Using artificial intelligence as a prophylaxis against stroke

SERIES (3): dpv-analytics‘ miniature ECG simplifies diagnosis, saves costs and eases burden on healthcare system
15 March 2022
dpv-analytics' ritmo

“Around 500,000 people in Germany suffer from atrial fibrillation and are often unaware of their condition. Cardiac arrhythmia often goes undetected and usually occurs at night," said Dr. Stephan Kranz, a specialist in nuclear medicine at Cardiologicum Hamburg. "Up to 20 per cent of all strokes are due to atrial fibrillation, according to current estimates," he added. If detected early, the risk could easily be averted: "Blood-thinning therapy can avert blood clots," Kranz pointed out. Ultimately, it's a matter of identifying patients at risk and so far that has beed done using time-consuming, and costly long-term ECGs. Kranz is among the trio of founders behind the Hamburg-based start-up dpv-analytics, which has come up with a simple and convenient alternative named ritmo. This miniature ECG, which is about the size of a passport photo, arrives in the post and is then attached to the patient's chest with a waterproof patch. Seventy-two hours later, it is checked using artificial intelligence.

Improved solution 

The certified medical device was developed by Dr. med. Ekkehard Schmidt, Dr. med. Matthias Glawe and Kranz at the Cardiologicum Hamburg. "We have not deemed conventional long-term ECG devices the ideal solution for a long time," said Kranz. The cable-connected devices are neither comfortable nor reliable to use, he noted. Especially at night, error messages occur easily because the patient turns over in their sleep and the cables become entangled . "When it came to investing in a new generation of devices, Dr. Ekkehard Schmidt, a founding partner of the Cardiologicum Hamburg, thought there had to be a better way," said Kranz. This led to ritmo which was developed in collaboration with the engineering firm livetec and the Hamburg-based AI start-up Psiori and scooped the German Medical Award 2021.

from l. to r. Dr. Kranz, Dr. Schmidt and Dr. Glawe
© dpv-analytics GmbH
from l. to r. Dr. Kranz, Dr. Schmidt and Dr. Glawe

Quality of AI hinges on data

AI is ideal for evaluating a long-term ECG. The self-learning algorithms scan every single heartbeat and detect deviations and anomalies accurately and rapidly. However, the quality of AI hinges on the data with which it is fed. "The Cardiologicum is one of the largest outpatient cardiology facilities in Germany. So we feed our AI high quality data," Kranz noted. "The evaluation jumps to the point of identifying a disorder. We validate the diagnosis so that we can start the appropriate therapy immediately, if needed."

Device for heart attack prophylaxis in preparation

The earlier a risk of stroke is identified, the better and not just for the patient, Kranz stressed. "Strokes are expensive. In the first year alone, the costs average EUR 19,000. More costs are incurred by the degree of disability and occupational disability. The more we screen, the more we save the German healthcare system. This form of diagnostic support is coming and is particularly interesting for home diagnosis and to counteract the lack of medical facilities in rural areas." The doctors are now working on their next project. "We are developing a device for heart attack prophylaxis that like ritmo relies on AI as well." However, it will probably take one and a half years before it is approved as a medical device, optimistically speaking, Kranz noted.

ys/pb

Read part 1 of our series: "Relieving phantom pain with virtual realityand part 2 "Business model: Fighting phobia in the virtual environment".