Modeling Surgical Patients? Tidal Volumes Set by Anesthesiologists-a Successful Regression Approach Based on 6 Body Parameters

During mechanical ventilation for a patient under surgery, the stroke volume of the ventilator needs to be set individually for adequate gaseous exchange. Such value of stroke volume (or Tidal Volume, V-T) is initially set by attending anesthesiologists, based primarily on their subjective experience. Depending on the patient's condition and surgical procedure, the initial V-T is often re-adjusted to a final V-T which is then used throughout the surgery. To provide more objective and rapid estimates of V-T, we here applied an approach with the flavor of artificial intelligence. Specifically, we first collected a dataset of 288 final V-T used clinically by 7 anesthesiologists, and then modeled V-T of individual patients based on their 6 body parameters (viz., body weight, height, chest circumference, sex, body mass index and age). Through linear regression, V-T was modeled rather satisfactorily (corresponding to an accuracy of 95%). Our results provided a more precise estimator and a major time-saver for anesthesiologists to set V-T of subjects, given only their 6 body parameters. Our study is especially valuable in this era of Covid-19 pandemic when mechanical ventilators are typically required for those patients seriously infected.