By: John Robert Dew, Ed.D.
(Originally published in the Proceedings of the 56th Annual Quality Congress in
Denver, CO in 2002, and in the ASQ Health Care Division Newsletter, Spring, 2003)
SUMMARY:
The Institute of MedicineÕs Cross the Chasm Report identifies six key areas for
systematic improvement in health care. One of these areas involves patient safety,
ensuring that Òpatients should not be harmed by the care intended to help them.Ó
This presentation examines the manner in which root cause analysis can contribute
to patient safety.
There should be no doubt about the need to continue to examine the causes of adverse
and sentinel events that threaten the safety of patients using health care services. A quick review of the Joint Commission for Accreditation of Health Care OrganizationsÕ
Sentinel Event Alerts in 2001 provides confirmation for continued examination of this
topic. JCAHO issued alerts concerning patients catching on fire using oxygen equipment in their homes, transmission of blood bourn pathogens among staff through
needlesticks and sharps injuries, deaths caused by mix-ups in medical gases, cases of
transmission of a rare disease due to instruments used during brain surgery, and a
summary of causes of medication errors. JCAHO states that their sentinel event
database includes 150 reported cases of wrong site, wrong person or wrong procedure
surgery. Fifty-eight percent of the cases occurred in either a hospital-based ambulatory
surgery unit or freestanding ambulatory setting, with 29 percent occurring in the
inpatient operating room and 13 percent in other inpatient sites such as the Emergency
Department, according to JCAHO. Seventy-six percent involved surgery on the
wrong body part or site; 13 percent involved surgery on the wrong patient; and 11
percent involved the wrong surgical procedure. (1)
Patient safety depends on a complex system that includes appropriate and accurate testing, correct diagnosis, appropriate treatment, and avoidance of mistakes that can easily occur in administrative, laboratory, pharmaceutical, and treatment settings. Threats to patient safety are often categorized as adverse events and sentinel events. The Institute of Medicine defines adverse events as unintended injury to patients resulting from a medical intervention, which includes any action by healthcare workers, including clerical and maintenance staff.(2)
JCAHO defines sentinel events as Òan unexpected occurrence involving death or serious physical or psychological injury or risk thereof.Ó(3) The two terms describe different degrees of severity of an error, mistake, or unexpected event. It is generally useful to think of adverse events as the small to medium errors that are the pre-cursors to a sentinel event. Attention to the causes of the smaller problems can prevent the occurrence of a sentinel event. Adverse events, then, include a wide range of problems that are considered small when they do not involve death or serious injury. The use of an incorrect medication, incorrect dose of medication, or the misleading of a specimen are considered adverse events as long as they do not cross the threshold of causing death or serious injury. Sentinel events include the headline grabbing errors, such as mistaken amputations, sending an infant home with the wrong family, infant abduction, performing surgery on the wrong patient, and patient suicide. It is common for a sentinel event to be proceeded by smaller, adverse events and Ònear-missÓ occurrences that should have alerted people to weaknesses in the work system.
IMMEDIATE CAUSES OF ADVERSE OR SENTINEL EVENTS
Charles Kepner and Benjamin Tregoe have described problems as situations where there is a deviation that has occurred between what should happen and what actually happened. According to Kepner and Tregoe, a problem exists when a deviation occurs, the cause of the deviation is unknown, and it is important to discover the cause of the deviation.(4) Regardless of the severity of the event (which classifies it as an adverse or a sentinel event), it can be regarded as a deviation between what should have occurred and what actually occurred. It is important to know the immediate cause of the deviation.
To discover the immediate cause of this deviation, it is advisable to ask a structured set of questions that define the scope of the problem in terms of what occurred and when it did not occur, where the problem happened (and where it did not) and the extent of the problem. Kepner and TregoeÕs research has shown the benefits of asking questions that describe what the problem ÒisÓ and what the problem Òis notÓ in order to rationally identify the most probable cause, which can then be verified, either thru evidence or a test.(5) It is important, for both adverse and sentinel events, to quickly identify the immediate cause of a problem in order to take corrective actions and to identify other circumstances where the same immediate cause could affect patients or staff.
ROOT CAUSES
In order to keep adverse events from growing into sentinel events, and in order to prevent the recurrence of both adverse and sentinel events, it is important to move beyond the understanding of the immediate cause of a deviation to understand the root cause.
A root cause is the most basic casual factor, or factors, which, if corrected or removed, will prevent recurrence of a situation. There is honest disagreement as to whether or not an error can be attributed to a single root cause (something that has the absolute effect of a light switch) or whether there will be a cluster of causes. This may depend on the taxonomy of root cause definitions adopted by an organization. The methods of inquiry that constitute root cause analysis are useful for both the diagnosis and the prevention of adverse and sentinel events.
Root Cause Analysis is a questioning process that provides a structured method
to enable people to recognize and discuss the beliefs and
practices in an organization.
Root causes reside in the values and beliefs of an organization. Until the analysis
moves into this level, it has not begun to grapple with root
causes.
