Explainable AI for Predicting Hospital Admissions in Heart Failure: ExplAIn-HF

Explainable AI for Predicting Hospital Admissions in Heart Failure: ExplAIn-HF

The goal of this observational study is to develop and validate an XAI based model that predicts HF events and identifies modifiable con-tributing factors, and to evaluate the added value of integrating high frequency smartwatch data compared to usual care in patients with heart failure. The main objectives of the study are:

1. To develop and validate an XAI based model that predicts HF events and identifies modifiable contributing factors, and to evaluate the added value of integrating high frequency smartwatch data compared to usual care. In the Netherlands usual care includes remote monitoring of heartrate, blood pressure, weight and symptoms.
2. To include insights of the smartwatch into activity patterns, impact on quality of life (KCCQ-12), and patient satisfaction with net promoter score (NPS).

Participants will wear a smartwatch for six months and perform an I-lead ecg with the smartwatch weekly.

Inclusion Criteria:

• Adult HF patients
• NYHA class II, or NYHA III
• Recent rehospitalization for HF
• Usual care with active participation in the telemonitoring program "Zorg Bij Jou" (ZBJ)
• Possession of an iPhone 11 or newer and compatible with the study apps

Exclusion Criteria:

• Lack of digital literacy or inability to use the device
• Not capable to sign consent
• Referral to hospice