Idoven is a health technology company advancing early detection and precision medicine for cardiovascular diseases. Idoven’s AI platform delivers substantial improvements to the speed, consistency and accuracy of electrocardiogram interpretation. Idoven partners with leading medical device and pharmaceutical companies on AI-driven innovations to develop a new standard of cardiovascular care.
There are 6 ways Idoven’s AI platform represents a breakthrough in diagnostic technology: - Most comprehensive diagnostic solution: Willem standardises and automates the identification of over 80 of the most frequent cardiac patterns, affecting 90% of the global population. - Cardiologist-level accuracy: Willem aids clincians to make an accurate diagnosis for every patient, every time, from anywhere (in the hospital, at home or on the move). Willem operates at 90% global accuracy. - Prediction of atrial fibrillation: Our AI models can predict the evolution of atrial fibrillation for a patient within the next 6 months, even for patients with no prior history, with just physiological data. - Device-neutral approach: Our AI platform can read both proprietary and standard ECG signal format (JSON, DICOM, EDF, CSV among others). - Highest safeguards for patient data: We are certified for the highest internationally-recognized standards for information security management systems (ISO 27001, 27017, 27018, 27701). Our analysis is performed on de-identified studies and anonymized/pseudo anonymized patient data. - Delivery in seconds/minutes, not hours/days: Willem can deliver a diagnosis in: real time for short-duration ECGs (vs. 1-5 minutes for a cardiologist, if available); less than 3 minutes for 24-hour ambulatory ECGs (vs. 20-40 min for a cardiologist, with an average waiting list of 1 to 3 months); and less than 10 minutes for 7-21 days long-duration ambulatory ECGs (vs. 2-8 hours for a cardiologist, with an average waiting list of 1 to 3 months).
Cardiovascular disease is expected to be a €1 trillion public health burden by 2035. Managing this public health challenge starts with early diagnosis, and the most ubiquitous point-of-care test to detect heart problems is the ECG. However, today’s systems are not sufficient to keep up with the demand for ECG interpretation and current procedures can be fairly subjective, leading to wide variations in interpretation and treatment decisions, and delays in clinical trials. Moreover, with the use of wearables and other biosensors rapidly accelerating, new approaches to analyse data at unprecedented speed and scale. Globally, there are 300 million ECGs performed per year, growing at 6% annually, and the ECG interpretation market by 2023 is estimated at €11.2 billion. This results in over 1 million hours spent by physicians just in Europe every day analysing patients’ ECG data to diagnose arrhythmias and other heart conditions. With the use of wearables and other biosensors rapidly accelerating, the healthcare industry requires new approaches to analyse data at unprecedented speed and scale. Non-AI based software has progressed but still requires significant human intervention to reliably diagnose and produce reports. Willem analyses ECG data from any-lead wearable, mobile and implantable device, in an automatic and standardised way, augmenting both cardiologist and non-cardiology experts clinician’s ability to identify, triage and diagnose patients at scale.
Our software has learned cardiology from more than 49,000 patients and 1,2 million hours of real heart records. Our deep neural network model has over 400 billion nodes, larger than some of the world’s largest AI models on natural language recognition. Through exclusive partnerships and research collaborations with world-class research institutions, Idoven has developed one of the world’s largest ECG databases for AI development and scientific innovations, of both asymptomatic individuals and patients, that has been structured and labeled by cardiologist experts for the sole purpose of training state-of-the-art AI algorithms to see patterns in ECGs that humans cannot. We have also developed the first of its kind predictive ECG-based AI models as prognostic tools. Our first model identifies patients that are at increased risk of developing persistent atrial fibrillation within the next 6 months. We have 4 more prediction algorithms in the pipeline.
