Search

Summary: Objective: Patients emerging from cardiac surgery can display varying degrees of cardiovascular instability arising from potentially complex, multi-factorial and interlinked causes. Stabilization and control of the cardiovascular system are currently managed by healthcare experts using experiential knowledge, and, in some centers, manually inputted decision pathway algorithms. This paper describes a clinical trial undertaken to determine the basic functioning of a clinical decision support system (CDSS) designed and constructed by the authors to facilitate the control of the major cardiovascular components in the early post-operative phase. Part II follows Part I''s description of the software and simulation testing of the CDSS, and describes the hardware setup of a patient monitoring and CDSS. The system is evaluated on three post-cardiac surgery intensive care patients whom had all undergone cardio-pulmonary bypass. Methods: The study was approved by the Sheffield Teaching Hospitals National Health Service (NHS) Foundation Trust Research Ethics Committee and conducted at the North Trent cardio-thoracic surgical unit and cardiac intensive care unit (CICU), Northern General Hospital, Sheffield (UK). Patients considered as ‘very likely’ to require active intervention to support the cardiovascular function following routine cardiac surgery were recruited during pre-operative surgical and anesthetic assessment, giving written informed consent when admitted for their operation. These patients underwent routine induction and maintenance of anesthesia by a non-study consultant anesthetist and the operation performed. There were no restrictions placed on the types of invasive monitoring used, on the use of trans-oesophageal echocardiography, drug selection, or the anesthetic agents selected by the clinicians performing the operations. All patients had full, routine invasive and non-invasive monitoring applied, including electrocardiography, central venous and peripheral arterial catheterisation, urine outputs and central temperature. After chest closure the patients were transferred to the CICU, sedated and ventilated, and the study commenced by the study anesthetist (1st author). The patients were in a clinically stable condition when admitted to the unit, and were attended by the treating clinicians until the handover to the study anesthetist occurred. The LiDCOplus ® (lithium dilution cardiac output) monitor (LiDCO Limited, Flowers Building, Granta Park, Cambridge CB1 6GU, United Kingdom) was calibrated after attachment to the patient''s arterial line, and the patient''s beat-to-beat hemodynamic data transferred to the host laptop computer. The CDSS graphical interface displays the patient''s clinical details and specific cardiovascular data and prompts the anesthetist to input the target ranges for each parameter, and select a suitable advisor frequency. This is the frequency with which the therapeutic advice is displayed on screen with an audible prompt for a control inputs from the anesthetist. In each case this was selected to be 30s. When the study anesthetist agreed with the CDSS advice (administration of fluid, commencing a drug, altering the drug infusion rate) the syringe motif on the “Advisor Infusion Rates” panel of the graphical interface was ‘clicked’ on and the infusion rate immediately and manually inputted to Graseby® 3400 pumps. If any disagreement between the anesthetist and the computer''s advice arose, the syringe motif on the “Expert Infusion Rates” panel of the preferred drug was ‘clicked’ on and the expert''s therapeutic decision (e.g. infusion rate) was entered in the corresponding data field and then applied to the pump. During all trials, data was stored for off-line analysis. Results: The CDSS successfully selected suitable drug therapies for each case and advised reasonable and appropriate infusion rates such that the study anesthetist did not have to override the suggested CDSS instructions and infusion rates. Under differing clinical conditions the system was able to maintain clinically appropriate and stable control of the cardiovascular system (CVS), with good profiles under noisy physiological measurements, and was readily able to regain control following transient deterioration of the patient hemodynamic parameters (coughing, or during blood sampling). [Copyright &y& Elsevier]

Summary: Objective: Survival probability predictions in critically ill patients are mainly used to measure the efficacy of intensive care unit (ICU) treatment. The available models are functions induced from data on thousands of patients. Eventually, some of the variables used for these purposes are not part of the clinical routine, and may not be registered in some patients. In this paper, we propose a new method to build scoring functions able to make reliable predictions, though functions whose induction only requires records from a small set of patients described by a few variables. Methods: We present a learning method based on the use of support vector machines (SVM), and a detailed study of its prediction performance, in different contexts, of groups of variables defined according to the source of information: monitoring devices, laboratory findings, and demographic and diagnostic features. Results: We employed a data set collected in general ICUs at 10 units of hospitals in Spain, 6 of which include coronary patients, while the other 4 do not treat coronary diseases. The total number of patients considered in our study was 2501, 19.83% of whom did not survive. Using these data, we report a comparison between the SVM method proposed here with other approaches based on logistic regression (LR), including a second-level recalibration of release III of the acute physiology and chronic health evaluation (APACHE, a scoring system commonly used in ICUs) induced from the available data. The SVM method significantly outperforms them all from a statistical point of view. Comparison with the commercial version of APACHE III shows that the SVM scores are slightly better when working with data sets of more than 500 patients. Conclusions: From a practical point of view, the implications of the research reported here may be helpful to address the construction of cheap and reliable prediction systems in accordance with the peculiarities of ICUs and kinds of patients. [Copyright &y& Elsevier]

Health care effectiveness and efficiency are under constant scrutiny especially when treatment is quite costly as in the Intensive Care (IC). Currently there are various international quality of care programs for the evaluation of IC. At the heart of such quality of care programs lie prognostic models whose prediction of patient mortality can be used as a norm to which actual mortality is compared. The current generation of prognostic models in IC are statistical parametric models based on logistic regression. Given a description of a patient at admission, these models predict the probability of his or her survival. Typically, this patient description relies on an aggregate variable, called a score, that quantifies the severity of illness of the patient. The use of a parametric model and an aggregate score form adequate means to develop models when data is relatively scarce but it introduces the risk of bias.This paper motivates and suggests a method for studying and improving the performance behavior of current state-of-the-art IC prognostic models. Our method is based on machine learning and statistical ideas and relies on exploiting information that underlies a score variable. In particular, this underlying information is used to construct a classification tree whose nodes denote patient sub-populations. For these sub-populations, local models, most notably logistic regression ones, are developed using only the total score variable. We compare the performance of this hybrid model to that of a traditional global logistic regression model. We show that the hybrid model not only provides more insight into the data but also has a better performance. We pay special attention to the precision aspect of model performance and argue why precision is more important than discrimination ability. [Copyright &y& Elsevier]