Safety Analyses Take on Engineer's Perspective at JCAHO's Urging
"Think like an engineerthe bad will happen; how can we stop it?" said Darryl S. Rich, Pharm.D., M.B.A., FASHP, JCAHOs associate director of surveyor development and management.
Rich offered his advice during a March 20 teleconference on failure-mode and effects analysis sponsored by the Institute for Safe Medication Practices (ISMP). The teleconference was organized to help health care practitioners comply with the JCAHO requirements, in effect since July 2001, that call for accredited hospitals to use failure-mode and effects analysis as part of an institutions patient safety program. The requirements are found primarily in the leadership chapter of JCAHO's Comprehensive Accreditation Manual for Hospitals.
The term failure-mode and effects analysis refers to exploring each step of a process and determining where an undesirable variation could cause the process to break down. "Engineers call that the 'failure mode,' " Rich said.
ISMP President Michael R. Cohen, M.S., D.Sc., said that failure-mode and effects analysis differs from root-cause analysis, another tool used to improve patient safety.
"Unlike root-cause analysis, which is really retrospectiveit takes place after an errorfailure-mode and effects analysis...anticipates errors and designs a system that will minimize their impact," Cohen said.
JCAHO requires hospitals to select for their preventive effort a high-risk process identified in the accrediting agency's publication Sentinel Event Alert.
"The organization has to, at least annually, select one high-risk process for proactive risk assessment," Rich said.
That selection, he said, must be based, in part, on the most frequently occurring types of sentinel events identified by JCAHO.
Rich added that pharmacy is a tempting target for failure-mode and effects analysis.
"Medication use is one of those high-risk processes," Rich said. "And if the organizations do select medication usewhich, at least based on our experience over the past eight months, a large number of organizations are doingthen clearly the pharmacy is going to be directly involved in that."
Rich cautioned hospital staff not to be overly ambitious when choosing a process for failure-mode and effects analysis.
"I think it's really wise to start out with something small, and then work your way up," he advised. "What you pick for your high-risk process doesn't have to be the entire medication-use process. It could be parts of it."
Pharmacist Mary Burkhardt, M.S., of the Department of Veterans Affairs (VA) National Center for Patient Safety (NCPS), likewise stressed the importance of keeping the project manageable.
"In Veterans Health Administration, we are looking at our bar-code drug administration program and what to do if that system is unavailable," she said. "Our experience here at NCPS suggests that, by narrowing the scope of the review, the team stands a better chance of [producing] a quality analysis that yields specific and effective actions."
Burkhardt said that VA relies on multidisciplinary groups in each hospital to conduct a failure-mode and effects analysis. In addition to people who have experience with the process that is being analyzed, VA includes employees who are unfamiliar with the subject.
"You want somebody who is not ingrained into the project and can actually step back and look at it with a fresh view," Burkhardt said. "It's also helpful to include someone with good flow-charting and software skills to assist the team in diagramming."
Rich concurred that working with flow charts is a critical part of conducting a thorough failure-mode and effects analysis. "There's a lot of software out there that can really help you with that," he noted.
After identifying the failure modes in a process, the team should look for ways to make the process safer without introducing new hazards into the system, and then check to see if the process really works.
"I'm familiar with a hospital that eliminated vials of sodium chloride [injection] because they were being used multiple times, and they were single-use vials," Burkhardt recalled. "So they went to syringes that came out on an automated dispensing device."
But eliminating the single-use vials created a new problem, Burkhardt said. "The staff in the patient care area used a work around and started drawing sodium chloride solution out of an i.v. bag." In other words, the patients were less safe after the small vials were removed, she said, a situation that could have become serious had the hospital not analyzed whether the procedural change had been effective.
Rich described hospitals' initial attempts to perform a failure-mode and effects analysis as "a learning process."
"Most of the organizations that we have surveyed on their first attempt at a failure-mode and effects analysis really do not do a good job," he said. "Some of them thought it was the same as a root cause analysis."
What is important to JCAHO, Rich said, is that each hospital does its own work and analyzes a critical process that is specific to the institution.
"We're not going to score you on how good it is at this point. We may, in a few years from now," he said. "But we will look to see that you actually did the process, and that is going to be critical on the survey."