BIOMARKERS TO DETECT ALTERATIONS OF THE CLINICAL CONTROL OF ASTHMA USING AUDIO SIGNALS

• Grant number: 25/01809-5
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: May 01, 2025
• End date: April 30, 2026
• Field of knowledge: Health Sciences - Physiotherapy and Occupational Therapy
• Principal Investigator: Celso Ricardo Fernandes de Carvalho
• Grantee: Mariana Lopes Pestana
• Host Institution: Faculdade de Medicina (FM). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Associated research grant: 23/00488-5 - SPIRA-BM: biomarkers for respiratory conditions on mobile devices using audio analysis with artificial intelligence, AP.TEM

Abstract

Introduction: Asthma is a very prevalent disease in Brazil and the lack of its clinical control is among the main factors that cause people to seek emergency care and lose quality of life. Aim: To develop biomarkers using audio signals (voice and speech) to detect the lack of clinical control in patients with moderate to severe asthma. Methods: Seventy-two patients with moderate to severe asthma will be assessed during a routine outpatient medical visit, regardless of the level of asthma control. Patients will be divided into two groups: asthma with (ACC) and without clinical control (ASC). Asthma control will be assessed using the Asthma Control Questionnaire (ACQ). A group of people without asthma will be matched by age, sex and body mass index (BMI) and will be included as a control group (GSA). Patients and people without asthma will be asked to participate in the study. If they accept, a recording will be made for audio recordings of three components: a sustained vowel, a sentence and a nursery rhyme. The recording will be made on a cell phone obtained with resources from the FAPESP thematic project and with the application "System for Early Detection of Respiratory Failure by Audio Analysis" (SPIRA), developed especially for this project. Analyses will be carried out using small and big data. The small data analysis will include voice variables (fundamental frequency, standard deviation of fundamental frequency, jitter, shimmer and signal-to-noise ratio) and speech variables (total sentence duration, number of pauses and duration of pauses); while the big data analysis will be carried out using artificial intelligence by researchers from the SPIRA project. Seventy-two patients will be assessed and the sample calculation was made based on a voice variable (fundamental frequency, fo; in Hz) and a speech variable (total sentence duration, sec).