CMAAO Coronavirus Facts And Myth Buster: Future of lung function screening |
Editorial
eMediNexus Coverage from: 
CMAAO Coronavirus Facts And Myth Buster: Future of lung function screening
Dr KK Aggarwal,  29 October 2020
Coronavirus Live Count Map India

remove_red_eye 1547 Views
#Multispeciality

3 Read Comments                

With input from Dr Monica Vasudev

1120: Voice-Based Screening Predicts Lung Function

  1. A preliminary study has revealed that novel voice and breath analysis using a smartphone app could be a useful measure of lung function.
  2. It may serve as a valuable tool for the identification and monitoring of respiratory disease.
  3. Automated voice and breath analysis was shown to be effective for predicting lung function, with an 82% accuracy for predicting patients with and without obstructive lung disease.
  4. The study was presented recently at the virtual CHEST conference.
  5. The ongoing, prospective, cross-sectional study included 128 initial participants (76 women, 52 men), enrolled during appointments for pulmonary function testing conducted at Allegheny General Hospital in Pittsburgh. Around 16.4% had lung obstruction.
  6. A voice collector app collected voice data. Participants read a phonetically-balanced passage with 199 words to optimize findings. It was later reduced to around 50 words.
  7. Voice and breath sound samples were recorded prior to and after pulmonary function testing, which corresponded with pre-and post-bronchodilator samples.
  8. Pre- and post-pulmonary function test forced expiratory volume in one second (FEV1) and forced vital capacity (FVC) results were obtained from each patient. Voice and breath audio samples were recorded on a smart tablet, with the help of the proprietary software app. The voice data were analyzed using cloud-based software. Voice audio recordings were calibrated to create a customized noise profile.
  9. The phonetically balanced reading passage was used to assess respiration, phonation, articulation, and resonance. A long vowel word list helped detect speaking-related dyspnea during articulation of long vowel sounds.
  10. Machine learning compared the voice-based screening to spirometry data.
  11. The automated voice analysis yielded good diagnostic accuracy for the prediction of FEV1 and FVC.
  12. Obstruction classification demonstrated an accuracy of 98%, and a sensitivity of 96%.

Source: Ashraf O, et al "Voice-based screening and monitoring of chronic respiratory conditions" CHEST 2020.

 

Dr KK Aggarwal

President CMAAO, HCFI and Past National President IMA

To comment on this article,
create a free account.
Sign Up to instantly read 30000+ free Articles & 1000+ Case Studies
Create Account

Already registered?

Login Now