MCRB: A Multiclassifier Tool for Risk of Bias Assessment in a Systematic Review to Produce Health Evidence to Decision Making

Healthcare is receiving many improvements from real world evidence. The advances in data management and machine learning models are driving better resources to support decision making. One of the most common techniques to develop recommendations and guidelines to physicians is through systematic rev...

Full description

Saved in:
Bibliographic Details
Published in2020 IEEE 33rd International Symposium on Computer-Based Medical Systems (CBMS) pp. 1 - 6
Main Authors Goncalves Pereira, Ramon, Zanon Castro, Giulia, Azevedo, Pamela, Torres, Lucas, Zuppo, Isabella, Rocha, Tulio, Afonso Guerra, Augusto
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.07.2020
Subjects
Clinical trials
Data models
Guidelines
Machine learning
multiclass classification
natural language processing
systematic review
Systematics
Training
Online AccessGet full text

Cover

More Information
Summary:Healthcare is receiving many improvements from real world evidence. The advances in data management and machine learning models are driving better resources to support decision making. One of the most common techniques to develop recommendations and guidelines to physicians is through systematic reviews. To evaluate the quality of the evidence, among other methods, the researchers uses the Risk of Bias Assessment. Besides, that method is widely used and provides good results it has been done manually by researchers. This work provides the MCRB a multiclassifier tool for the risk of bias assessment using machine learning techniques. The software was tested with four machine learning models: Logistic Regression, SVM, Naive Bayes, and XGBoost. The results show that is quite possible to classify well the seven dimensions of the problem with a macro average AUC Score of 75%.
ISSN:2372-9198
DOI:10.1109/CBMS49503.2020.00008