4380Vocal biomarker predicts long term survival among heart failure patients

Abstract

Home telemedicine holds the potential to reduce costs and improve outcome among patients with congestive heart failure (CHF). We recently reported a possible relationship between vocal biomarkers and coronary artery disease. The purpose of the current study was to test the hypothesis that vocal biomarkers are associated with long term survival among CHF patients. Method(s): Study cohort included 2,282 patients who were registered to a large HMO's call center of patients suffering from chronic conditions including CHF. 940 acoustic features were extracted from 20 secs of speech for each of the patients. A biomarker was then developed based on a train cohort (N=1,391) using machine learning techniques optimized according to an age cutoff of 75. The biomarker was then tested on a final mutually exclusive CHF study cohort (N=891). The biomarker was evaluated as a continuous variable, as well as dichotomized to two groups: high (upper quartile) and low. Result(s): Mean age of the CHF study population was 74+/-12, 67% were men. During an average follow-up of 32+/-15 months, 246 (28%) patients died. Kaplan Meier survival analysis showed significantly higher cumulative probability of death among subjects in the higher vocal biomarker group (41%+/-19% vs. 30%+/-20%, p<0.001; FIGURE). Consistently, cox regression analysis of the vocal biomarker as a continuous variable in the model, and with adjustment to known predictors of poor survival, demonstrated that each standard deviation increase in the vocal biomarker was associated with a significant 15% increased risk of death during follow-up (95% CI 1.02 - 1.3, p=0.02). Conclusion(s): This is the first study to document a relationship between a vocal biomarker and long-term survival among CHF patients. The results have important clinical implications for telemedicine and CHF patient care. (Figure Presented).