L. Samal (1), T. A. Stavroudis (2), R. E. Miller (3), H. P. Lehmann (4), C. U. Lehmann (5)
(1) Division of General Internal Medicine and Primary Care, Brighman and Women's Hospital; (2) Division of Newborn and Critical Care, Children’s Hospital of Los Angeles; (3) Department of Pathology, Johns Hopkins University School of Medicine; (4) Division of Health Sciences Informatics, Johns Hopkins University School of Medicine; (5) Division of Neonatal-Perinatal Medicine, Johns Hopkins University School of Medicine
Medical informatics applications, clinical laboratory information systems, reminder systems, hospital communication systems, time factors
Background: A computerized laboratory result paging system (LRPS) that alerts providers about abnormal results (“push”) may improve upon active laboratory result review (“pull”). However, implementing such a system in the intensive care setting may be hindered by low signal-to-noise ratio, which may lead to alert fatigue.
Objective: To evaluate the impact of an LRPS in a Neonatal Intensive Care Unit.
Methods: Utilizing paper chart review, we tallied provider orders following an abnormal laboratory result before and after implementation of an LRPS. Orders were compared with a predefined set of appropriate orders for such an abnormal result. The likelihood of a provider response in the post-implementation period as compared to the pre-implementation period was analyzed using logistic regression. The provider responses were analyzed using logistic regression to control for potential confounders.
Results: The likelihood of a provider response to an abnormal laboratory result did not change significantly after implementation of an LRPS. (Odds Ratio 0.90, 95% CI 0.63–1.30, p-value 0.58) However, when providers did respond to an alert, the type of response was different. The proportion of repeat laboratory tests increased. (26/378 vs. 7/278, p-value = 0.02)
Conclusion: Although the laboratory result pager altered healthcare provider behavior in the Neonatal Intensive Care Unit, it did not increase the overall likelihood of provider response.
T.M. Lehmann1, T. Aach2, H. Witte3
IMIA Yearbook 2006 1: 57-67
T. M. Lehmann(1), H. P. Meinzer(2), T. Tolxdorff(3)
Methods Inf Med 2004 43 4: 308-314
C. M. Couris (1), M. Rabilloud 2 , C. Colin (1) , R. Ecochard (2)
Methods Inf Med 2002 41 4: 349-356