Heart disease is the leading cause of death in the United States. Though there are several types of heart disease, one of the main ways the heart actually dies is through congestive heart failure, when the heart fails to pump effectively. Ineffective pumping causes blood and other fluids to build up throughout the body. Eventually, the fluid buildup may keep the heart from functioning at all, a condition called acute decompensated heart failure. This is an emergency condition, and without rapid medical treatment quickly leads to death. Fortunately, with early intervention, several patients can recover – at least temporarily. Clearly, one of the keys to saving people is early intervention.

The vocal cords and lungs may be affected by congestive heart failure. As these fill with fluid, there are changes in how the voice sounds. These changes are difficult for people to hear, but technology might be able to help out. One type of technology looking to tackle the topic is a deep neural network called HearO.

Deep neural networks are a subset of artificial intelligence. These systems learn how to make predictions from examples. The HearO system, made by the company Cardio Medical, analyzed the voices of people with congestive heart failure. The system learned by comparing people’s voices while they were “wet” (fluid-filled while in the hospital) and “dry” (after hospital treatment and discharge). Using this, the system learned to detect voice differences in the severity of the condition.

The HearO system has now been packaged as a smartphone app. Patients talk into the app every day and it compares their voice to itself. It scans for changes that indicate a fluid build-up (and danger!). The hope is that HearO will help detect changes in the voice before acute decompensated heart failure occurs. Clinical trials are currently underway to test the HearO’s effectiveness and some of our ENCORE Research Group sites are enrolling for this. 

Written by: Benton Lowey-Ball, B.S. Behavioral Neuroscience

Sources:

Amir, O., Abraham, W. T., Azzam, Z. S., Berger, G., Anker, S. D., Pinney, S. P., … & Edelman, E. R. (2022). Remote speech analysis in the evaluation of hospitalized patients with acute decompensated heart failure. Heart Failure, 10(1), 41-49.

Brooks, M. (2015, December 15). Smartphone app could give voice to congestion in heart failure. Medscape. https://www.cordio-med.com/wp-content/uploads/2021/12/smartphone-app-could-give-voice-to-congestionin-heart-failure.pdf

Tomov, N. S., & Tomov, S. (2018). On deep neural networks for detecting heart disease. arXiv preprint arXiv:1808.07168.