To cite this version:Michel Dojat, Laurent Brochard. Knowledge-based systems for automatic ventilatory management.. Respiratory Care Clinics, 2001, 7 (3), pp.379-96, viii.
IntroductionDespite their enormous potential to facilitate bedside management, the practical role of computers in critical care environments is generally restricted to the storage and the retrieval of data coming from electronic medical devices and hospital information networks. Benefits of the use of computers in health care may be extended by the design of computerized medical assistants that can efficiently discharge the clinical staff of repetitive tasks (which, in practice, often are not performed) and, importantly, help practitioners to make efficient decisions in time. In intensive care and anesthesia, the demand for computerized medical assistants is potentially considerable, in order to filter and synthesize the growing mass of clinical parameters and information available. The progressive introduction of computerized protocols has been proposed to standardize the bedside decision making process for mechanical ventilation and to reduce unnecessary variation among practitioners 32 , reinforcing the potential impact of computerized medical assistants. Designing such assistants for intelligent monitoring, diagnosis and therapy planning tasks in Intensive Care and Anesthesia is a challenging goal that requires the modeling of several levels of knowledge ranging from low level data interpretation to high level cognitive tasks, such as planning (for a review of recent research work in this field see 12,25 ). Our goal in this special issue is to show how knowledge-based computerized assistants can be used to practically improve the closed-loop control of mechanical ventilation.
Control and Planning: Two Key Points for Automatic Ventilation
Management
3Modern methods of mechanical ventilation partially assist the patient's ventilation by adding a variable amount of mechanical support to his/her spontaneous activity. In this context, since the needs of the patient are evolutive, it is essential to continuously control the ventilatory support, in order to avoid excessive work of breathing and effort, discomfort and dyspnea on the one hand, or excessive support, hyperinflation and dyssynchrony on the other hand. In parallel to this ideal automatic adaptation, it may be necessary to plan the long term adaptation of the therapy according to specific medical goals. For instance, it may be indicated to gradually decrease the level of assistance in order to facilitate the weaning from the ventilator or to take into account large variations of physiological needs during the patient wake-up from anesthesia or drug intoxication.Planning and control are two different tasks that have a common goal: choosing actions over time to influence a process, based on some model of that process 9 . Control is a local task to determine what to do the next instant. Planning is a strategic task to regulate the process evolution. For control and plann...