Journal
STOCHASTIC ENVIRONMENTAL RESEARCH AND RISK ASSESSMENT
Volume 31, Issue 2, Pages 417-430Publisher
SPRINGER
DOI: 10.1007/s00477-016-1227-5
Keywords
Clostridium difficile infection (CDI); Predictive model; Decision tree; Fuzzy synthetic evaluation
Categories
Funding
- Interior Health Authority (IHA), Kelowna, BC
- Natural Sciences and Engineering Research Council of Canada (NSERC)
Ask authors/readers for more resources
Several risk factors associated with the increased likelihood of healthcare-associated Clostridium difficile infection (CDI) have been identified in the literature. These risk factors are mainly related to age, previous CDI, antimicrobial exposure, and prior hospitalization. No model is available in the published literature that can be used to predict the CDI incidence using healthcare administration data. However, the administrative data can be imprecise and may challenge the building of classical statistical models. Fuzzy set theory can deal with the imprecision inherent in such data. This research aimed to develop a model based on deterministic and fuzzy mathematical techniques for the prediction of hospital-associated CDI by using the explanatory variables controllable by hospitals and health authority administration. Retrospective data on CDI incidence and other administrative data obtained from 22 hospitals within a regional health authority in British Columbia were used to develop a decision tree (deterministic technique based) and a fuzzy synthetic evaluation model (fuzzy technique based). The decision tree model had a higher prediction accuracy than that of the fuzzy based model. However, among the common results predicted by two models, 72 % were correct. Therefore, this relationship was used to combine their results to increase the precision and the strength of evidence of the prediction. These models were further used to develop an Excel-based tool called C. difficile Infection Incidence Prediction in Hospitals (CDIIPH). The tool can be utilized by health authorities and hospitals to predict the magnitude of CDI incidence in the following quarter.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available