Unexpected increased mortality after implementation of a commercially sold computerized physician order entry system
These are the reviews for Han's 2005 article, Unexpected Increased Mortality After Implementation of a Commercially Sold Computerized Physician Order Entry System. 
The authors of this article describe their experience with PowerOrders (Cerner) at Children’s Hospital of Pittsburgh. The system was "implemented hospital-wide over 6 days."  The mortality rates were compared 13 months before and 5 months after computerized physician order entry (CPOE) activation .
The methods section notes that they studied all children admitted to their hospital “via interfacility transport for specialized, tertiary-level care during an 18 month period”. Severity of illness was assessed by the PRISM (Pediatrics Risk of Mortality) score. This population is equivalent to pediatric intensive care unit (PICU) /critical care patients.
The methods section further discusses the use of the CPOE system. The staff had a 3-hour teaching/practice session roughly 3 months prior to the system activation. The system provides decision support related to drugs. The orders must be activated by nursing prior to processing by the pharmacy.
The results section outlines the patient characteristics in table 1. There was no significant difference in the pre and post CPOE patients except that there were more patients with central nervous system disease prior to the CPOE implementation. The mortality rate was 2.8% (39 of 1394) prior to CPOE and 6.57% (36 of 548) after CPOE. The P value was less than 0.001. There were multiple factors that were independently associated with risk of mortality including CPOE.
The authors contrast their experience with other staff at their hospital (1) who noted a decrease in adverse drug events. This type of improvement has been noted by other groups (2). The authors also note that entering orders by CPOE takes more time and suggest that this may have adverse consequences in their group of acutely ill patients. Bergeron (3) has raised similar concerns about time. They also note changes in the dynamics of interaction among the staff, i.e. less face-to-face physician-nurse communication.
This article raises a number of important issues regarding CPOE.
- Direct measurement of actual adverse events (death) may have greater validity than measuring potential adverse drug events.
- The amount of training that the staff received (3 hours) was probably inadequate for most of the staff.
- The amount of time after CPOE implementation may have been too short. It would be worthwhile to re-examine this issue 2-3 years post CPOE implementation.
- Improvement in the user-system interface may decrease some of the time problems. The intensive care unit is especially problematic in the time area.
- Ongoing training is important in ensuring optimum use of CPOE. This could be a situation where a biomedical informatician could be helpful in seeing the problems and suggesting possible solutions. The use of custom order sets would be one possible suggestion for the time problem.
Based upon the report by Institute of Medicine and safety initiatives promoted by Leapfrog group, the Han et al wrote that, Department Critical care Medicine, University of Pittsburg School Of Medicine, Pittsburg, PA, implemented a commercially sold computerized physician order entry (CPOE) systems. Accordingly, the attempt was an effort to reduce medical errors and mortality rate. The idea behind the implementation was to test the hypothesis that CPOE results in the reduction of mortality rate among children who are transported for specialized care. 
The method employed was, gathering demographic, clinical, and mortality of all children that were admitted via interfacility transport to the authors’ children’s hospital for a period of 18 months. According to the authors, a commercially purchased CPOE was used during the time frame in general and medical-surgical operations. The implementation of the software was completed in six days. Retrospective analysis of pre-CPOE (15 months) and post-CPOE implementation was done on 1942 children.
The authors claimed that 75 children, among the 1942 who were referred for specialized care during the study period, which result in 3.86%. Analysis revealed that the mortality rate increased significantly from 2.80 % (39 of 1394) before the CPOE implementation to 6.57% (36 of 548) after CPOE. With respect to the observations of the authors, they conclude that the unexpected increase in mortality and the implementation of the CPOE seem to be coincidental. Accordingly, while the CPOE holds great promise, the authors advise that, when implementing CPOE, systems, institutions should continue to evaluate mortality effects, in addition to medical error rates for “children who are dependent on time-sensitive therapies.”
The authors did not totally disagree with the findings of those who were promoting the efficacy of CPOE in clinical settings. It appears that their observations could be prone to errors as well. Moreover, the environment in which the examination was made could not be representative of the total environment in which CPOE systems have been implemented. The unexpected increase in the mortality rate of the patients could also be attributed to the time of the sickness or the degree of illness of the patients involved. Moreover, the authors did not indicate any demonstrable bias against or for the use of CPOE by the clinicians. And this could also be a factor in not using the systems efficiently, which could have had impact on the clinicians’ job performances. Are all these clinicians computer literate? That question was not answered. Therefore, further studies are needed to justify the efficacy of CPOE.
- Computerized provider order entry implementation: no association with increased mortality rates in an intensive care unit
- Perceived increase in mortality after process and policy changes implemented with computerized physician order entry
- Prescription Errors with or without CPOE
- Upperman JS, Staley P, et al. The introduction of computerized physician order entry and change management in a tertiary pediatric hospital. 
- Potts AL, Barr FE, et al. Computerized Physician Order Entry and Medication Errors in a Pediatric Critical Care Unit. 
- Bergeron BP. Medical Errors: Computers Are No Panacea. J Med Pract Manage. 2005 Jul-Aug;21(1):31-4. 
- Han YY, Carcillo JA, et al. Unexpected Increased Mortality After Implementation of a Commercially Sold Computerized Physician Order Entry System. Pediatrics. 2005 Dec; 116(6):1506-12.
- Upperman JS, Staley P, et al. The introduction of computerized physician order entry and change management in a tertiary pediatric hospital. Pediatrics. 2005 Nov;116(5):e634-42. http://www.ncbi.nlm.nih.gov/pubmed/16263977
- Potts AL, Barr FE, et al. Computerized Physician Order Entry and Medication Errors in a Pediatric Critical Care Unit. Pediatrics. 2004 Jan;113(1):59-63. http://www.ncbi.nlm.nih.gov/pubmed/14702449
- Bergeron BP. Medical Errors: Computers Are No Panacea. J Med Pract Manage. 2005 Jul-Aug;21(1):31-4. http://www.ncbi.nlm.nih.gov/pubmed/16206802