St. Barnabas Health System’s Breakthrough Around Wrong-Patient CPOE Medication Errors
St. Barnabas Health System treats the most critically ill and severely injured patients. As a New York State-designated Stroke Center and AIDS Center, SBH Health System provides access to much-needed services. To reduce the risk of errors when using computerized provider order entry (CPOE), SBH Health System convened a multidisciplinary team to review its ordering process, finding several contributing factors--including interruptions, alert fatigue and the lack of an active patient identification process--could lead to wrong-patient errors.
On Tuesday, June 4, at the Mid-Atlantic Health IT Summit, sponsored by Healthcare Innovation and held at the Hilton City Avenue in Philadelphia, Zane Last, Pharm.D., director, healthcare analytics, a and business intelligence, and director, IT-care management at the Bronx-based St. Barnabas Health System, provided a featured presentation at the Summit, entitled “A Validated Strategy to Reduce Error in Electronic Orders.”
The SBH Health System, founded in 1866, is a not-for-profit, nonsectarian, 435-bed teaching hospital and a level 2 trauma center. It is the fifth-largest PPS (provider performing system in New York state; it’s payer mix is 90 percent Medicaid and Medicare. On the ambulator side of operations, the organization delivers over 500,000 outpatient visits annually, and is one of the largest providers of mental health services in the Bronx.
The object of the initiative, Last told his audience, was to determine the proportion of incorrect medication errors being placed in the CPOE (computerized physician order entry) system, and to implement mechanisms to stop those errors while they are in process. “It’s well-known that CPOE systems have a lot of positive effects; they can prevent a lot of errors. But they can also facilitate existing errors, such as wrong-patient errors,” he noted. As a result, he said, “our recognition of our errors led us to build out an activity reentry process, wherein we changed our process to immediately reidentify patients before the next meds order.”
What Last and his colleagues did was to develop alerts in their electronic health record (EHR) that required the ordering clinician to verify the identity of the patient. Doing so enabled a significant reduction of near-miss, wrong-patient CPOE orders. The findings of this project were published in the Journal of Clinical Outcomes Management in December 2016.
The first step, Last explained, was to create an active-reentry hard-stop alert and build it into the system. What that means is that ordering clinicians were prompted to reaffirm the patient’s identity. Specifically, every clinician was challenged to enter the initials of the patient, last name initial first, and then the patient’s year of birth, in placing orders. What’s more, if the clinician makes a second error, that clinician is locked out, and has to try again, or has to change which patient he or she is looking at. And in developing that challenge/hard stop, Last and his colleagues first assessed the problem using data analytics, and determining where the likelihood of erroneous entry was occurring, and devising a workflow, which led to the hard-stop alert.
The requirement placed upon prescribers was in line with TJC’s National Patient Safety Goal,” calling for two patient identifiers when providing care, treatment, and services, he noted.
Meanwhile, once the hard-stop alert was built into the workflow, it underwent one week of beta testing, and then an eight-week pilot project for testing, during October-December 2014. It went fully live across the organization in November 2015.
Speaking of the process, Last told his audience that “We needed to do a good process review and understand what was going on, where the workflows were, and where we could interject changes.” And, in that, he and his colleagues got buy-in from senior leadership, from emergency department leadership (where the initiative was piloted), and from front-line staff. What’s more, they had representation from all disciplines involved with the CPOE process, including nursing, pharmacy, radiology, clinical laboratory, information technology, and analytics. “It’s really important to bring in the appropriate stakeholders, including senior leadership, ED leadership, and front-line staff,” he noted. What’s more, he said, “Senior leadership were on board from the beginning.”
What were the critical success factors involved? “One of the things that we did,” Last told his audience, “was to go through a really rigorous process of really understanding the workflows, and end up with a fishbone diagram. And we found there were too many orders relative to the number of patients involved. And in our information system, providers were only allowed to order one med at the had shown a 35-percent reduction in wrong-patient orders during the we same time. There were multiple orders. Too many orders, and alert fatigue. Environment of care issues: interruptions to providers; ED design; there were EHR environment issues: too many clicks were needed on the part of prescribing clinicians.”
What’s more, Last noted that, at the outset of this initiative, “On the physician side, there was no active process to ID patients; also, there were multiple log-ins allowing two patients to be open on two different screens, with potential for picking up the wrong patient. And lastly, in the EHR, there is a specific status board within the ED; no active ID process within the EHR, and the opportunity for multiple log-ins.”
The results were quite startling: the ID reentry requirement decreased near-miss wrong-patient orders in the ED by 35 percent during the eight-week pilot period. The system was also successful in bringing down the percentage of all CPOE near-miss events by 49 percent across the organization. With 37 percent of medication order errors occurring in the ED prior to the pilot, that meant one near-miss per day in the ED.
Meanwhile, Last told his audience, “We also did time studies. We aware that any new requirement added time. The reentry process at first was taking 6.2 seconds to do; with experience, that brought the delay time down to 4.0 seconds.”
Looking back on the experience of the initiative, Last told his audience that:
Ø Communication with all departments was important
Ø Other IT initiatives were going live at the same time, so we had to be aware of that for the IT department
Ø The ED had been experiencing 37 percent of all errors, so that area naturally was the main focus of effort, with the other being the ICU (intensive care unit)
Meanwhile, Last noted, errors can still occur, as residents sometimes create workarounds—“but we were able to keep those workarounds down to a minimum.” And in that regard, he added, “Workplace interruptions were a significant factor; we’re working to create no-interruption zones.”
As for multiple sign-on sessions, he said, “We looked at that really carefully. The problem is that if we disallow providers from having more one session at a time, that would affect providers across the whole organization. We couldn’t block it only for the ED, so we decided to leave it open.”
Among the key lessons learned, Last said, are that “It’s important for organizations to develop strategies for these kinds of situations. Developing the continuous monitoring retract and reorder events tool was something we refined. We found that the another aspect of this involves involving the physicians to engag10-minute timeframe was the most effective strategy.” In addition, he said, “I think that involving key stakeholders from the beginning is one of the most important critical success factors” in an initiative like this one.
