CPOE: Promise and Progress

Ever since the Institute of Medicine (IOM) published "To Err Is Human: Building a Safer Health System" in November 1999, pressure has been mounting for hospitals and clinics to implement computerized physician order entry (CPOE) systems to improve patient safety. The highly publicized IOM report concluded that medical errors kill as many as 98,000 people and cause complications in one million hospital patients every year.

Ultimately, CPOE will be used for all diagnostic and treatment orders from physicians. However, most of the immediate interest in CPOE is focused on its ability to reduce medication errors. 

While many agree that CPOE systems can eliminate such errors, the technology hasn't been widely adopted because it's complicated, expensive, and can take years to implement. So far, just over 3 percent of hospitals use CPOE systems, according to a recent study by The Leapfrog Group, the coalition of Fortune 500 companies and other large healthcare purchasers that has jump-started the patient safety movement.

Nevertheless, CPOE "is coming to a hospital near you," says Thomas Yackel, MD, associate medical information officer at the Oregon Health & Science University (OHSU) in Portland. "Hospitals presumably are paying lots to put these systems in, and there will be enforcement of the use of these systems."

Indeed, The Leapfrog Group found that 30 percent of hospitals plan to implement CPOE by 2004, and some hospitals are making its use mandatory. For example, when Cedars Sinai Health System in Los Angeles launched its CPOE system in mid-2002, it required doctors to demonstrate their ability to use the system or risk losing hospital privileges. And a new law gives California hospitals and clinics until Jan. 1, 2005 to eliminate or substantially reduce medication errors, implying widening use of CPOE.

Proven benefits

A study in the September/October 2002 issue of the Journal of the American Medical Informatics Association (JAMIA) concluded that CPOE can be an effective tool for improving the delivery of healthcare. The results, from Ohio State University (OSU) Hospitals and The Arthur G. James Cancer Hospital (The James) in Columbus, Ohio, determined that their CPOE and electronic medication administration record (eMAR) systems, "enhanced patient care by improving turnaround time, reducing transcription errors, and improving verbal order countersignature by physicians," along with other positive findings.

According to Hagop Mekhjian, MD, lead author of the study, "The objective was go to an electronic medical record, not for its own sake, but to meet the needs of quality care in a multi-hospital system with physician clinics and satellite centers. Integrating all parts of the medical record was essential in providing complete information for all care," he says. "One of the crown jewels was physician order entry."

The scope of CPOE

But CPOE isn't simply about replacing paper orders with electronic ones. Unlike order communication tools, which simply capture and transmit orders, CPOE uses rules-based logic to provide relevant information at the time of the order, which helps physicians make appropriate ordering decisions. That's why CPOE requires organizational restructuring that addresses existing information technology (IT), physician workflow, and integrating the processes and technologies of other entities, such as pharmacy and laboratory.
The problem, according to Don Rucker, MD, vice president and chief medical officer at Siemens Health Services, is that "typically, IT departments won't get as much doctor input as they need. IT departments need to understand that it's process engineering, not just software installation."

In most cases, physicians will be required to adopt new clinical workflow procedures and use new computer workstation interfaces, voice recognition units, or data entry units similar to personal digital assistants (PDAs). The first step toward success, Rucker says, is that physicians must agree to work online. Then they must be involved in the design, implementation, and support of the system to ensure that the CPOE application fits the clinical workflow.

To guarantee that physicians took an active role in its system design, OSU Hospitals established teams of physician consultants who signed a contract that outlined their responsibilities. In exchange, medical and administrative leadership empowered the doctors to approve system design and operational policy. The method was such a success, that other hospitals are copying it.

"We are looking to 'hire' physicians in our practice for this purpose to have dedicated time to work on the issues involved," says OHSU's Yackel. "This will likely take from two to four hours per week during the design and implementation phases of the project."

Positioned for success

In response to pressure from patient safety advocates and payers, large urban academic hospitals generally have been among the first to install CPOE systems. Most community hospitals are in the planning stages while they try to secure funding. Typically, once hospitals have their systems in place, they extend CPOE to satellite facilities and physician offices.

Hospitals that are best positioned to adopt CPOE quickly are ones with clinical information systems implemented from a vendor that offers CPOE, according to Jane Metzger, research director for First Consulting Group in Boston. Organizations that need to replace their information systems should evaluate CPOE as part of their decision, but they need to do it within the framework of their overall clinical goals.

