Login / Signup

A Pervasive Pulmonary Function Estimation System with Six-Minute Walking Test.

Ming-Feng WuChi-Min TengTz-Hau KuoWei-Chang HuangChih-Yu Wen
Published in: Biosensors (2022)
Self-monitoring for spirometry is beneficial to assess the progression of lung disease and the effect of pulmonary rehabilitation. However, home spirometry fails to meet both accuracy and repeatability criteria in a satisfactory manner. The study aimed to propose a pervasive spirometry estimation system with the six-minute walking test (6MWT), where the system with information management, communication protocol, predictive algorithms, and a wrist-worn device, was developed for pulmonary function. A total of 60 subjects suffering from respiratory diseases aged from 25 to 90 were enrolled in the study. Pulmonary function test, walking steps, and physical status were measured before and after performing the 6MWT. The significant variables were extracted to predict per step distance (PSD), forced vital capacity (FVC) and forced expiratory volume in one second (FEV 1 ). These predicted formulas were then implemented in a wrist-worn device of the proposed pervasive estimation system. The predicted models of PSD, and FVC, FEV 1 with the 6MWT were created. The estimated difference for PSD was-0.7 ± 9.7 (cm). FVC and FEV 1 before performing 6MWT were 0.2 ± 0.6 (L) and 0.1 ± 0.6 (L), respectively, and with a sensitivity (Sn) of 81.8%, a specificity (Sp) of 63.2% for obstructive lung diseases, while FVC and FEV 1 after performing the 6MWT were 0.2 ± 0.7 (L) and 0.1 ± 0.6 (L), respectively, with an Sn of 90.9% and an Sp of 63.2% for obstructive lung diseases. Furthermore, the developed wristband prototype of the pulmonary function estimation system was demonstrated to provide effective self-estimation. The proposed system, consisting of hardware, application and algorithms was shown to provide pervasive assessment of the pulmonary function status with the 6MWT. This is a potential tool for self-estimation on FVC and FEV 1 for those who cannot conduct home-based spirometry.
Keyphrases
  • lung function
  • machine learning
  • healthcare
  • deep learning
  • randomized controlled trial
  • physical activity
  • pulmonary hypertension
  • mental health
  • social media
  • health information