Root cause analysis (RCA) — the process by which health care organizations identify and learn from errors and potential errors — has been a pillar in the U.S. health system for some time, but experts note the results have been inconsistent at best. Hospital medical errors are now the third leading cause of death in this country, according to the Agency for Healthcare Research and Quality.
To find out how traditional RCA processes are falling short, the National Patient Safety Foundation (NPSF) assembled a panel of experts to examine the approach in depth and offer guidance aimed at elevating the RCA process so that it results in accurately identifying what’s behind errors or near-misses, and leads to actions that eliminate or drastically reduce the chances that the problem will happen again.
The results of this investigative process are a new set of guidelines aimed at putting real teeth in the RCA process and making sure efforts to identify the causes and potential causes of patient harm lead to sustainable system improvements. In fact, noting that too often RCA efforts fall short when it comes to implementing sustainable solutions, the authors of the guidelines have changed the name of RCA process to Root Cause Analysis and Action or RCA squared (RCA2).
While the complete guidelines are available via the NPSF website (www.npsf.org/RCA2), the co-chairs of the improvement effort recently held a web conference to answer questions about the process and offer direction on how health care organizations can get a much better return from their error-prevention programs.
Go beyond analysis
James Bagian, MD, PE, project co-chair and director of the Center for Health Engineering and Public Safety at the University of Michigan noted that too often RCA processes have failed because they lack a standardized approach.
“When we even use the term ‘root cause analysis’ it means different things to different people, which causes problems in learning from each other,” he says.
Bagian added that RCA processes often lead to superficial solutions such as telling someone to be more careful, which he noted is not very helpful or sustainable.
“There is often a lack of follow-up even when we have done a good analysis, which often we have not,” he says. “We don’t have good, robust action plans to go along with that.”
One of the primary reasons why the guideline authors decided to enhance the term used for RCA to RCA2 is because the former term implies that there is only one root cause, which is not generally the case, Bagian offers. He also suggests the RCA process should be more than just an analysis.
“If all we do is an analysis, then that doesn’t make any difference to the patient,” he says. “The point is what action do we take, and that is why we call it RCA squared.”
The new process, outlined in the guidelines, also addresses the issue of sustainability and follow-up, Bagian adds. “It is not enough to have good ideas. You must implement them and make sure that they really had the desired effect,” he says. “You also have to make sure you didn’t create new problems with your solution to the current problem.”
Incorporate transparency
A first step toward reducing harm is coming up with a way to identify which problems or errors you are going to put through the RCA2 process. Bagian notes that for too many organizations, the prioritization process is based on the level of patient harm that occurred.
“The problem is if you take such a risk-based approach, then by definition you will not be looking at close calls or near misses of a serious sentinel event,” he says. “That is not a good thing, because we first have to actually harm someone before we understand why it happened and do something about it. We don’t live our own lives this way. It doesn’t make sense, and yet, for the vast majority of institutions, that is exactly what they do.”
Instead, it is important to look at the severity of what did or could have occurred, and the likelihood that it might occur, Bagian explains.
“There are studies that show that a close call occurs — depending on the type of event — anywhere from 10 to 300 times more often than the event it is the harbinger of,” he says. “That is 10 to 300 opportunities you have to recognize it as a vulnerability and do something about it.”
The NPSF report offers guidance about how to develop risk matrices that incorporate severity categories and probabilities. However, when building this model for determining which errors or near misses to address through RCA2, the criteria need to be transparent so that there is no need for debate, Bagian advises.
“If you have that, there should be one individual or their backup who makes that determination, because it is a fairly straightforward thing,” he says. “Everybody understands how it is done, everybody should be striving toward the same goal, and you don’t have to dissuade the perception that you are cherry-picking just what you want to look at as an organization.”
Assemble a team
Who should serve on an RCA2 team?
“We suggest that at least one of these people be a subject matter expert in the area involved, and one should clearly be competent in the RCA process,” explains Doug Bonacum, CSP, CPPS, a project co-chair and vice president for quality, safety, and resource management for Kaiser Permanente in Oakland, CA. “One should be a frontline staff member who does the work but was not involved in the actual event. And — in a new recommendation — one should be a patient representative.”
The patient representative should not be someone involved in the actual event, notes Bonacum, but he adds that the authors thought that having this component would bring value to the investigative process.
“We think this is going to enable the RCA2 team to gain a more complete understanding of the circumstances surrounding the event, and may offer additional perspectives on how to reduce the risk of recurrence, which is our goal,” he explains.
The number of team members should be limited to between four and six members, Bonacum adds.
The idea behind this recommendation is that the group should be nimble and able to overcome organizational inertia, he says.
Further, he notes it takes into consideration the fact that health care organizations have limited resources.
