Daniel M. Keller, PhD

November 16, 2012

SAN DIEGO, California — An epidemic of cognitive drift (CD) is affecting intensive care unit (ICU) physicians. In a small study reported here at Kidney Week 2012: American Society of Nephrology 45th Annual Meeting, Macaulay Onuigbo, MD, MBA, a nephrologist and transplant physician in the Department of Nephrology at Mayo Clinic Health System in Eau Claire, Wisconsin, described a 100% affliction rate of this computer-induced malady.

"The big problem was the fact that I noticed that not just me but other colleagues...were getting very frustrated with the time delays between the click of a mouse and the time you get to see what you expect to see on the PC screen," Dr. Onuigbo told Medscape Medical News.

CD is what happens in the time between clicking a computer mouse and when new information appears on the screen, according to Steven D. Levitt and Stephen J. Dubner, authors of Superfreakonomics. The interval can be as short as 1 second. If the lag is 10 seconds, the authors posited that the user's attention can be lost completely.

With the introduction and now deep penetration of electronic health records (EHRs) and computerized physician order entry (CPOE), it was hoped that medication errors would be reduced and charting made more efficient. The Affordable Care Act mandates "meaningful use" of EHR systems.

One study of physicians' overall satisfaction with an EHR system showed that 27% were very dissatisfied, 30% were dissatisfied, 23% were neutral, and 20% were satisfied or very satisfied.

In his poster presentation, Dr. Onuigbo said that the hope of reduced errors has not been realized and that there is evidence of some unintended consequences of increased medication errors under some circumstances. From a literature review, he said he found 73 references to CD related to neurology and psychiatry topics but none related to physicians and EHRs or to medication errors.

For his pilot survey, Dr. Onuigbo performed oral interviews with 10 randomly selected ICU physicians at a community hospital in northwestern Wisconsin to assess the frequency of CD during EHR use. He used a 10-second interval to assess the potential for cognitive drift.

He first asked survey participants to count off 10 seconds. "And then I asked the doctor, 'From your experience on a daily basis...when you work on the [EHR], how often do you think you get to wait for more than 10 seconds before the next screen comes on?' and [the answer] was literally, 'All the time, very common,' " Dr. Onuigbo said. All the respondents said the wait was frustrating, distracting, "and sometimes they simply quit. They left the PC."

He found that all 10 of the surveyed physicians (100%) reported experiencing CD several times a day. In their opinions, CD was a source of significant frustration, stress, possible burnout, and user resistance to EHR implementation.

"As far as the drift component goes, it was clear that for most doctors, more than 10 seconds meant that they would begin to lose focus," Dr. Onuigbo said. He pointed out that physicians are very busy and have many tasks to do in seeing just 1 patient. Any delay can be a "time sink" and a distraction. Pulling up a patient's laboratory values, fluid input and output, and other items can take more than a minute on some systems. "That's totally frustrating, distracting, and before you know it, you lose focus, and then the chances for errors begin to creep up fairly quickly," he said.

He concluded that CD is common, unreported, and unrecognized as a source of physician stress and causes distraction that can lead to medication errors.

Kenneth Schild, MD, an emergency physician at Scripps Clinic in La Jolla, California, agreed with Dr. Onuigbo's findings. "Talking to some of my colleagues, and many of them work in other institutions, all of the [EHR] systems have similar issues," he commented to Medscape Medical News. "Some, of course, are better than others." He said the lag time varies, and it can be especially long at night when the systems are doing automatic backups.

"The other issue, which isn't brought up by this [poster is]...since you're spending so much time working with a system that was not part of your training or part of your practice, you're thinking so much about the software that you think less about the patient, which is a problem. And also it takes you away from patient care as far as time. So we're spending in a 12-hour shift maybe 1 to 2 hours in front of the computer system, which used to be spent mostly in front of the patient," Dr. Schild said.

In his view, a major cause of poor system performance is that new systems are installed onto old computers and old networks, and the networks are often tying together legacy systems.

"These links, depending upon how good your [EHR] system is, are very slow. There's conversion," he said.

For example, laboratory reports from clinical chemistry and images from radiology need to be converted into a system-compatible format. "So everybody has a similar issue that there's these links between servers, so not only are you having a problem with your own network just going to the [EHR] server, but the [EHR] is picking up information from all of these other servers within your clinic or hospital, and it slows things down," Dr. Schild explained.

At big institutions, a handful of people decide which EHR system to purchase based on "cost and presentation and everything else, and it becomes difficult," he said. In addition, different parts of different EHR systems may function better or worse depending on the supplier, and different departments may have different needs. So systems are a result of compromise to some degree.

Although the lag time between screens is a problem, Dr. Schild said so is the number of mouse clicks needed to access information or functions, as well as the design of the interface, which varies from system to system.

"Many of the systems were purchased and grouped together, so even within the system that we use at Scripps, there're probably 3 or 4 databases because there were 3 or 4 programs that were merged together. Each one actually functions somewhat differently. You can tell it was a hodgepodge," Dr. Schild said.

Scripps is not unique, he said. "If you look at any of the surveys, most physicians are unhappy with their EHR system."

Dr. Onuigbo said that in discussions at his poster with physicians from the United States and elsewhere, he found that "it's not an uncommon problem."

He suggested that including practicing physicians in all phases of EHR design, development, and implementation may help to lessen the CD from using the systems. Optimizing connectivity; faster and more robust servers, networks, and workstations; and minimizing the number of mouse clicks to perform a task will all be needed to make the systems work seamlessly and decrease the chances of CD.

The study did not receive any commercial funding. Dr. Onuigbo has disclosed no relevant financial relationships. Dr. Schild was not involved in the study and has disclosed no relevant financial relationships.

Kidney Week 2012: American Society of Nephrology 45th Annual Meeting. Abstract FR-PO070. Presented November 2, 2012.