Built own AI model to detect heart failure

Elias Stenhede, an engineer and postdoctoral researcher at the AI Innovation Centre at Akershus University Hospital (Ahus) in Norway, has built an AI model that looks for signs of heart failure in ordinary ECG recordings. The model analyses standard ECG measurements and searches for patterns that may indicate the condition. A photo with the report shows Stenhede having an ECG taken himself.

The aim is early detection. Henrik Schirmer, a heart specialist at Ahus, explains that effective medicines exist for heart failure. Finding the disease early means patients can start treatment before it progresses further, and this is one of the main goals of the model.

The material available gives only this introduction to the project. It does not say how accurate the model is, how it was trained or tested, or whether and when it will be used in clinical practice at Ahus or elsewhere.

The work matters because it uses a routine, widely available test, the ECG, to flag a serious condition that responds well to treatment when caught in time. The story was published on 31 October 2026 in Norwegian, under the health technology heading.