Lessons from early adopters: How healthcare IT and people interact
ORLANDO–The deployment of healthcare information technology (HIT) to reduce medical errors, improve patient safety, and achieve economic efficiencies is not a wholesale panacea for these challenges. “E-iatrogenesis,” a term that has enjoyed currency of late, is widely used to describe the unintended consequences generated by HIT systems and how they impact workflow and affect quality.

“The problems with HIT have resulted in several journal articles examining e-iatrogenesis,” said Blackford Middleton, MD, in a presentation today at the 2008 HIMSS conference.

Middleton, chairman of the Center for Information Technology Leadership at Brigham and Women’s Hospital, Harvard Medical School in Boston, shared some of the latest research on e-iatrogenesis during a joint nursing informatics, pharmacy IT, and physician’s IT symposia this morning.

According to Middleton, “approximately 60 to 70 percent of clinicians have reported some problems with HIT errors.”

For example, in a study published in the Journal of the American Medical Association, (JAMA, Vol. 293:10, March 2005), 22 categories of increased medication risk were identified due to the utilization of computerized physician order-entry (CPOE) systems. Approximately one-third of the possible problems were assessed as information errors, while the remaining two-thirds were identified as input or workflow errors. He noted that the study survey was based on user’s perception of risk and not actual events.

Perception of CPOE errors was not the basis of a study presented in the journal Pediatrics (Vol. 116:6, Dec. 2005); rather, mortality figures before and after implementation of the system in a critical-care department formed the endpoint. Researchers found a pre-CPOE mortality rate of 2.8 percent compared with a post-deployment mortality rate of 6.57 percent, Middleton reported.

He said that the study authors found there were order delays due to lack of pre-registration in the system; nurses were away from the beside and were performing data entry at a CPOE workstation; it delayed pharmacy administration and there were problems with order timing for subsequent medication doses; and that there was an up-front time cost to enter the orders.

Lastly in a multi-site analysis of five CPOE mature programs (JAMA, Vol. 13:5, Sept./Oct. 2006), researchers uncovered 79 unintentional adverse consequences noted in CPOE use by 95 providers. They refined this data into nine categories of unintended consequences of HIT adoption that Middleton shared with his audience.
1. Work for clinicians, 19.8 percent frequency
2. Unfavorable workflow issues, 17.6 percent frequency
3. Never-ending systems demands, 14.8 percent frequency
4. Problems related to paper persistence, 10.8 percent frequency
5. Untoward changes in communication patterns and practices, 10.1 percent frequency
6. Negative emotions, 7.7 percent frequency
7. Generation of new kinds of errors, 7.1 percent frequency
8. Unexpected changes in the power structure, 6.8 percent frequency
9. Overdependence on the technology, 5.2 percent frequency
Middleton does not advocate against adopting HIT products such as CPOE systems; instead his focus on presenting an overview of e-iatrogenesis research findings is to share the lessons learned by early adopters of CPOE.

“Clinical decision support does not necessarily equal truth,” he said. “All systems are vulnerable: if you put garbage in, you’ll get garbage out.”