Do VoIce Changes Predict Heart Failure

There are two specific aims for this first phase of the project:

1. Collect short-term acoustic and accelerometric data on groups of patients with volume overload before and after successful diuresis of amounts of fluid that are large enough to permit robust statistical testing of derived measures.
2. Determine whether direct measures (based on previously-developed approaches for acoustic and ACC-based monitoring of vocal function can differentiate between hypervolemic and optivolemic patterns.

2.2 Justification for Clinical Study The study will evaluate if acoustic and accelerometric measures of voice and speech production have the sensitivity to detect a clinically significant change in lung fluid status for patients with confirmed acute heart failure syndrome with pulmonary congestion. Changes in lung fluid in acute heart failure syndrome are difficult to diagnose and evaluate. A blinded observational methodology was chosen as the most appropriate study design to understand changes in lung fluid values in patients with acute heart failure syndrome with pulmonary congestion. The primary endpoint is significant correlation of changes in voice and speech characteristics with change in total body water as measured by change in weight during the hospital admission.

2.3 Measurement Precision Exploratory endpoints will be to determine the distributions of lung fluid values for patients with confirmed acute heart failure syndrome and to determine if acoustic and accelerometer-based measurements have sufficient sensitivity to detect changes in lung fluid status in a symptomatic hospitalized population. The establishment of the underlying etiology of the heart failure will help determine how it may contribute to symptomology and changes in acoustic and accelerometric measures of voice and speech production. The types of analyses that can be performed on the airborne acoustic (microphone) signal are somewhat limited by the levels of background noise in the recording environment, particularly with respect to some of the acoustic measures related to voice quality. For this reason it is proposed that recordings of voice/speech production for this pilot study be obtained simultaneously with both a microphone and a neck-placed miniature accelerometer (ACC). The ACC is relatively immune to environmental noise, and we have been able to extract some voice quality-related measures from the ACC signal. The ACC is easily placed on the anterior neck just above the clavicular notch using double-sided medical grade adhesive tape (see Figure 1). The recording protocol will be comprised of: 1) sustained vowels, 2) standard sentences and a reading passage, and 3) 30 seconds of spontaneous speech. Recordings will be made before, during, and after an acute episode of Congestive Heart Failure (CHF). Each recording session should take less than 10 minutes.

3 STUDY PROTOCOL 3.1 Study Design and Objective - see above