Personalised algorithms in heart failure telemedicine to improve self-management

Heart failure is a significant and growing clinical problem. Currently, more than 225,000 people in the Netherlands suffer from this condition, resulting in 30,000 hospitalizations per year. This leads to high costs. There is an increasing need for alternative and affordable strategies to prevent hospitalizations and improve the quality of life and self-management. Telemonitoring is a crucial option where patients can collect data in their home environment and be remotely supported to detect and treat the worsening of their heart failure early. Unfortunately, the use of telemonitoring is limited by unnecessary alarms that cause a lot of anxiety and pose a burden on the healthcare system.

In this project, we have developed an innovative algorithm for patients with heart failure based on a combination of biometric measurements such as weight, heart rate, and blood pressure. The goal was to detect early deterioration with minimal false alarms. We collected data from two hospitals, monitoring patients with telemonitoring to map the occurrence of worsening heart failure. This study has shown that, contrary to clinical practice where weight is primarily used, heart rate is the most significant predictor of deterioration. The algorithm performs better than when manually set threshold values are used.

The program code for how we developed the algorithm has been published and is publicly available for others to further develop and ultimately implement.