Data Mining to Predict the Use of Vasopressors in Intensive Medicine Patients

Abstract

The role that decision making process plays in Intensive Medicine is very critical essential due to the bad health condition of the patients that go to Intensive Care Units (ICU) and the need of a quick and accurate decisions. Therefore each decision is crucial, because it can help saving endangered lives. The decision should be always taken in the patient best interest after analyzing all the data available. In the eyes of the intensivists, the ever growing amount of available data concerning the patients, makes it each time more difficult for them to make a decision based on so many information. It is based on this ideal of improving the decision making process, that this work arises and Data Mining models were induced to predict if a patient will need to take a vasopressor, more specifically: Dopamine, Adrenaline or Noradrenaline. This work used real data provided by an Intensive Care Unit and collected in real-time. The data mining model were induced using data from vital sign monitors, laboratory analysis and information about the patient’s Electronic Health Record. This study was based in clinical evidences and provided very useful results with a sensitivity around 90%. These models will reduce the need of vasopressor drugs by helping intensivists to act and take accurate decision before the vasopressor be need by the patient. It will improve the patient condition because when the time comes the predicted necessity of the vasopressor will cease to exist due to the early care provided by the intensivist. The decisions can be for example change the therapeutic plan. Overall, the decision making process becomes more reliable and effective and the quality of care given to patients is better.

Hospital readmission prediction with hybrid‐sampling and self‐paced balance learning

“…20 Braga et al evaluated the risk index of discharged patients using methods such as support vector machines (SVM) and decision trees. 21 These methods can accurately predict readmission risk and aid in identifying patients who may require readmission and return to the intensive care unit (ICU).…”

“…For example, Golmohammadi et al improved readmission prediction accuracy by introducing neural networks, decision trees, and chi‐squared automatic interaction detection methods 20 . Braga et al evaluated the risk index of discharged patients using methods such as support vector machines (SVM) and decision trees 21 . These methods can accurately predict readmission risk and aid in identifying patients who may require readmission and return to the intensive care unit (ICU).…”

Pervasive Patient Timeline for Intensive Care Units

“…This feature can help reducing costs and manage resources. In this case Data Mining models were developed to predict the vasopressors need [20]. The figure is presenting the probability of a patient need vasopressors in the next hour.…”

Clustering of Drug Sampling Data to Determine Drug Distribution Patterns with K-Means Method : Study on Central Kalimantan Province, Indonesia

“…Many studies have been conducted in health sector. Data mining has been able to provide many contributions to the decision-making system for health services [8] [19], to find out the pattern of a disease [11], to find the cause of the spread of a disease [14], to predict or early diagnose of a disease [6] [12], to be used in finding great information on the costs of care and treatment of patients who have certain diseases [9] [16] and can provide alternative recommendations in treatment [7][10] [17]. However, study on knowledge discovery in drug surveillance data has never been done.…”