Congestion assessment using a self-supervised contrastive learning-derived risk index in patients with congestive heart failure (CONAN): protocol and design of a prospective cohort study

Bojan Hartmann1,†, Niels-Ulrik Hartmann1,†, Julia Brandts, Marlo Verket,
Nikolaus Marx, Niveditha Dinesh, Lisa Schuetze, Anna Emilia Pape,
Dirk Müller-Wieland, Markus Kollmann, Katharina Marx-Schütt, Martin Berger,
Andreas Puetz, Felix Michels, Luca Leon Happel, Lars Müller, Guido Kobbe,
and Malte Jacobsen

The CONAN study is a prospective, investigator-initiated cohort study Corsano CardioWatch 287-2. This study was designed to evaluate whether continuous wearable monitoring can help detect episodes of acute decompensated heart failure earlier and more reliably in a real-world heart-failure population. The study plans to include 290 patients with clinically confirmed acute decompensated heart failure, who will be followed for 90 days while receiving standard of care plus continuous remote monitoring. Its central objective is to assess whether a self-supervised contrastive learning-derived risk index can identify worsening congestion from continuously recorded physiological time-series data, rather than relying only on intermittent clinic assessments.

Importantly, the patients in this study are specifically monitored with the Corsano CardioWatch 287-2, making Corsano’s device a core part of the study design rather than a peripheral accessory. The enrolled patients receive a registered smartphone together with the CardioWatch 287-2 wrist wearable from Corsano Health, and the wearable is used for continuous recording of vital signs and activity parameters throughout the monitoring period. The platform captures multiple streams of data generated by the Corsano watch, including photoplethysmography-derived signals, heart rate, respiratory rate, temperature, SpO2, RR intervals, sleep-related data, and physical activity, creating a rich longitudinal dataset for algorithmic analysis.

A major strength of the study is that it uses the continuous, multi-parameter data obtained from the Corsano CardioWatch to move beyond rigid single-threshold monitoring. Instead, the study evaluates whether deviations from each patient’s own “regular” physiological patterns can signal impending decompensation. This is particularly relevant in heart failure, where deterioration often develops gradually and may be better captured through subtle changes across several parameters over time. By basing the analysis on data collected passively from the Corsano wrist-worn device in both inpatient and outpatient settings, the study highlights the practical value of unobtrusive continuous monitoring in supporting more proactive heart-failure care.

The publication also makes clear that this is an independent academic study using the Corsano CardioWatch 287-2 in a serious clinical context. In the conclusion and conflict-of-interest section, the authors explicitly note that the trial uses the Corsano wearable, while also stating that Corsano does not provide financial support and has no influence on the writing of the manuscript. That strengthens the credibility of the work and underlines that Corsano’s CardioWatch strength for prospective evaluation in a university-led heart-failure monitoring study focused on clinically meaningful detection of congestion and usability in daily practice.