Journal
IFAC PAPERSONLINE
Volume 54, Issue 15, Pages 127-132Publisher
ELSEVIER
DOI: 10.1016/j.ifacol.2021.10.243
Keywords
Mechanical ventilation; PEEP; Respiratory mechanics; Elastance; Prediction; VILI; Basis function; System identification; Virtual patient
Categories
Funding
- NZ Tertiary Education Commission (TEC) fund MedTech CoRE (Centre of Research Excellence) [3705718]
- NZ National Science Challenge 7, Science for Technology and Innovation [2019-S3-CRS]
- EU [872488]
- Marie Curie Actions (MSCA) [872488] Funding Source: Marie Curie Actions (MSCA)
Ask authors/readers for more resources
This research extends a single compartment lung mechanics model with physiologically relevant basis functions to predict patient-specific response to changes in mechanical ventilation care. The prediction outcome is highly accurate and validated with data from volume-controlled and pressure-controlled ventilation trials.
Mechanical ventilation (MV) is used in the intensive care unit (ICU) to treat patients with respiratory failure. However, MV settings are not standardized due to significant inter- and intra- patient variability in response to care, leading to variability in care, outcome, and cost. There is thus a need to personalize MV. This research extends a single compartment lung mechanics model with physiologically relevant basis functions, to identify patient-specific lung mechanics and predict response to changes in MV care. The nonlinear evolution of pulmonary elastance as positive-end-expiratory pressure (PEEP) changes is captured by a physiologically relevant, simplified compensatory equation as a function of PEEP and pressure identification error at the baseline PEEP level. It allows both patient-specific and general prediction of lung elastance of higher PEEP. The prediction outcome is validated with data from two volume-controlled ventilation (VCV) trials and one pressure-controlled ventilation (PCV) trial, where the biggest PEEP prediction interval is a clinically unrealistic 20cmH(2)O, comprising 210 prediction cases over 36 patients (22 VCV; 14 PCV). Predicted absolute peak inspiratory pressure (PIP) errors are within 1.0cmH(2)O and 3.3cmH(2)O for 90% cases in the two VCV trials, while predicted peak inspiratory tidal volume (PIV) errors are within 0.073L for 85% cases in studied PCV trial. The model presented provides a highly accurate, predictive virtual patient model across multiple MV modes and delivery methods, and over clinically unrealistically large changes. Low computational cost, and fast, easy parameterization enable model-based, predictive decision support in real-time to safely personalize and optimize MV care. Copyright (C) 2021 The Authors.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available