Voice, Dyspnea and Acute Respiratory Failure

Not Applicable

Withdrawn

• Conditions: Acute Respiratory Diseases
• Interventions: Other: Voice registration

• Registration Number: NCT05340933
• Lead Sponsor: Assistance Publique - Hôpitaux de Paris

Brief Summary

Breathing is an automatic vital function that has the peculiarity of being controllable voluntary for actions other than breathing. Speech production is a characteristic example of use of the respiratory system for nonrespiratory purposes. A healthy respiratory system is necessary for speech to be adequately produced and modulated. In patients with respiratory diseases, it becomes difficult to interfere with an automatic control of breathing that is intensely active to compensate for the respiratory deficience. Speech production is impeded, and, reciprocally, speech can generate dyspnea. This study explores the hypothesis that longitudinal changes in speech characteristics will parallel the clinical evolution of acute respiratory episodes. The aim is to validate such changes as prognostic indicators, in the perspective of future telemedicine applications. The hypothesis tested is that of an association between :

* vocal abnormalities at inclusion (assessed in relation to known data within a normal population (database of holy subjects already constituted) and the initial clinical severity (assessed according to the usual clinical and gasometric criteria):

* the evolution of vocal abnormalities during the stay and the clinical evolution.

Detailed Description

In the conceptual framework describe in the "brief summary" section of this document, this observational longitudinal monocentric study will include consecutive patients admitted in a specialised respiratory medicine ward for acute respiratory episodes. Any such episode will be considered be it "de novo" or complicating an underlying chronic respiratory disease. Vocal recordings will be performed daily, and will be analysed according to standard in the fields. Clinical parameters will also be recorded daily (vital signs, treatment intensity, outcome -including requirement for treatment intensification, transfer to the ICU, death, discharge to rehabilitation facility, discharge to home). The clinical follow-up and the vocal follow-up will be confronted to determine if voice analysis has an intrinsic prognostic value, alone, or in combination with clinical signs.

Study Timeline

• Study First Submitted: 2022-01-10
• Study First Submitted QC: 2022-04-15
• Study First Posted: 2022-04-22
• Status Verified: 2024-09
• Last Update Submitted: 2024-09-30
• Last Update Posted: 2024-10-02
• Study Start: 2025-06-01
• Primary Completion: 2026-02-01
• Study Completion: 2026-09-01

Recruitment & Eligibility

• Status: WITHDRAWN
• Sex: All
• Target Recruitment: 150

Inclusion Criteria

• patients hospitalised in the Pitié-Salpêtrière Pneumology Department with an acute respiratory illness (pneumonia of any cause, COVID pneumonia depending on the epidemic context, COPD decompensation, etc);
• whose condition allows conversational exchanges with the nursing staff within the framework of usual care;
• adults, not protected;
• understand and speak French fluently;
• affiliated to the social security system;
• having read and understood the information leaflet;
• do not object to the use of their data;

Exclusion Criteria

• a clinical condition on admission that is too severe to allow the patient to answer the usual questions of the anamnestic and clinical examination
• patients with uncorrected hearing problems
• patients with neurological, otorhinolaryngological or psychiatric pathology

Study & Design

• Study Type: INTERVENTIONAL
• Study Design: SINGLE_GROUP

Arm && Interventions

Group	Intervention	Description
Arm 1	Voice registration	intervention correspond to the voice registration

Primary Outcome Measures

Name	Time	Method
Characterize voice analysis as a biomarker of respiratory status and its evolution in patients hospitalized in pneumology using machine learning algorithmshospitalized in pneumology	1 month	machine learning algorithms trained on the audio database obtained from patients discussion with medical staff. Voice parameters: respiratory rythms and intensity, and articulatory performances, will be extracted from voice recording, combined and analysed by the algorithms.

Secondary Outcome Measures

Name	Time	Method
Correlation of the used of algorithms based on voice and medical diagnosis.	1 month	Medical diagnosis based on physiological parameters (heart rate (bpm) ; oxygen saturation (%) ; respiratory rate (cycle/min)) will be carried out in the routine care and correlated with the algorythms results.

Trial Locations

Locations (1)

Departement of Respiratory Medicine , Pitié-Salpêtrière Hospital; 🇫🇷 Paris, France