Clinical readiness of virtual patient models for individualised mechanical ventilation: a systematic literature review

Purpose: Individualised optimisation of mechanical ventilation (MV) remains a challenging task in modern intensive care medicine, which calls for clinical decision support technologies. Model-based decision support systems could help assessing the complex interactions between the mechanical ventilator and the individual patient's pathophysiology, providing thorough mechanistic insight in the dynamic clinical course. These so called computational physiological models (CPMs), or virtual patients models, consequently allow for tailoring MV settings accordingly. To assess the current advancement of CPMs for individualised MV, we critically appraised the current literature on the clinical readiness of these models in the ICU.
Methods: A systematic literature search was conducted on February 13, 2023 in MEDLINE ALL, Embase, Scopus and Web of Science to identify original research articles describing CPMs for individualised MV in the ICU. The modelled physiological phenomena, clinical applications, and clinical level of readiness were extracted.
Results: Out of 6,333 identified unique publications, 149 articles were included. CPMs for individualised MV emerged since the 1970s with increasing levels of readiness (Figure 1). Models incorporated one (n=109, 73%) up to four (n=6, 4%) physiological phenomena in different combinations. 131 articles (88%) modelled lung mechanics, mainly to optimise lung protective ventilation. Gas exchange (n=38, 26%) and gas homeostasis (n=36, 27%) models had mainly applications in controlling oxygenation and ventilation. Models of respiratory muscle function for diaphragm protective ventilation emerged recently (n=3, 2%).
Discussion: CPMs are evolving towards clinical application to address the current main challenges of individualised MV. The majority of CPMs are focused on lung mechanics, to attain lung protective strategies. CPMs to improve oxygenation and ventilation are advancing by integrating models of gas exchange, gas homeostasis and respiratory control. Recently, also models dedicated to respiratory muscle function for diaphragm protective ventilation have emerged.
Conclusion: Computational physiological models are advancing towards clinical application as an explainable tool to optimise individualised MV.
Trial number registration: PROSPERO – CRD42022301715