A root cause analysis is defined as a retrospective approach to error analysis — the investigation of the direct or original error that led to an adverse event. In healthcare, such an analysis is typically reserved for tracing the origin of serious adverse events. “Near misses” are not reviewable under the sentinel event policy of the Joint Commission on Accreditation of Healthcare Organizations (JCAHO).
The “root” of the analysis of human error — not system error — dates back to the late 1970s, when an accident spawned an entirely new field of psychological study. As the story goes, “James Reason was making tea, and the cat was clamoring to be fed. He efficiently opened the tin of cat food — and put it in his teapot. The two components got mixed up. Both the teapot and the cat’s feeding dish afforded the same opportunity — putting stuff in,” writes Renee Lertzman, associate editor of saferhealthcare, an online patient safety resource that is a partnership among the British National Patient Safety Agency, the British Medical Journal Publishing Group, and the Institute for Healthcare Improvement.1
Reason, a cognitive psychologist, “suddenly realized a new research topic was literally under his nose. In tracing the causes of absent-minded incidents, Reason began an exploration of human error. Three decades later, Reason has become a leading expert on error and one of the recognized architects of the tools used to improve patient safety.”
After performing hundreds of interviews about absent-mindedness, Reason discovered that error results from under-specification. “In order to get something done correctly, you have to specify your plan of actions accurately,” Lertzman writes. “But if we under-specify, we are looking at a hazy picture; we go into ‘autopilot’ mode and fill in the gaps.”
Reason determined that instead of blaming the individual, it was more sensible to focus on the circumstances that led to the error.
What a “Root Cause” Is
JCAHO president Dennis O’Leary presented testimony about medical errors and adverse events to a subcommittee of the U.S. Senate in the year 2000.2 In his statement, he observed, “The solution to reducing the number of medical errors resides in developing mechanisms for collecting, analyzing, and applying existing information.” This, he explained, required five steps:
1. Creating a blame-free environment in which people would not be afraid to report a sentinel event (an occurrence resulting in the death or serious injury to a patient).
2. Performing root cause analyses to determine all the potential causes of the event.
3. Implementing actions to reduce the potential for additional (similar) errors.
4. Establishing patient safety standards that must be met.
5. Publicizing what was learned from the experience to all organizations at risk for a similar event.
In 1997, the Joint Commission began issuing occasional Sentinel Event Alerts, which disseminated the information that was learned from the event. These alerts include not only the problematic behaviors but also best practices and solutions.
“A root cause is the most basic causal factor or factors which, if corrected or removed, will prevent recurrence of a situation,” writes John Robert Dew, EdD, in an article published in the proceedings of the 56th Annual Quality Congress in 2002.3
“There is honest disagreement as to whether or not an error can be attributed to a single root cause ... or whether there will be a cluster of causes,” Dew adds. Dew presents five basic root causes:
1. Putting budget before quality
2. Putting schedules before quality
3. Putting politics before quality
5. Lack of understanding of knowledge, research, and education.
Dew also describes the difference in types of safeguards — physical (locks and walls) are best, followed by natural safeguards (distance) and time (limited exposure). And what he calls “information safeguards,” such as posters, alarms, etc., can also be useful. Other safeguards include measurement processes (tests, visual inspections), knowledge (charts, checklists), and administration (policy statements). The latter, however, he says, “are only as reliable as the administrators who enforce them.”
“When you’re looking at a root cause analysis, you have the ‘lumpers’ and the ‘splitters,’”explains Mark F. Gan, DO, medical director at El Paso Day Surgery. The splitters “split” everything into minutiae, and the lumpers “lump” data into easy composites of information, he adds. “For a lumper, a surgical case started around 9 a.m. For the splitter, it started at precisely 8:57 a.m. The people who do well in those root cause committees are splitters. If a patient falls out of bed in the OR, it goes to committee, and they have to look at the whole chain of events, including every department that was involved. When considering housekeeping, was the floor slippery?
