Abstract

Background and Aims:
In intensive care units (ICUs), clinicians use various mathematical models to assess and monitor respiratory illness. However, evidence suggests that some pulmonary internal and external physiolog¬ical factors are not integrated in the calculation of these models. This research aimed to demonstrate the variation in these models within a short period in critical care settings.

Methods:
A retrospective cohort study was conducted from 2017 to 2019 with adults admitted to an ICU and mechanically ventilated for more than 24 hours. Two different arterial blood gas samples were selected within a mean of 120 minutes. The coefficient of variation percent [(CV%)=SD/Mean] was calculated.

Results:
In total, 211 intubated ICU patients were included. The mean age was 57 (SD +/- 19.62) years, and 33% were female. The degree of variation of the average CV% for all the mathematical models was 14%-30%. The mean CV% for the arterial oxygen tension to inspired oxygen fraction (PaO2/FiO2) and arterial/alveolar oxygen tension ratio (PaO2/ PAO2) was 14%, which is a low variation compared to the respiratory index (RI) for which the mean CV% was 30%.

Conclusions:
The findings of this study indicate that most of the existing oxygen mathematical models used in clinical settings lack physiological realism, necessitating a new mathematical model that integrate both assess illness severity and enhance other models’ accuracy.

Key words: Mathematical model - Pulmonary oxygen indices - Respiratory index - Oxygen index - Oxygen sat¬uration index.