Clinical Laboratory Strategies: April 29, 2010

Strategies logo

Abnormal Result Notification in Electronic Medical Records
Study Finds Disconnect Between Lab Alerts and Follow-Up
By Bill Malone


In an effort to improve quality and safety, electronic medical record systems frequently include special alerts and messages to providers that were not possible with paper methods. This issue of Strategies examines new research that highlights the ways in which the lab’s efforts to communicate electronically with ordering providers can have unexpected results.

The advent of electronic medical records (EMR) poses dramatic changes to communication between labs and ordering providers, an area that both parties struggle with even when limited to paper and telephone. As specified by CLIA, it is the lab’s responsibility to get test results to the right person at the right time. Yet EMRs have opened up myriad possibilities beyond these two longstanding communication modes to enhance and organize data for providers.

Aiming to find whether one type of enhanced electronic message—automated abnormal result notifications—made a difference in follow-up of patients, a new study examined provider behavior in response to four lab alerts: hemoglobin A1c ≥15%, positive hepatitis C antibody, prostate-specific antigen ≥15 ng/mL, and thyroid-stimulating hormone ≥15 mIU/L (Am J Med 2010;123:238-44). The study was conducted at the Michael E. DeBakey Veterans Affairs (VA) Medical Center and five satellite clinics in Southeast Texas, all of which use the same VA EMR. The researchers examined whether or not providers acknowledged alerts via the EMR interface within 2 weeks of receipt, and further, whether providers took appropriate follow-up actions within 30 days.

The results present a perplexing picture of communication issues. About 10% of the alerts went unacknowledged, with timely follow-up lacking in 6.8% of cases, a figure that was not statistically different for acknowledged versus unacknowledged alerts. The researchers also discovered that 17.4% of the alerts arose from redundant tests, but oddly, alerts for new diagnoses were  more likely to lack timely follow-up compared with alerts for tests that were redundant.

Even though the findings seem to undermine the value of such alerts, the study did not include a control scenario in which there were no alerts, and so a conclusion that alerts are not completely effective should not and cannot be drawn, cautioned Hardeep Singh, MD, MPH, the lead author of the study and assistant professor of medicine at Baylor College of Medicine in Houston, Texas. “What we can conclusively say is that when healthcare providers receive these alerts and acknowledge them, there is much more to patient safety and timely follow-up than just a provider looking at an alert on the computer screen,” Singh said. “It’s the follow-up actions that they take after acknowledgment that are really important.” Singh is also an investigator at the Houston VA Health Services Research and Development Center of Excellence and director of the VA Center of Inquiry to Improve Outpatient Safety Through Effective Electronic Communication.

Because of the way the VA EMR works, understanding the “added” value of alerting is not straightforward. Though providers see all of their alerts—whether from the lab or otherwise—when they log onto the EMR, they can also see an abnormal test result by directly viewing a patient record, even if they did not click on an alert to get there. In the study, most of the alerts were acknowledged, and it’s impossible from the study’s data to know if the alerts were the factor that nudged providers to follow-up when they did so.

Most bewildering to the investigators was the finding that timely follow-up was higher in redundant or unnecessary tests instead of alerts indicating a new diagnosis. “It is concerning,” said Singh. “It appears that alerts about new findings were more at risk of getting lost to follow-up versus some of the older information that the providers already knew about. We could not completely explain it.”

Why did providers seem to overlook following up important information when faced with lab alerts?  The authors speculated that the high number of redundant tests may have distracted providers from the more important information, undermining the perceived value of all laboratory alerts in general. While the study did not test this hypothesis directly, other laboratorians who work with EMRs agreed that a lot of work remains to be done before the many capabilities of electronic systems consistently improve care and communication rather than distract from it.

“For some of us, this study proves what we already felt and recapitulates a lot of the problems many of us face launching an EMR,” said Albert Jekelis, PhD, a lab consultant, retired director of laboratory services at UBHC of the University of Medicine and Dentistry of  New Jersey, and past chair of the AACC Lab Information Systems and Medical Informatics Division. “But I think it’s still very important because it can lead to more improvement. Communication between the end user and the laboratory, which has never been great, may now need to get even better.”

Because responsibility lies both with providers accessing lab information in an EMR as well as the lab that provides it, laboratorians and providers have to work even harder to understand one another in order to make electronic systems work, Jekelis emphasized. “One of the biggest complaints I get from the physician community is that they’re just overwhelmed with information,” he said. “The lab has to be cognizant that the providers’ job is not just to review data: they have to see the patient.”

Joanna Baker, a laboratory information officer at Moncrief Army Community Hospital in Fort Jackson, South Carolina agreed. Beyond just the software, those responsible for an EMR must also carefully consider ‘people-ware,’ how users learn and interact with an EMR. “This could have been just a teaching issue, that the people that did not follow up had not been using the VA EMR long enough or didn’t understand how to use it efficiently,” she observed. “Going into the 21st century, all of us are going to have to figure out why all of these improvements that we are making on the technical end of healthcare are not working the way they should on the other end for the providers.” Neither Baker nor Jekelis were involved in the study.

One area where EMRs clearly need attention is in education, noted Baker. “Some of the new people that come on board are easy to teach, but we’re also hiring providers that are older and tend to be more set in their ways,” she said. “These providers are harder to convince that you really need to follow through stepwise on the EMR. Some healthcare providers think the system is slow because they don’t have good functional knowledge of it, as simple as which button to push and when, or which window to use.”

Going forward, how providers respond to lab alerts and other features of the EMR will need to be tackled with a different strategy, Baker suggested, such as from a patient safety viewpoint that would involve standards for hospital accreditation. Similarly, Jekelis would like to see future research examine the concept of personalized alerts, where different providers in different specialties with different needs can receive more focused alerts from the lab. “I would like to see how we can avoid over stimulating the end user, going back to the question from this study of whether overstimulation promotes cynicism and a disregard of the alerts.”

 

Rate this page:       
Page Access: