Human versus Computer Controlled Selection of Ventilator Settings: An Evaluation of Adaptive Support Ventilation and Mid-Frequency Ventilation

摘要

Background. There are modes of mechanical ventilation that can select ventilator settings with computer controlled algorithms (targeting schemes). Two examples are adaptive support ventilation (ASV) and mid-frequency ventilation (MFV). We studied how different clinician-chosen ventilator settings are from these computer algorithms under different scenarios. Methods. A survey of critical care clinicians provided reference ventilator settings for a 70 kg paralyzed patient in five clinical/physiological scenarios. The survey-derived values for minute ventilation and minute alveolar ventilation were used as goals for ASV and MFV, respectively. A lung simulator programmed with each scenario's respiratory system characteristics was ventilated using the clinician, ASV, and MFV settings. Results. Tidal volumes ranged from 6.1 to 8.3 mL/kg for the clinician, 6.7 to 11.9 mL/kg for ASV, and 3.5 to 9.9 mL/kg for MFV. Inspiratory pressures were lower for ASV and MFV. Clinician-selected tidal volumes were similar to the ASV settings for all scenarios except for asthma, in which the tidal volumes were larger for ASV and MFV. MFV delivered the same alveolar minute ventilation with higher end expiratory and lower end inspiratory volumes. Conclusions. There are differences and similarities among initial ventilator settings selected by humans and computers for various clinical scenarios. The ventilation outcomes are the result of the lung physiological characteristics and their interaction with the targeting scheme.