¹Department of Pediatrics, Stanford University, Stanford, CA, USA

Correspondence: Dr HC Lee, Department of Pediatrics, Division of Neonatology, University of California, San Francisco, 533 Parnassus Ave., U503, San Francisco, CA 94143, USA. E-mail: LeeHC@peds.ucsf.edu

²Dr. HC Lee's affiliation has changed since completion of the study to the University of California, San Francisco.

Received 15 September 2007; Revised 6 January 2008; Accepted 4 February 2008; Published online 26 June 2008.

Abstract

Objective: To review the results of a quality improvement (QI) project to improve admission temperatures of very low birth weight inborn infants.

Study Design: The neonatal intensive care unit at Lucile Packard Children's Hospital underwent a QI project to address hypothermic preterm newborns by staff education and implementing processes such as polyethylene wraps and chemical warming mattresses. We performed retrospective chart review of all inborn infants with birth weight <1500 g during the 18 months prior to (n=134) and 15 months after (n=170) the implementation period. Temperatures were compared between periods. Multivariable logistic regression was used to account for potential confounding variables. We compared mortality rates and grade 3 or 4 intraventricular hemorrhage rates between periods.

Result: The mean temperature rose from 35.4 to 36.2 °C (P<0.0001) after the QI project. The improvement was consistent and persisted over a 15-month period. After risk adjustment, the strongest predictor of hypothermia was being born in the period before implementation of the QI project (odds ratio 8.12, 95% confidence interval 4.63, 14.22). Although cesarean delivery was a strong risk factor for hypothermia prior to the project, it was no longer significant after the project. There was no significant difference in death or intraventricular hemorrhage detected between periods.

Conclusion: There was a significant improvement in admission temperatures after a QI project, which persisted beyond the initial implementation period. Although there was no difference in mortality or intraventricular hemorrhage rates, we did not have sufficient power to detect small differences in these outcomes.