Systems have come a long way, but there’s still room for innovation
January 7, 2014
Ten years ago, the Center for Information Technology Leadership at Partners HealthCare in Boston published a widely heralded report suggesting that nationwide adoption of ambulatory CPOE, tightly coupled with electronic health records and clinical decision support, could prevent 2 million adverse drug events, 130,000 life-threatening medication errors and 190,000 hospital admissions per year – all while saving $44 billion, largely from avoiding duplication.
“Those were the days when electronic medical records were kind of a gleam in some people’s eyes,” recalled Jan Walker, RN, who served as executive director of the now-defunct CITL at the time the report came out.
Those also were the days when the Leapfrog Group, a Washington-based coalition of large healthcare purchasers, was heavily promoting CPOE adoption as a way to boost patient safety.
Most notably, though, hospitals were treading carefully after Cedars-Sinai Medical Center in Los Angeles turned off its CPOE system in early 2003. Medical staff there rebelled after a house-wide “big bang” rollout, realizing that the poorly planned system increased their workloads and encouraged cutting corners. It took another eight years before Cedars successfully implemented CPOE.
In the past decade, several studies have been published suggesting that CPOE can introduce and even magnify errors without proper safeguards, particularly against physicians suffering from “alert fatigue” turning off notifications.
meaningful use in 2011, even as some remain wary.
A February 2013 paper in the Journal of the American Medical Informatics Association estimated that nearly a third of U.S. acute care hospitals had fully implemented CPOE by 2008.
Leapfrog used to give hospitals credit if they were merely planning on installing CPOE. Now, the technology not only has to be operational, it has to work properly.
In 2008, when Leapfrog began evaluating CPOE systems, 108 hospitals met the organization’s minimum standards. Five years later, 847 reported having functional CPOE and 523 passed the test. That represents about half of all hospitals reporting to the Leapfrog Group, and is up from 292 reporting and 181 passing in 2011.
“It’s a completely new ball game,” said Binder. She said meaningful use, which began in 2011, is the primary driver, even if the Stage 1 CPOE requirement of at least one medication order for 30 percent of patients was so low.
In Stage 2, the minimum increases to 60 percent of patients for medication orders, 30 percent for lab orders and 30 percent for radiology orders.
Leapfrog’s criteria are more stringent. The organization tests to make sure clinical decision support mechanisms are in place. A hospital has to show that more than 75 percent of medication orders are entered electronically in at least one inpatient unit because Leapfrog members want to emphasize prevention of adverse drug events.
“We have no better measure right now in this country for medication errors,” Binder explained. Medication error is the “No. 1 patient safety problem by far,” she continued. “We use CPOE as a proxy measure.”
From the provider side, CPOE can be a big help – or, as some learned years ago, an untenable burden.
“I’ve lived it and I have the scars,” Hughes quipped.
Today, what Hughes describes as the “PlayStation-Xbox generation” of physicians enter practice wondering why the level of automation has been so poor, said Hughes. “The difficulty with CPOE has been making it part of a physician’s natural workflow,” he suggested.
Longtime medical informatics professional Howard Landa, MD, tried to implement CPOE at Loma Linda University Medical Center in California in 2001, two years before the Cedars debacle. He brought it up on one unit, then took it down shortly thereafter because physicians complained of having to do too much work.
[See also: CPOE remains a challenge for many, surveys show.]
“We weren’t accounting for the workflows,” said Landa, now the chief medical information officer for Alameda County Medical Center in Oakland, Calif.
Landa, vice chairman of the Association of Medical Directors of Information Systems, knows today that there is a balance between facilitating workflow and disrupting it.
“We tried to do too much without understanding the integration of decision support into the workflow,” Landa said of earlier CPOE efforts. “All that does is frustrate.”
He believes CPOE has improved markedly in the last few years after a long period of stagnation. Landa used a Technicon Data Systems CPOE system as an intern at New York University in 1983.
“Between then and the early 2000s, you didn’t see much difference,” he said. There was little in the way of clinical decision support or workflow support for order entry, he said.
“I look at CPOE as a piece of decision support,” said Landa. “Rules and alerts require CPOE.” He added, “It is hard to do meaningful use without an aggressive CPOE program.”
Hughes said that there is kind of a spectrum of alert fatigue. System designers have begun to distinguish between subtle guidance and “you need to act now” kind of guidance.
“There is becoming stratification of alerts,” he said. And context matters. For example, test results often are abnormal in patients in intensive care, so it is almost counterproductive for a CPOE to keep flagging abnormal values in an ICU, he noted.
“I would say were in the second, maybe third generation of CPOE systems, we have a lot more refining to do,” Hughes said. “There’s a huge amount of opportunity for innovation still ahead.”