Date of Award
2008
Degree Type
Thesis
Degree Name
Master of Science
Program
Epidemiology and Biostatistics
Supervisor
Dr. Orlando da Silva
Second Advisor
Dr. Karen Campbell
Third Advisor
Dr. GuangYong Zou
Abstract
A population-based retrospective cohort generated by linking Obstetric and NICU databases at St Joseph’s Health Care was used to generate a tool capable of predicting respiratory disorders in term and late preterm infants. Singletons were included, while multiples, congenital anomalies, and infants that were small (<3 percentile) and large (>97 percentile) for gestational age were excluded from analyses. Descriptive statistics and risk ratios were used to compare morbidity rates, intervention use and NICU admission rates. Multivariable logistic regression was used to construct the predictive model. Morbidity rates, intervention use and NICU admission rates decreased with increasing gestational age. Little difference was apparent in discrimination and calibration across the ten models constructed and thus model 10 which uses 1/3 of the predictors while maintaining better than good discrimination and calibration may be the most clinically applicable. Ultimately, this tool may assist clinicians in differentiating high from low risk infants who may appear healthy or stable at birth.
Recommended Citation
Prendergast, Patrick M., "Predicting Respiratory Disorders in Term and Late Preterm Infants" (2008). Digitized Theses. 4784.
https://ir.lib.uwo.ca/digitizedtheses/4784