Modeling zero-inflated and overdispersed count data: application to in-hospital mortality data

Hyperchloremia (high serum chloride level) is frequently observed in critically ill patients in the intensive care unit (ICU). Clinical evidence shows that hyperchloremia is associated with increased in-hospital mortality. Length of hospital stay (LOS) is often used as an indicator of hospital efficiency, a proxy of resource consumption and is especially important in organizing hospital services. Such data often have a highly right-skewed distribution for non-zero values and possible excess zero counts. Our study aims to examine the association of serum chloride levels at different time points with hospital mortality and to model the length of hospital and ICU stays in conjunction with zero-inflated and overdispersed count data. This research will consider the use of several univariate and multivariate models to evaluate the effects of serum chloride as it pertains to patient mortality. This research resulted from application to more than 1700 critically ill patients from a local hospital.

Grantor dc:publisher

University of Tennessee at Chattanooga

Author dc:creator

• Nanni, Lisa

Contributors dc:contributor

• Gao, Cuilan
• Wang, Jin; Gunasekera, Sumith; Nichols, Roger
• College of Arts and Sciences

• Inflation (Finance) -- Mathematical models
• Medical statistics

Licence

• https://rightsstatements.org/page/InC/1.0/?language=en

Language dc:language

English, eng

Repository record dc:identifier

https://scholar.utc.edu/theses/615

OAI identifier oai:identifier

oai:scholar.utc.edu:theses-1765