|
Title:
|
A STUDY TO CLASSIFY HOSPITAL ACQUIRED COMPLICATIONS BASED ON ELECTRONIC HEALTH RECORD: THE PREPROCESSING PHASE |
|
Author(s):
|
Denise Bandeira da Silva, Taís Rabello, Cristiano André da Costa, Diogo Schmidt and Rodrigo da Rosa Righi |
|
ISBN:
|
978-989-8533-69-2 |
|
Editors:
|
Pedro Isaías and Hans Weghorn |
|
Year:
|
2017 |
|
Edition:
|
Single |
|
Keywords:
|
EHR, HAC, MIMIC, Health Informatics, Machine Learning |
|
Type:
|
Full Paper |
|
First Page:
|
39 |
|
Last Page:
|
46 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
The conception of measures to prevent illnesses acquired by patients when they are hospitalized, called Hospital Acquired Complications (HACs), are a challenge these days. Prevention can be achieved through the creation of standard policies and procedures to be followed by hospital agents, as well as by algorithmic analyzes of available hospitalization data, to anticipate cases of HACs. This work extends the preprocessing phase of an application of the machine-learning model for predicting the occurrence of HACs using temporal clinical data. The original work achieved good results in the analysis phase, but they argue that it would be necessary to increase the volume of data with which the results were made so that those results could be generalized. They used the MIMIC II database (Multiparameter Intelligent Monitoring in Intensive Care) and this work aims to convert the application developed to make use of the MIMIC III (Medical Information Mart for Intensive Care) database, which has a much larger data volume. However, this article focuses only on the preprocessing phase. The conversion was successful and the application now retrieves data from the MIMC III according to the needs of the prediction model. The next step of this work is to use the retrieved data and re-test the original model and verify the possibility of generalization of the results. |
|
|
|
|
|
|