"The right answer for one hospital may not be the answer for the hospital next door," says Metzger. "People get fixated on picking a CPOE vendor. They need to remember that CPOE has to fit within the clinical system strategy and a much larger effort."

Workflow changes must be planned carefully to address all caregivers and ancillary departments -- physicians, nurses, laboratories and pharmacy -- and to allow multiple operation changes during CPOE rollout. According to "A Primer on Physician Order Entry," written by First Consulting Group, some of these changes include:

• Each care environment and ancillary department switching from paper to CPOE;
• Patients and providers moving between CPOE-supported units to non-CPOE units; and
• CPOE implementation extending from medicine to surgery or other settings.

During the transition, hospitals may need to maintain many of their existing communication protocols to minimize interruptions to physician workflow, says Donald Levick, MD, MBA, the physician liaison for information services at Lehigh Valley Hospital (LVH) in Allentown, Pa. The community hospital has been implementing CPOE in stages to various medical-surgical units since its system went live in June 2001. By phasing in CPOE one unit at a time, Levick says LVH was able to avoid a "massive intrusion" into the physicians' existing workflow.

"We're using [CPOE] in some units, but not all. This allows doctors to use it in some settings, but doesn't force them to use it in all settings," Levick says.

Support and integration

Currently, use of the system at LVH is voluntary, and about 200 of the hospital's 750 physicians have been trained to use it. So far, reaction to the new system has been mixed, Levick says. Many physicians embraced the concept and have become adept at using it, while others who are still learning the system perceive it as a burden.

Still, physicians are encouraged to use it. Three dedicated trainers are available for one-on-one initial training and, when necessary, retraining. Technical support is available on-site for at least 10 hours per day for most patient units, and 24-hour support is available on-site for up to two weeks when a new unit is added to the system.
LVH is also in the process of providing each rounding physician with his or her own data entry unit -- a two-pound wireless sub-notebook computer -- that allows physicians freedom of mobility and eliminates the hassle of repeatedly having to log in and out of the system.

Data integration is perhaps the biggest technical challenge. For example, many hospitals use clinical and pharmacy records systems from two different vendors. Additionally, the databases that contain patient data such as laboratory results, pharmacy, and vital statistics such as weight, are usually independent of the CPOE product. All of these must be properly coupled to use CPOE, particularly the high-value functions like clinical decision support.

Yackel says electronic data definitions must be standardized to permit use with a rules engine or other decision support logic. Such definitions should be developed in part by the physician-led implementation team. Organizations also need to define the clinical decision support "rules" to be used by the system to determine orders and when to issue alerts (e.g., for potential interactions between drugs and allergies or drugs and laboratory values). Many vendor products contain starter sets of rules that can be customized.
"At the end of the day, don't reinvent from scratch. It's easier to copy than to invent," Rucker says. "Building order sets from scratch is the biggest single time cost on the implementation side."

Timelines and costs

A typical implementation takes about 18 months under the best of conditions, according to Metzger. A speedy startup is likely when the organization is committed, has developed its CPOE plan, and has a sufficient information technology network to handle CPOE and clinical decision support.

The Leapfrog Group surveyed hospitals using CPOE and found that system implementation can cost anywhere from $500,000 to $15 million, depending upon the size of the hospital and the status of its existing information system. Annual operating costs can range from $200,000 to $2 million. Generally, the Leapfrog Group concludes, CPOE could save hospitals between $180,000 and $900,000 annually, by reducing medication errors and adverse drug events; however, some hospitals reported annual savings of up to $5 million by reducing drug errors and using their CPOE systems to identify medical substitutions, increase use of clinical pathways, and promote clinical efficiency.

The JAMIA study found little change in total cost per hospital admission, which dropped to $5,661from $5,967 at OSU Hospitals, and rose to $6,518 from $6,427 at The James during the 10 to 12 month study period. Nevertheless, the study concluded that CPOE provides a good return on investment, so upgrades will be put in place over the next 18 months. The "next generation" system will include Web-based applications and voice recognition technology, and will be expanded to include outpatient CPOE as well.

"We weren't able to discern why the cost per admission didn't change significantly," Mekhjian says. "Another way of looking at the data is that this did not increase care. We believe safe care is cost effective because [it reduces] mistakes. The fact remains that I haven't heard too many people say they're against paying for safety of care."

Sandy Campbell can be reached at editor@physicianspractice.com.

This article originally appeared in the January/February 2003 issue of Physicians Practice.