Heart failure prediction in patients with remotely monitored implanted cardiac devices: a multiparametric model

Heart failure (HF) and atrial fibrillation (AF) are widely spread among the global population, especially in elderly patients. Usually, they co-exist because they share the same risk factors, but they may also be characterized by a complex cause-effect mechanism, responsible for worsening in the patient’s clinical status, which can lead to recurrent hospitalisations. In fact, it is well known that HF-related hospitalisations represent the 1-2% of all hospital admissions and that they have the highest readmission rate in the short term. Therefore, in order to try to reduce the clinical and economic impact of HF hospitalisations on the national healthcare system, it is fundamental to implement a predicting model to timely identify HF worsening before they require to be managed in a hospital set-up. This project is a continuation of a retrospective study on patients of the Ausl Romagna, which started few years ago at the cardiology department of the Morgagni-Pierantoni hospital in Forlì. The novelty tested in this actual version is the use of the AF burden as a new parameter employed in the final risk-score computation, together with daily and nightly heart rates and physical activity; hence candidates to be selected for this study are only patients implanted with a device carrying a working atrial catheter. The predicting algorithm is realized in Matlab and it combines the daily data trends, remotely collected by ICDs and CRTs devices, to build-up a risk index, which is compared to an upper nominal threshold in order to activate an alarm, indicating a risk to experience a hospital admission in a short time. The results show a good algorithm performance and the obtained metrics are in line with both the version not including the AF burden in the score computation, also applied to devices with only a ventricular catheter, and the alert algorithms designed by the main biomedical devices companies.