Estimating Center Effects Using Multiple Methods in a Pediatric Glycemic Control Study Open Access

Shepler, Samantha Hastings (2015)

Permanent URL:


Blood sugar levels can increase to unsafe levels following trauma through injury or illness resulting in prevalent hyperglycemia in pediatric ICUs. This condition has been associated with many negative health outcomes including longer lengths of stay and increased mortality. Glycemic control protocols have been presented as a treatment for critical illness hyperglycemia. This multi-center study created a group of six pediatric ICUs to implement this type of protocol. The protocol consisted of blood glucose checks every 12 hours, if the initial blood glucose reading was >140 mg/dl an additional reading was done within two hours. If the second reading was also >140 mg/dl then insulin was delivered. This analysis aims to assess the relationship between the ICUs and length of stay and mortality for the hyperglycemic pediatric ICU patients. Log-normal, gamma, and Cox proportional hazards models were proposed to model length of stay with random intercepts to allow for variation by ICU. The gamma regression model was selected as the principle method for modeling length of stay due to the distribution of the data and easy interpretation. Emory CICU, the single pediatric cardiac ICU in the study, had significantly shorter estimated length of stay adjusting for pediatric logistic organ dysfunction (PELOD) score, a measure of baseline severity. Weight and if the patient was an infant were also found to be significantly associated with length of stay. Additionally, a logistic model was created to model mortality with random intercepts to allow for variation between ICU. The results showed no significant differences in odds of mortality across the ICUs but showed a significant association between mortality and PELOD score. Our results suggest there is a difference at the ICU level that our data set is unable to ascertain which is driving significantly different lengths of stay for pediatric ICU patients.

Table of Contents

Chapter I

Introduction. 1

Chapter II

Review of Literature. 4

Chapter III

Methodology. 11

Chapter IV

Results. 16

Chapter V

Discussion. 20

References. 25

Tables. 27

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
Subfield / Discipline
  • English
Research Field
Committee Chair / Thesis Advisor
Committee Members
Last modified

Primary PDF

Supplemental Files