A variety of taxonomies exist to categorize root causes into headings such as less
than adequate management oversight, operator inattention to detail, and other groupings.
Sectors that have extensive experience with root cause analysis, such as nuclear power
operations, often experience exasperation with root cause categories that simply stop
with recognizing that management oversight is less than adequate. In seeking to categorize root causes, organizations often end up with a large number of causes in the
management category, with no further analysis.
An appropriate rule of thumb for knowing how deep to dig in conducting a root cause
analysis is to dig until you reach the point of admitting something that is really embarrassing about the organization, but not to go so far that your are in the domain of
theology.(6) In order to enable organizations to dig more bravely, a new taxonomy of
root causes that focuses on management issues that people are rarely courageous enough to discuss until a post mortem of an organization is being conducted, would be useful.
DEWÕS TAXONOMY OF ROOT CAUSES
1. Placing budgetary considerations ahead of Quality. (Fundamentally the organization does not understand the cost of poor quality.)
2. Placing schedule considerations ahead of quality. (We donÕt have time to do it right. Ready, Fire, Aim)
3. Placing political considerations (internal politics and external marketing) ahead of quality.
4. Arrogance. (The laws of nature and the laws of Congress should not apply to us)
5. Fundamental lack of understanding of the knowledge, research, education.
(Usually supplemented by belief in magic)
This taxonomy seeks to unearth the truly fundamental problems with management
systems in any organizational setting.
QUESTIONING TO THE VOID
One approach to root cause analysis is referred to as Òquestioning to the void,Ó or as ÒThe Five WhysÓ in Japan. It means to repeatedly ask how it is so that something happened. In some cases, asking, Òhow it is so?Ó five times can reveal the deeper root cause that made it possible for an event to occur. We donÕt simply ask why, but pose the question in a written of ways such as Òhow is it thatÉ?Ó
EVENT AND CAUSAL FACTOR ANALYSIS
Through interviews and review of logs and records, construct a time line that defines each discreet action and significant thought that led up to the adverse or sentinel event. Then, examine each action or thought, using the five why method in order to identify the root cause.
BARRIER/SAFEGUARD ANALYSIS
ÒBarrierÓ analysis was introduced by engineers in the Naval Nuclear Power program
to identify failures in processes and systems that led to accidents. As quality practitioners have focused on removing barriers to achieving quality, it has seemed contradictory to talking about introducing new barriers, hence, Meri Curtis suggested re-naming barrier analysis and calling it safeguard analysis for the health care sector.(7)
The steps are:
1. Identify the potential (or actual) source of the event and the target or victim.
2. Determine which safeguards were already in place to protect the target/victim from the source, and decide which safeguards are effective and which have either failed or appear weak.
3. Determine what additional safeguards could or will prevent the occurrence.
4. Develop a plan to strengthen existing safeguards and/or develop new safeguards.
HIEARCHY OF SAFEGUARDS
Not all safeguards are equal. Some are highly reliable and require a great deal of effort to circumvent. Others depend greatly on whether or not people are willing or able to follow rules.
In terms of effectiveness and reliability, there is a hierarchy of safeguards.
Physical
Natural
Information
Measurement
Knowledge
Administrative
Physical safeguards, such as locks and walls, are the most reliable safeguards. Natural safeguards such as distance (placing things out of reach or contact) and time (limiting exposure) can be effective. Information safeguards that caution people about hazards, such as labels, signs, and alarms, help people avoid problems, if their meaning is clearly understood. Measurement processes can become safeguards that ensure that work is being performed within safe limits. This includes tests, visual inspections, and other types of data collection. Knowledge safeguards, such as posting of information, checklists and charts, ensure that information is constantly available to staff. Administrative safeguards, such as policy statements, are only as reliable as the administrators who enforce them.
Some safeguards are easier to subvert than others. Administrative safeguards can be easily rendered useless under lax management. Knowledge safeguards can be undermined due to a lack of investment in quality education and continued training. Measurement safeguards can be circumvented when people falsify data. People can ignore information safeguards and can even cut through locks.
For these reasons, health care practitioners will carefully consider potential problems and will study actual adverse and sentinel events to develop a defense in depth. It is highly advisable to have a series of different types safeguards in place to ensure an event does not occur.
REFERENCES
1. ÒSentinel Event Alert: Issue 24,Ó JCAHO ,December 5, 2001.
2. Kohn, Linda et. al., To Err Is Human. Institute of Medicine, National Academy Press,
1999.
3.
Sentinel Events: Evaluating Cause and Planning
Improvements. JCAHO 1998.
4.
Kepner, Charles and Benjamin Tregoe. Problem Analysis and Decision
Making. Princeton Research Press. Princeton, N.J. 1979.
5. Ibid
6. Dew, John R. ÒIn Search of the Root Cause.Ó Quality Progress. March 1991.
7.
Dew, John R and Meri Curtis. Diagnosing and Preventing Adverse and Sentinel
Events. Opus Communications. Marblehead, MA 2001.