When considering the OR staff, how many people were in the room? Were they looking at the patient? Were they all looking at X-rays on the opposite wall? Were they all talking to each other and not observing the patient? Was the patient asleep or awake? They’ll delineate every variable. It’s like that old game that uses a series of contraptions to get a single result — ‘Mousetrap.’ Just like that game, they’re backtracking every step to figure out the actual cause of the outcome,” he says.
In a surgical center or specialty hospital, there is still a quality assurance (QA) committee, and that committee still follows up after sentinel events, but it is all included under the QA umbrella. “In a hospital, they’re big enough and have enough staff to split up into subcommittees — morbidity and mortality, a tumor board, a root cause committee,” Gan adds. “In a surgical center, you have the economy of size thing, so you don’t have the time and physical staff to have too many committees. Instead, you make use of what you have, and QA handles all that.”
Reconstructing the Error
The Joint Commission has created a database of sentinel events; the organization collects and analyzes data from the review of the events, as well as the root cause analysis, action plans, and any follow-up activities or surveys.
JCAHO requires that the root cause analysis be both “thorough and credible,” meaning that the individuals completing the analysis have firsthand knowledge of the incident or processes that led to it. The leadership must support the findings, which should blame system-related defects rather than individuals; the analysis should be consistent; and the investigation should include a review of relevant literature, according to Briefings on Adverse and Sentinel Events by Opus Communications.4
“To be credible, root cause analysis requires rigorous application of established qualitative techniques,” write Heidi Wald, MD, of the University of Pennsylvania School of Medicine, and Kaveh G. Shojania, MD, of the University of California, San Francisco School of Medicine.5
After a sentinel event is identified, “a multidisciplinary team is assembled to direct the investigation. The members of this team should be trained in the techniques and goals of root cause analysis, as the tendency to revert to personal biases is strong. Multiple investigators allow triangulation or corroboration of major findings and increase the validity of the final results,” they write.
There are some drawbacks to the root cause analysis, however. “They are limited by their retrospective and inherently speculative nature,” write Wald and Shojania. “There is insufficient evidence in the medical literature to support root cause analysis as a proven patient safety practice; however, it may represent an important qualitative tool that is complementary to other techniques employed in error reduction ... The use of root cause analysis merits more consideration, as it lends a formal structure to efforts to learn from past mistakes.”
How to Do It
It is essential to have a plan of action in case a sentinel event occurs. The goal is to be proactive rather than reactive. JCAHO now requires that an organization it accredits have a definition of “sentinel event” in place so that it can institute procedures to identify, report, and managed such events. The commission also asks its accredited facilities to perform a root cause analysis and apply improvements to reduce the risk of a further event. It also requires the facility to monitor the efficacy of all of these developments.
“A root cause analysis focuses primarily on systems and processes, not individual performance,” according to the Joint Commission’s Sentinel Events policies and procedures.6
The action plan resulting from the root cause analysis should include all the strategies that will be used, and should address who is accountable for implementation, oversight, pilot testing, timelines, and follow-up reporting on efficacy of these actions.
The sentinel events that are subject to review by JCAHO include:
- Any incident that results in unanticipated death or major permanent loss of function (unrelated to the patient's underlying disease or condition)
- inpatient suicide
- Unexpected death of a full-term infant
- Discharge of an infant to the wrong family
- Rape or other sexual assault
- Blood transfusion error
- Wrong-site surgery
- Foreign object retention after surgery or other medical procedure
- Severe neonatal hyperbilirubinemia
- High radiation exposure, or radiotherapy to the wrong part of the body
When developing the root cause analysis, JCAHO recommends the facility incorporate the following elements:
- "A determination of the human and other factors most directly associated with the sentinel event and the processes and systems related to its occurrence;
- Analysis of the underlying systems and processes through a series of 'Why' questions to determine were redesign might reduce risk;
- Inquiry into all areas appropriate to the specific type of event as described in the current edition of ‘Minimum Scope of Review of Root Cause Analysis’
- Identification of risk points and their potential contributions to this type of event
- A determination of potential improvement in processes or systems that would tend to decrease the likelihood of such events in the future, or a determination, after analysis, that no such opportunities exist."