“We feel team members need to be grounded in some basic understanding of human factors and safety science, and they absolutely need a team leader who is not only skilled in the RCA2 process, but also in dealing with people, in dealing with emotions, in communicating both in writing and orally as an expert, and, most importantly, staying on task,” Bonacum says.
Graphical descriptions
While many health care organizations like to map processes as they are intended to occur, the report authors recommend RCA2 teams also make a graphical description of the event as it actually occurred.
“By mapping what actually happened, causes and corrective actions may be more easily identified, and your goal of implementing sustainable, systems-based improvements more easily achieved,” Bonacum says.
The RCA2 team should interview anyone who might know anything about why the event or near miss occurred, but the report authors noted it is important to probe why the event occurred, not just what precisely took place.
Team members should also take advantage of the literature, their own past experience, and any other resources that might help them in pinpointing the true causes of the problem under investigation, Bagian notes.
When a problem or error emerges that merits analysis, the report authors noted the process should begin within 72 hours, and the investigation should be completed in 30 to 45 days.
“The longer you wait to figure out what the problem is and come up with countermeasures or mitigation actions to implement, the longer your patient population is at risk,” Bagian observes. “It certainly is more than unfortunate to have a problem, and then while you are waiting and not taking action, another patient is harmed.”
It makes sense to have established times when the RCA2 teams will meet so that if people are on call, they can take steps to ensure they are available ahead of time.
“Often people who don’t plan ahead and then try to bring up [meetings] in an ad hoc way have trouble getting teams together, and this delays their success,” Bagian says.
“The point is that you want to get your results as quickly as you can and deliver a good, well-thought-out job,” he adds.
Include specifications
When team members are ready to prepare the final causal statements regarding the event, they need to be mindful of the fact they will probably not be the ones who are approving the final implementation of any solutions or funding the necessary resources.
“Those individuals may not have the foresight that you have, so look at [this task] as if you are writing specifications,” Bagian says. “Describe your causes in a way so that … it actually lays out for you what the solution would really be.”
What does not work? Using negative descriptors such as saying that someone is poorly trained or that someone failed to act does not spell out a solution, Bagian warns.
“If a physician didn’t know it was [his or her] responsibility [to act], that is not the point,” he stresses. “Why didn’t they know it was their responsibility? Why didn’t [the health care system] set the systems so that they knew that?”
Take action
The authors stressed that the most important part of the RCA2 process is what happens as a result of the analysis.
“It is the implementation of actions to eliminate or control system hazards or vulnerabilities that have been identified in those causal statements,” Bonacum explains. “We want to ensure the actions and causes are tightly coupled.”
The NPSF report introduces the concept of an action hierarchy in which several actions may be required, but some are deemed stronger or longer-term than others.
“The action hierarchy is based on human factors and our understanding of error prevention and of systems safety,” Bonacum says.
For example, in the case of a piece of equipment that continually flashes an error code, Bonacum explains that a short-term fix might be to post instructions on how to clear that error code right on the device.
However, a permanent solution would be to work with the manufacturer to change the software or the device.
“That basically eliminates this thing from ever happening again, and it is the type of systems solution we want for our practitioners and absolutely for our patients,” Bonacum says.
Designate responsibility
To ensure actions or solutions are carried out, it is critical for specific individuals to be made responsible, Bagian advises.
“Committees aren’t responsible for anything. A person is,” he says. “It could be the chair of a committee, the chief medical officer, or a charge nurse … but the person needs to understand the buck stops with [him or her].”
Bagian adds it is not just a matter of designating who is responsible. There needs to be a specific date for when the action must be accomplished.
“People have to know what is going to be measured, how it is going to be measured, by whom, and what date,” he says. “Those are really important things. They are methodical things. But most experience will show this is seldom ever done.”
When measuring the effectiveness of RCA2-devised actions, Bagian says it is important to look at both process and outcome measures.
“If you only look at outcome, and the outcome didn’t change, and then you just suppose that people did the process, that is a big assumption,” he notes.
Further, Bagian emphasizes that the results need to be widely disseminated.
“Leadership has to know,” he observes. “The staff should know too because they spent their own blood, sweat, and tears reporting things, and some participated in the investigation or implemented countermeasures. They want to know if they made a difference; same thing with patients and families.”
When people realize that what they did made a difference, it will whet their appetite to do more quality improvement, Bagian observes.
“Celebrate the wins. Everything is not going to be totally successful, but then you learn and make iterative changes to make it work,” he notes. “Only by doing follow-up measurement do you even know about that, so these are really important things to kind of close the loop and make sure that you can continue to do better.”
-
James Bagian, MD, PE, Director, Center for Health Engineering and Public Safety, University of Michigan, Ann Arbor, MI. E-mail: [email protected].
-
Doug Bonacum, CSP, CPPS, Vice President, Quality, Safety, and Resource Management, Kaiser Permanente, Oakland, CA. Phone: (510) 271-5910.
