BOX 2-1
Unintended Consequences of Health IT: A Look at Implementing CPOE

Two pediatric intensive care units (ICUs) implemented the same electronic health record (EHR) system with computerized provider order entry (CPOE) in Pittsburgh and Seattle. The Pittsburgh experience led to a significant increase in mortality, while the same system implemented in Seattle did not (Del Beccaro et al., 2006; Han et al., 2005). Later, several other children’s hospitals introduced the same CPOE system, leading to no change or even lower rates in mortality (Longhurst et al., 2010).

The differing impact on mortality rates may be due to the hospitals’ differences in the implementation and use of the CPOE system. These differences, as illustrated by the Pittsburgh and Seattle pediatric ICUs, are highlighted below:


•   Specific order sets designed for critical care were not created.

•   Changes in workflow were not sufficiently predicted, resulting in a breakdown of communication between nurses and physicians.

•   Orders for patients arriving via critical care transportation could not be written before the patients arrived at the hospital, delaying life- saving treatments.

•   Changes, unrelated to the CPOE system, were made in the administration and dispensing of medication that further frustrated the clinical staff, for example:

image   At the same time the CPOE system was installed, the satellite pharmacy serving the neonatal ICU was closed and medications had to be obtained from the central pharmacy, delaying treatment.

image   Emergency prescriptions were required to be preapproved, and all drugs were moved to the central pharmacy.


•   Researchers visited Pittsburgh to learn about problems associated with their implementation of the CPOE system.

•   Intensive care staff was actively involved during the design, build, and implementation stages.

•   Specific order sets were designed for ICU and pediatric ICU before implementation.

•   New order sets, based on the most frequently used orders, were created to help reduce the time it takes a clinician to enter orders (Del Beccaro et al., 2006; Han et al., 2005).

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement