Ulf Ludwigs

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

Evaluation of knowledge-based systems differs from that of conventional systems in terms of verification and validation techniques. Furthermore, evaluating medical decision-support systems is difficult because the field is thus far comparatively unexplored. This paper presents an evaluation of a medical knowledge-based system called VentEx that supports decision-making in the management of ventilator therapy. Real patient data from 1300 hours of patient care involving 12 patients with 6 diagnoses are used to validate the knowledge base. The results range from 4.5% to 15.6% disagreement between the setting recommendations produced by VentEx and a gold standard, and 22.2% disagreement for recommendations for weaning. A comparison between the standard and two physicians showed that VentEx produced advice of the same quality as the physicians.

A Placebo-Controlled Experimental Study of Steroid Inhalation Therapy in Ammonia-Induced Lung Injury

BACKGROUND The use of corticosteroids in toxic lung injury caused by exposure to an irritating gas such as ammonia has not been adequately studied. OBJECTIVE To evaluate the effects of budesonide inhalation in a rabbit model of toxic lung injury induced by ammonia. DESIGN Randomized, blind placebo-controlled laboratory investigation employing 16 New Zealand White rabbits. Lung injury was induced by inhalation of a defined amount of aerosolized ammonia. Thirty minutes later, the rabbits were randomized to receive either inhalation therapy with 0.5 mg budesonide or placebo. After another 2 hours, a second treatment inhalation, identical to the first one, was administered. RESULTS Airway pressures, hemodynamics, and gas exchange were measured at baseline, 5, and 15 minutes after ammonia administration and every 30 minutes during a 6-hour period after the first blind inhalation of corticosteroids or placebo. The ammonia inhalation resulted in an acute severe lung injury, detected after 15 minutes as a decrease in Pao2 from 23.3 (+/- 3.6) to 11.0 (+/- 3.6) kPa (p < 0.005) and an increase in peak airway pressure from 13 (+/- 2) to 17 (+/- 2) cm H2O (p < 0.005). During the 6-hour observation period, the blood gas parameters improved gradually in all rabbits. In comparison with placebo, budesonide did not result in improved gas exchange or reduced airway pressure levels during the observation period. CONCLUSION In this animal model corticosteroid inhalation therapy had no effect on ammonia-induced lung injury.

Integrating knowledge-based technology into computer aided ventilation systems

A knowledge-based decision support system for respirator treatment, the KUSIVAR system, has been designed in cooperation between hospital, university and industry. Changes in patient data from respirator and monitoring equipment trigger a computer program that generates advice to the staff concerning e.g. therapy modes and respirator settings using expert systems and process control technology.A prototype has been built on an advanced development workstation, the Unisys Explorer, using the software Knowledge Engineering Environment (KEE). The clinical version is implemented on an Intel 80396-based microcomputer connected on-line via a data-acquisition processor to the respirator. The decision support software is implemented as a module under the Microsoft Windows multitasking environment and communicates with modules for data acquisition, database, handling and data presentation by means of message passing using the Windows Dynamic Data Exchange protocol. The modules present coherent user interfaces by conforming to Microsoft Windows standards.The knowledge base is being extensively validated by an expert group in the ICU and the system will be evaluated through animal experiments and clinical studies.

Prognosis in Patients Presenting with Non-Traumatic Coma

BACKGROUND Studies of patients presenting with coma are limited, and little is known about the prognosis of these cases. OBJECTIVE The aim of this study was to investigate the acute and long-term prognosis after an episode of non-traumatic coma. METHODS Adults admitted consecutively to an emergency department in Stockholm, Sweden between February 2003 and May 2005 with a Glasgow Coma Scale (GCS) score of 10 or below were enrolled prospectively. All available data were used to explore the cause of the impaired consciousness on admission. Patients surviving hospitalization were followed-up for 2 years regarding survival. RESULTS The final study population of 865 patients had the following eight different coma etiologies: poisoning (n = 329), stroke (n = 213), epilepsy (n = 113), circulatory failure (n = 60), infection (n = 56), metabolic disorder (n = 44), respiratory insufficiency (n = 33), and intracranial malignancy (n = 17). The hospital mortality rate among the 865 patients was 26.5%, varying from 0.9% for epilepsy to 71.7% for circulatory failure. The accumulated total 2-year mortality rate was 43.0%, varying from 13.7% for poisoning to 88.2% for malignancy. The level of consciousness on admission also influenced the prognosis: a GCS score of 3-6 was associated with a significantly higher hospital mortality rate than a GCS score of 7-10. CONCLUSION The prognosis in patients presenting with non-traumatic coma is serious and depends largely on both the level of consciousness on admission and the etiology of the coma. Adding the suspected coma etiology to the routine coma grading of these emergencies may more accurately predict their prognosis.