The root cause analysis should include participation by leadership in the organization and also by the individuals involved in the processes under review.
JCAHO emphasizes that all the root cause analyses and action plans will be treated as confidential.
Private companies may also provide root cause analysis, which, although not required by JCAHO for a “near miss” event, is still a good idea, to prevent an actual sentinel event from the same cause. For example, Medical Risk Management Associates (MRMA), LLC, offers software, training, and consultation to assist in risk management. This group suggests appointing both a root cause analysis facilitator and a separate team leader. The former is the expert on how to conduct a root cause analysis, while the team leader focuses on the content matter. These two roles determine what data to obtain, who should obtain it, and where it can be found. The company identifies those locations as “medical records, statements from personnel, maintenance records, instruction manuals, policy manuals, literature, etc.”7
The facilitator and team leader then set up a meeting time to begin the analysis; MRMA recommends inviting anyone from the affected department and creating a multidisciplinary team, and asking each invitee to bring along a written chain of events or observations, as well as suggestions for reasons behind the sentinel event and possible solutions. The meeting can begin with an initial list of each step in the process leading to the event, then a discussion of causes for each step.
The team leader and facilitator then develop a flow chart with all the elements leading to the sentinel event. Ask “why” repeatedly for each element, tracing each step back to the initial cause of that step — which becomes a “contributing factor.”
The next step is a second meeting with the previous participants, asking them for input on the flow chart. For the action plan, determine at least one corrective action for each contributing factor. Also create a list identifying who is responsible for the corrective action, by what date, and any follow-up activity that will be necessary.
In a third meeting, show the event sequence, flow chart, and the list to the team, allowing them an opportunity to provide feedback or suggestions for changes. “Identify as a team who should receive copies of the entire work, and who is responsible for distributing those copies,” suggests MRMA. “At a minimum, the communication in whole or part should include the heads of involved departments, all involved persons, the office of risk management and the office of continuous performance improvement. Don’t forget to at least discuss the findings with the person who reported the event.”
Make sure that follow-up work is done to determine that every element of the action plan has been satisfied.
Preventing Future Errors
Healthcare organizations are not required to report any sentinel event to the Joint Commission, but are strongly encouraged to do so. If the facility does not, and the incident is reported to JCAHO by someone else, such as a patient or the media, the healthcare facility is expected to perform a root cause analysis and determine an action plan within 45 calendar days, and submit these items to the Joint Commission. If JCAHO finds either element unacceptable, the facility may be placed on “Accreditation Watch” by the relevant committee.
If the events are reported by the facility, however, this carries with it several benefits — allowing the “lessons learned” to be added to the JCAHO database of sentinel events; opening a dialogue between the organization and JCAHO and allowing consultation to take place while preparing an action plan and performing the root cause analysis; and good old-fashioned public relations, demonstrating to the public that the facility is doing its best to ensure the problem does not happen again.
Once the Joint Commission determines that these steps have been met — root cause analysis and action plan — it will issue an Official Accreditation Decision Report that will state the commission’s intentions to continue or modify the current accreditation status of the facility, and will assign a follow-up activity to determine success based on the action plan.
1. Lertzman R. From absent minded to error wise: a conversation with James Reason. Published: 03 January 2006. www.saferhealthcare.org.uk/IHI/Topics/SafetyCulture/Literature/AbsentMinded.htm
3. Dew, John Robert. Using Root Cause Analysis to Make the Patient Care System Safe. Proceedings of the 56th Annual Quality Congress in Denver, Colo, 2002. http://bama.ua.edu/~st497/UsingRootCauseAnalysispercent20.htm
4. Opus Communications. Briefings on Adverse and Sentinel Events. July 1999. www.sentinelevent.com/credible.php
5. Wald H, Shojania K. AHRQ Evidence reports and summaries; AHRQ Evidence Reports, No.43. Making Health Care Safer: A Critical Analysis of Patient Safety Practices. Chapter 5. Root Cause Analysis. www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=hstat1.section.59569