Authors:
Sharareh R. Niakan Kalhori
and
Xiao-Jun Zeng
Affiliation:
University of Manchester, United Kingdom
Keyword(s):
Predicting, Tuberculosis, DOTS, Demographic Data, Logistic Regression.
Related
Ontology
Subjects/Areas/Topics:
Biomedical Engineering
;
Health Information Systems
;
Healthcare Management Systems
;
Support for Clinical Decision-Making
Abstract:
About fifteen years after the start of WHO’s DOTS strategy, tuberculosis remains a major global health threat. Patients vary considerably in their performance in completing treatment course of tuberculosis. Defect in treatment completion have serious undesirable consequences. Although several studies have predicted outcome of treatment for pulmonary tuberculosis, few tools are available to identify high risk patients in finishing treatment course and getting cure prospectively. A logistic regression model proposed to predict the given outcome applying patient demographic characteristics related to just less than 10,000 tuberculosis patients diagnosed by Iranian health surveillance system in 2005. Several tests validate the developed model, X2 (6) = 351.902, P < 0.0001. Also, the model confirmed the significant role of considered factors, calculating the odds ratio of outcome occurring based on each category of variables and explaining the possibility of using the model in other simil
ar patient population. In brief, to support the decision of how intensive the carrying out of DOTS should be for each patient, the predictive models like logistic regression could be useful.
(More)