ucdlabsanalytics.wordpress.com

Team members: Saul Lugoye, Daniele Iannarelli, Brian Henry, Wen-Yu Gu (Forrest)

St. Vincent’s process more than 20000 lab requests on a daily basis, which consists of an average of 5 – 10 different types of test in their biochemistry department, and expressed the need of extracting knowledge or intelligence from the historical laboratory information with the goal of preventing readmission, reduction in patients waiting times, reducing healthcare running costs as well as improving patient care.

At present, laboratory tests that are commonly conducted for patients with kidney disease include measuring the ratios of Electrolytes such as Sodium, Potassium, Chlorine and Bicarbonate in the body. However, for some reasons, the bicarbonate anion is rarely measure, and experts in the field suggest that it should be measured if either the Sodium or Potassium levels fall higher or lower than clinically predefined threshold.

We need to use our technical and data mining skills to analyse the existing electronic historical data sets, and formulating a model or rule that will determine when the bicarbonate anion could be measured if the threshold is either slightly or significantly abnormal. The rule will then be included onto they lab analytical system for automation of such required tests.

Our intention is to design and develop a data-mining tool tailored to the healthcare environment, so that they may be able to analyse and gain knowledge from any other historical clinical information. This system could be used as extra quality control tool for hospital in order to determine the efficiency and effectiveness of new protocols.

To compile & run, open two terminal windows: mvn -P server package java -jar target/veintobrain-0.0.1-SNAPSHOT.jar server veintobrain.yaml

mvn -P client compile mvn -P client gwt:run-codeserver

then visit http://localhost:8080