Remote Speech Analysis in the Evaluation of Hospitalized Patients With Acute Decompensated Heart Failure

• DOI: 10.1016/j.jchf.2021.08.008
• Publication Date: 2021-12-08T20:45:41Z
• Authors: Offer Amir; William T. Abraham; Zaher S. Azzam; Gidon Berger; Stefan D. Anker; Sean P. Pinney; Daniel Burkhoff; Ilan D. Shallom; Chaim Lotan; Elazer R. Edelman
• Keywords: Decompensation

This study assessed the performance of an automated speech analysis technology in detecting pulmonary fluid overload patients with acute decompensated heart failure (ADHF).Pulmonary edema is main cause (HF)-related hospitalizations and a key predictor poor postdischarge prognosis. Frequent monitoring often recommended, but signs decompensation are missed. Voice sound technologies have been shown to successfully identify clinical conditions that affect vocal cord vibration mechanics.Adult ADHF (n = 40) recorded 5 sentences, 1 3 languages, using HearO, proprietary processing application, upon admission (wet) discharge (dry) from hospital. Recordings were analyzed for distinct measures (SMs), each time, frequency resolution, linear versus perceptual (ear) model; mean change baseline SMs was calculated.In total, 1,484 recordings analyzed. Discharge tagged as distinctly different 94% cases, differences all 87.5% cases. The largest documented SM2 (218%). Unsupervised, blinded clustering untagged 9 further demonstrated SMs.Automated can voice alterations reflective HF status. platform expected provide valuable contribution in-person remote follow-up HF, by alerting imminent deterioration, thereby reducing hospitalization rates. (Clinical Evaluation Cordio App Adult Patients With CHF; NCT03266029).