Widgets Magazine

Computerized system may reduce hospital mortality

A recent study by Lucile Packard Children's Hospital and the Stanford School of Medicine has shown that mortality rates decline following the implementation of a computerized physician order entry (CPOE) system. The study provided evidence for correlation between the two variables, though causation was not determined. (Stanford Daily File Photo)

A new study conducted by Lucile Packard Children’s Hospital and the Stanford School of Medicine has demonstrated that hospital-wide mortality rates fall substantially with the introduction of a computerized physician order entry system (CPOE). Published in this month’s issue of “Pediatrics,” the study is the first to establish this correlation.

CPOE allows physicians to order tests and prescribe treatments and medications electronically, making their patients’ information readily accessible to authorized hospital staff as part of a broader electronic medical records (EMR) system.

CPOE was implemented in 2007 at Lucile Packard Children’s Hospital. Researchers compared mortality from 2001 to 2007 and then to 2009 in a perspective control study that analyzed historical data and isolated variables using an “autoregressive integrated moving average model,” said Jin Hahn, the co-author of the study. Hahn, a professor of pediatric neurology, served as co-director of the LPCH clinical transformation program from 2004 to 2008, when CPOE implementation took place.

Lead author Christopher Longhurst, an assistant clinical professor of pediatrics, and his colleagues concluded there was a 20-percent reduction in mortality rates, after controlling for the rapid response team, seasonal variability and other factors. Data analysis indicated two statically significant decreases in mortality: one decline after implementation of rapid response teams in 2005 and a second following the introduction of CPOE in 2007.

Prior to the implementation of CPOE, ordering medication electronically was somewhat problematic. Now, the CPOE systems integrated at Lucile Packard include windows with real-time suggestions and notifications of possible calculation errors on-screen.

“We were able to cut the time between critical medication order and time of administration<\p>–<\p>something we speculate was a huge cause in mortality,” Longhurst said.

Despite the promising results of the study, researchers hesitate to define a causal relationship.

“I think it’s important to emphasize correlation,” Longhurst said. “With this sort of data, we can’t guarantee any factor led to the results. On the other hand, we saw a meaningful decrease even after controlling for variables.”

Previous studies have shown neutral or negative effects in pediatric hospitals following implementation of CPOE, Hahn said. Longhurst’s study is the first to show a statistically significant evidence of a decrease in mortality.

“A 2005 Pittsburgh study on the effect [of a CPOE system] at a children’s hospital showed a slight bump up in mortality,” Longhurst said. “This was obviously very troubling. A follow-up study in Seattle showed no change, showing systems could at least be put in safely without hurting children.”

Longhurst found the trend toward more widespread implementation of CPOE systems, in conjunction with electronic medical records, to be promising. He cited the 2009 Health Information Technology for Economic and Clinical Health (HITECH) Act signed into law by President Obama as a manifestation of this trend. The HITECH Act allotted $19.2 billion to increase the use of electronic medical records by physicians and hospitals.

“Currently, I would estimate only 10 to 15 percent of hospitals are using computerized physician orders while 50 percent of children’s hospitals are,” Longhurst said. “These numbers will rise.”

Still, transitioning to the new computerized system may be difficult for older physicians who are accustomed to paper environments.

“For most physicians, CPOE has been the single biggest change,” Longhurst said. “Good adoption of the system is very challenging. We never said it would be easier or faster, but it would be safer for patients.”

Hahn agreed that the steep learning curve may be problematic at first and believes that the current software interface has much room for improvement.

“Many other hospitals are implementing CPOE and EMR and it is only a question of time when these will be the standards of care in hospitals in the U.S.,” Hahn said.

Despite its growing recognition, the study has not convinced all researchers. A three-year study conducted at Seattle’s Children Hospital found no significant change in mortality after CPOE was implemented in 2003.

“The study was not a controlled experiment that you might do in a lab,” wrote Mark Del Beccaro, chief medical information officer at Seattle’s Children Hospital, in an e-mail to The Daily. “In the laboratory setting, the investigator sets up the experiment and usually manipulates one variable at a time and measures the effect. In a complex hospital setting this isn’t possible with EMRs.”

“A group of children’s hospitals is considering looking at pooled data from those that have and have not implemented an EMR with CPOE and seeing if the Packard data is more generalizable to other institutions,” Del Beccaro added.

Looking toward the future, Longhurst said he plans to carry out a follow-up study that investigates which specific aspects of implementation lead to the lower mortality rates.