved they can estimate skin cancer with a portable system using advanced image analysis. Such systems cannot substitute the medical doctor, but they can advice people to see their doctor if a birthmark looks suspicious. You will be able to buy systems like that for identification of health problems. Moreover, they will be approved and regulated.
In addition, it is important to emphasize that artificial intelligence/AI is a mix of knowledge within medical and engineering science, and that this has enormous possibilities, but it also has its limitations. For example, the medical doctor will always be in charge of the final diagnosis.
In time, the portable wireless intelligent systems will become safer and deliver very high probability for correct assessment of a number of diseases for example skin disorders, heart incidents or abnormal heart signals. In addition, the number of false alarms will be very low. The systems will not be infallible – just like a medical doctor is not – we have for example invented several systems, which estimate diseases or symptoms with a similar accuracy compared to the best medical doctors, among these a system for assessment of narcolepsy.
I expect that more portable wireless devices for home-use will arrive at the market. Then you will for example draw a small blood sample, and the device provide a pre-diagnosis for a number of conditions and diseases. Intensive research is taking place in predicting diseases from a single miniscule blood sample, but this is not within my area of research.”
On the verge of a breakthrough
”A lot of things are in the pipeline, and the development in digital health technology means that systems will be accessible for estimation of your health conditions and possible early signs of illness at a price so cheap that they will be available in the supermarket.
As an example, as a research group, we have in collaboration with a Danish health technology company developed a wireless alarm for epilepsy for children and adults, where the unit is placed on the arm. When a seizure emerges, an intelligent algorithm sends an alarm to a receiver, so for example parents can save the child from harming itself and in worst-case scenario save it from mortal danger at night, if one of the particularly dangerous types of epileptic seizures hits.
We will see more compact systems such as headbands or other textiles, which contain electrical sensors, and maybe microphones and motion sensors etc. Based on the recorded activity, it will be possible to estimate motion disorders, brain diseases, sleep disorders etc.
Ground-breaking digital health technology based on medical signal processing, big data and AI is clearly spreading out. It is an enormous ongoing process, but it will take time to build up these systems based on a lot of patients and sufficient data about abnormal and normal incidents to get enough basic knowledge for the mathematical algorithms, which have to be developed. But it is on the way!”
Photo: Jesper Scheel