Data Quality and Critical Events in Ventilation: An Intensive Care study

Mechanical Ventilation

A patient needs to be connected to a ventilator when he cannot breathe. In this work, the mechanical ventilation (artificial) is used. Mechanical ventilation in Intensive Care Units is considered an essential, life-saving therapy for patients with critical illness and respiratory failures (Prevention, 2015). According to Evans (Evans et al., 2005), the ventilators were also developed to generate alarms when a patient becomes disconnected, or the ventilation values are critical. Despite these innovations, the ventilator is not capable of determining if a value is valid or not, i.e., if the patient presents a normal value or a noise value or how critical is the collected value. A valid analysis requires some human observations. However, the humans can only see the collected values, and it is tough to analyse several values in a few minutes. To minimize this problem a set of intelligent procedures, able to detect the data quality, can be defined. These procedures are usually based on the use of intelligent agents (Cardoso et al., 2014). In this field, a set of experiments was made to define and detect critical events in the respiratory system (Portela, Gago, Santos, Silva, & Rua, 2012; Portela, Gago, et al., 2013; Portela et al., 2016).