Evaluation of a knowledge-based decision-support system for ventilator therapy management

A model-based decision support system for critiquing mechanical ventilation treatments

A computerized system for critiquing mechanical ventilation treatments is presented that can be used as an aide to the intensivist. The presented system is based on the physiological model of the subject's respiratory system. It uses modified versions of previously developed models of adult and neonatal respiratory systems to simulate the effects of different ventilator treatments on the patient's blood gases. The physiological models that have been used for research and teaching purposes by many researchers in the field include lungs, body tissue, and the brain tissue. The lung volume is continuously time-varying and the effects of shunt in the lung, changes in cardiac output and cerebral blood flow, and the arterial transport delays are included in the system. Evaluation tests were done on adult and neonate patients with different diagnoses. In both groups combined, the differences between the arterial partial pressures of CO(2) predicted by the system and the experimental values were 1.86 ± 1.6 mmHg (mean ± SD), and the differences between the predicted arterial hemoglobin oxygen saturation values, S(aO2), and the experimental values measured by using pulse oximetry, S(pO2), were 0.032 ± 0.02 (mean ± SD). The proposed system has the potential to be used alone or in combination with other decision support systems to set ventilation parameters and optimize treatment for patients on mechanical ventilation.

Automated mechanical ventilation: adapting decision making to different disease states

The purpose of the present study is to introduce a novel methodology for adapting and upgrading decision-making strategies concerning mechanical ventilation with respect to different disease states into our fuzzy-based expert system, AUTOPILOT-BT. The special features are: (1) Extraction of clinical knowledge in analogy to the daily routine. (2) An automated process to obtain the required information and to create fuzzy sets. (3) The controller employs the derived fuzzy rules to achieve the desired ventilation status. For demonstration this study focuses exclusively on the control of arterial CO(2) partial pressure (p(a)CO(2)). Clinical knowledge from 61 anesthesiologists was acquired using a questionnaire from which different disease-specific fuzzy sets were generated to control p(a)CO(2). For both, patients with healthy lung and with acute respiratory distress syndrome (ARDS) the fuzzy sets show different shapes. The fuzzy set "normal", i.e., "target p(a)CO(2) area", ranges from 35 to 39 mmHg for healthy lungs and from 39 to 43 mmHg for ARDS lungs. With the new fuzzy sets our AUTOPILOT-BT reaches the target p(a)CO(2) within maximal three consecutive changes of ventilator settings. Thus, clinical knowledge can be extended, updated, and the resulting mechanical ventilation therapies can be individually adapted, analyzed, and evaluated.

Intelligent decision support systems for mechanical ventilation

OBJECTIVE

An overview of different methodologies used in various intelligent decision support systems (IDSSs) for mechanical ventilation is provided. The applications of the techniques are compared in view of today's intensive care unit (ICU) requirements.

METHODS

Information available in the literature is utilized to provide a methodological review of different systems.

RESULTS

Comparisons are made of different systems developed for specific ventilation modes as well as those intended for use in wider applications. The inputs and the optimized parameters of different systems are discussed and rule-based systems are compared to model-based techniques. The knowledge-based systems used for closed-loop control of weaning from mechanical ventilation are also described. Finally, in view of increasing trend towards automation of mechanical ventilation, the potential utility of intelligent advisory systems for this purpose is discussed.

CONCLUSIONS

IDSSs for mechanical ventilation can be quite helpful to clinicians in today's ICU settings. To be useful, such systems should be designed to be effective, safe, and easy to use at patient's bedside. In particular, these systems must be capable of noise removal, artifact detection and effective validation of data. Systems that can also be adapted for closed-loop control/weaning of patients at the discretion of the clinician, may have a higher potential for use in the future.

SIVA: A Hybrid Knowledge-and-Model-Based Advisory System for Intensive Care Ventilators

The Sheffield Intelligent Ventilator Advisor is a hybrid knowledge-and-model-based advisory system designed for intensive care ventilator management. It consists of a top-level fuzzy rule-based module to give the qualitative component of the advice, and a lower-level model-based module to give the quantitative component of the advice. It is structured to offer adaptive patient-specific decision support. It can be operated in either invasive or noninvasive modes depending on the availability of data from invasive clinical measurements. The user can choose between the full-advisory mode and the clinician-directed mode. The advice given by the top-level module has been validated against retrospective real patient data and compared with intensivists expertise and performance under simulation conditions. Closed-loop simulations were performed assuming various clinical scenarios including sudden changes in the patient parameters such as the shunt or deadspace with noise and disturbances. They have shown that the advice given was appropriate and the blood gases resulting from the closed-loop decision support were acceptable. The system was also shown to be tolerant to noise and disturbances. It is implemented in MATLAB/SIMULINK and LabVIEW.

NéoGanesh: a working system for the automated control of assisted ventilation in ICUs

Automating the control of therapy administered to a patient requires systems which integrate the knowledge of experienced physicians. This paper describes NéoGanesh, a knowledge-based system which controls, in closed-loop, the mechanical assistance provided to patients hospitalized in intensive care units. We report on how new advances in knowledge representation techniques have been used to model medical expertise. The clinical evaluation shows that such a system relieves the medical staff of routine tasks, improves patient care, and efficiently supports medical decisions regarding weaning. To be able to work in closed-loop and to be tested in real medical situations, NéoGanesh deals with a voluntarily limited problem. However, embedded in a powerful distributed environment, it is intended to support future extensions and refinements and to support reuse of knowledge bases.

Building intelligent alarm systems by combining mathematical models and inductive machine learning techniques

In this article a technique is described to develop knowledge-based alarm systems for ventilator therapy, using mathematical modeling and machine learning. With a mathematical model airway pressure, expiratory gas flow and CO2 concentration at the endotracheal tube are simulated for patients, undergoing volume-controlled ventilation with constant ventilator settings, during normal functioning of the breathing circuit and during breathing circuit mishaps (leaks and obstructions). Simulations were performed for 94 physiologically different 'patients', by varying airway resistance and lung/thorax compliance values in the model. Each simulated breath was described by a set of derived signal features and a label that constituted during which event (normal function or mishap) the breath was recorded. With an inductive machine learning algorithm rules, linking signal feature values to breathing circuit events, were created from data of 54 of the simulated patients. The resulting set of rules was able to classify 99% of events in the data of the remaining 40 patients correctly. Of signals, measured at a ventilated lung simulator, 100% of events were classified correctly.