VentEx: an on-line knowledge-based system to support ventilator management

This paper will demonstrate the clinical application of a knowledge-based decision-support system called VentEx for ventilator management. VentEx has been implemented using a knowledge-based development tool on a PC under the Microsoft Windows multitasking environment. It is integrated into a computer aided ventilator system including the Siemens Elema Servo Ventilator 900 C equipped with a Servo Computer Module 990 and the CO2 analyser 930. The system provides advanced ventilator monitoring with expert advice concerning ventilator strategy and settings based on data from on-line monitoring. The knowledge base has been primarily validated and the system has been clinically tested by the intensive care unit staff. Different approaches such as knowledge acquisition, representation and system integration have been outlined and discussed.

KAVE : a tool for knowledge acquisition to support artificial ventilation

A decision support system for artificial ventilation is being developed. One of the fundamental goals for this system is the application of the system when a domain expert is not present. Such a system requires a rich knowledge base. The knowledge acquisition process is often considered to be the bottleneck in acquiring such a complete knowledge base. Since no single available method, for example interviewing domain experts, is sufficient for removing this bottleneck, we have chosen a combination of different methods. The different backgrounds of knowledge engineers and domain experts could cause communication restrictions and difficulties between them, e.g. they might not understand each others knowledge domain and this will affect formulation of the knowledge. To solve this problem we needed a tool which supports both the knowledge engineer and the domain expert already from the initial phase of developing the knowledge base. We have developed a knowledge acquisition system called KAVE to elicit knowledge from domain experts and storing it in the knowledge base. KAVE is based on a domain specific conceptual model which is a result of cooperation between knowledge engineers and domain experts during identification, design and structuring of knowledge for this domain. KAVE includes a patient simulator to help validate knowledge in the knowledge base and a knowledge editor to facilitate refinement and maintenance of the knowledge base.

Multivariable optimization of mechanical ventilation. A linear programming approach

The proposed method aims at improved ventilatory care with reduced morbidity. It combines two important aspects of mechanical ventilation: gas exchange and lung mechanics. A single criterion was selected as optimization index of lung trauma: peak respiratory power (PRP) defined as the maximum product of pressure times flow during inspiration. Arterial blood gases reflect gas exchange and constitute the constraints of the problem. The constraints as well as the optimization index are expressed as linear functions of the input variables (frequency of breathing, tidal volume, and positive end expiratory pressure). A linear programming approach can therefore be used to determine the values of input variables that minimize PRP and at the same time keep arterial blood gases within the prescribed limits. The coefficients of the constraints and the optimization index equation are found by manipulating input variables in order to obtain four different values of PaO2, PaCO2 and PRP (there are four coefficients in each equation). The coefficients can then be calculated and the optimization procedure run. In a pilot study 5 patients suffering from diseases of varying pulmonary pathology were investigated with this method. In 4 out of 5 the ventilator treatment improved in terms of blood gas values (mean increase in PaO2 was 4.7%) and reduction of mechanical load on the lungs (mean PRP reduction was 20%). Lower PRP is accompanied by lower mean power and pressure values, which results in increased cardiac output. Presently, the main problem is the time it takes to determine the patient coefficients (approx one hour), a procedure that needs to be simplified.

Metabolic vs structural coma in the ED—an observational study

BACKGROUND Patients presenting unconscious may reasonably be categorized as suffering from a metabolic or structural condition. STUDY OBJECTIVE The objective was to investigate if some routinely recorded clinical features may help to distinguish between these 2 main forms of coma in the emergency department (ED). METHODS Adults admitted to an ED in Stockholm between February 2003 and May 2005 with a Glasgow Coma Scale (GCS) score less than 11 were enrolled prospectively. The GCS score was entered into a protocol that was complemented with available data within 1 month. RESULTS The study population of 875 patients was classified into 2 main groups: one with a metabolic (n = 633; 72%) and one with a structural disorder (n = 242; 28%). Among the clinical features recorded in the ED, 3 were found to be strongly associated with a metabolic disorder, namely, young age, low or normal blood pressure, and absence of focal signs in the neurological examination. Patients younger than 51 years with a systolic blood pressure less than 151 mm Hg who did not display signs of focal pathology had a probability of 96% for having a metabolic coma. The mean GCS score on admission was identical in the groups. Hospital mortality was 14% in the metabolic and 56% in the structural group. CONCLUSIONS These findings indicate that unconscious young adults who present without a traumatic incident with a low or normal blood pressure and without signs of focal pathology most probably suffer from a metabolic disorder, wherefore computed tomography of the brain may be postponed and often avoided.

Pulmonary Epithelial Permeability and Gas Exchange: A Comparison of Inverse Ratio Ventilation and Conventional Mechanical Ventilation in Oleic Acid-Induced Lung Injury in Rabbits

STUDY OBJECTIVE (1) To explore the interaction between mechanical ventilation and oleic acid (OA)-induced lung injury on indexes of pulmonary gas exchange and epithelial permeability, and (2) to compare this interaction using two different modes of ventilation: pressure-controlled inverse ratio ventilation (PCIRV) and volume-controlled ventilation with positive end-expiratory pressure (VCV PEEP). DESIGN Randomized animal study. SETTING Experimental laboratory investigation at Södersjukhuset, Stockholm, Sweden. ANIMALS Twenty-four New Zealand white rabbits. INTERVENTIONS (1) Ventilation with PCIRV (n=6) or VCV PEEP (n=6) for 6 h at equal end-expiratory alveolar pressure levels of 5 cm H2O followed by induction of lung injury (IV injection of OA 0.15 mL/kg). (2) Induction of lung injury followed by 6 h of ventilation with either PCIRV (n=6) or VCV PEEP (n=6) as described above. MEASUREMENTS AND RESULTS Lung mechanics, heart rate, BP, and gas exchange results were equal at baseline. In group A, after 1 h of ventilation, mean airway pressure was 11.9+/-4.4 with PCIRV and 8.3+/-1.0 cm H2O with VCV PEEP (p<0.05). Forty minutes after OA injection, PaO2/fraction of inspired oxygen (FIO2) was 24+/-10 kPa with PCIRV and 44+/-15 kPa with VCV PEEP (p<0.05). Mean airway pressure was higher and peak airway pressure was lower with PCIRV. In group B, after 6 h of ventilation, PaO2/FIO2 was 17+/-5 kPa with PCIRV and 43+/-8 kPa with VCV PEEP (p<0.01). Systemic BP was lower with PCIRV and mean airway pressure was higher. Technetium-99m diethylene triamine penta-acetic acid lung clearance: In group A, curves were monoexponential with both PCIRV (half-life time [T 1/2], 21+/-8 min and VCV PEEP (T 1/2, 126+/-59 min, p<0.005) until injection of OA. In the VCV PEEP-treated animals, a marked increase in clearance rate was observed within 60 s of OA injection (T 1/2, 13+/-9 min, p<0.001). Fifteen minutes after OA injections, T 1/2 had decreased to 38+/-17 min with VCV PEEP. In the animals treated with PCIRV, OA injection did not lead to a significant change in clearance rate, although the elimination pattern was observed to change from single-compartment to multicompartment type. In group B, clearance curves were monoexponential with both ventilatory modes. There was no significant difference in clearance rate between PCIRV (T 1/2, 25+/-9 min) and VCV PEEP (T 1/2, 36+/-16 min, not significant). CONCLUSIONS The observation that PaO2 was lower in the PCIRV-treated groups must be interpreted with caution in this animal study with relatively few observations. The finding may reflect differences in the effect of OA injection in the two ventilatory modes. It is also possible that externally applied PEEP is more effective than PCIRV in increasing oxygen tension, either because of a less inhomogenous distribution of ventilation and perfusion or for other reasons. The clearance results imply that PCIRV causes an alteration in lung epithelial or membrane function in comparison to VCV PEEP. This functional difference is most likely caused by the large time-weighted lung volume produced by pressure control in combination with a prolonged inspiration. Induction of high permeability lung injury with OA eliminates the difference between PCIRV and VCV PEEP. It remains to be established whether these findings are relevant with regard to ventilator-associated structural lung injury in man.

Knowledge base design for decision support in respirator therapy

A knowledge base is built for decision support applied to respirator therapy (the KUSIVAR project). The knowledge representation is object-oriented using frames to store multiple forms of knowledge: variable descriptions, transformation tables, rules and methematical models.The system is data-driven, generating and displaying advice automatically triggered by changes in data from the respirator and the patient. The inferenceing mechanism is forward-chaining i.e. a rule is evaluated as soon as its condition is satisfied. Temporal aspects of the reasoning are represented by a number of mechanisms, among others limited validity times for data, trend analysis and mathematical models.The knowledge base is organized according to disease groups and decision situation which simplifies knowledge acquisition and improves response times since it enables the system to focus on a limited set of rules in each situation.To test the feasability of the system design a prototype has been built using Knowledge Engineering Environment (KEE) from Intellicorp on an Explorer workstation from Unisys. The production system, which is interfaced to a Siemens Elema Servo Ventilator 900C, is currently being implemented under the Microsoft Windows multitasking environment on a microcomputer based on an Intel 80386 processor.