Posts Tagged ‘fatal medical mistakes’

Follow Up: A Deeper Look at Root Cause Analyses

November 28th, 2010 by avatar

Last week we took a look at root cause analysis (RCA) in a real case of hospital maternal death. We discussed the importance of repeatedly asking “Why?” at each step of the process (the 5-Whys approach) that led to the fatal error in which an epidural medication was accidentally infused into a laboring woman’s central bloodstream. We saw how grouping contributing factors into categories and depicting them on a fishbone diagram can be a helpful aid for understanding and identifying root causes.

Although widely used throughout the health care industry, the 5-Whys and Fishbone methods are not the only, nor even necessarily the best, assessment tools for conducting an RCA. There is no “one size fits all” root cause analysis approach. Some RCA experts argue that the 5-Whys method implies that failure results from only one linear sequence of events (e.g., the nurse hung the wrong bag because she was tired because she worked a double shift because hospital policy allowed for chronic staffing shortages) thereby missing other potential causes of system failure. Others point out that the Fishbone approach is also overly simplistic, failing to demonstrate the interactions between some root causes and potentially missing others.

A reader comment posted last week by a career RCA analyst spurred me to critically re-examine the scientific rigor and validity of my own blog post (never let it be said that Science and Sensibility bloggers are resistant to constructive feedback!).  Bob Latino is a nationally recognized expert in RCA and CEO of Reliability Center, Inc., a consulting firm specializing in improving equipment, process, and human reliability.  Mr. Latino is also the author of several books on root cause analysis for patient safety in health care settings. In order to address the shortcomings of the 5-Whys and Fishbone methods, Mr. Latino developed the PROACT Logic Tree as an alternative tool for use within RCA. He took the same St. Mary’s case that we used to construct the fishbone diagram and applied it to his Logic Tree (pdf).

Understanding the Logic Tree:

The Logic Tree is read from top to bottom as a reverse time line by following the red boxes. At the top of the tree is the Event (red box), which is the negative outcome that triggered the need for an RCA—in this case, a maternal death. Underneath the Event are Modes (red boxes), which are the manifestations of the failure that lead to the Event—in this case, cardiovascular collapse.  It is important to remember that the Event and the Mode(s) are known facts, and the Event is the last “effect” in the chain of factors.  From the Event and Mode(s), we go backwards in short increments of time to recreate the path to failure.

From the point of the Mode(s), instead of asking Why? as we did with the fishbone diagram, the Logic Tree instead asks How could? The answers comprise the tree’s branches and are called Hypotheses. For example,

How could the patient have experienced a sudden cardiovascular collapse?

  • Hypothesis 1: Because she experienced a fatal allergic reaction to medication.
  • Hypothesis 2: Because she had an underlying heart condition.
  • Hypothesis 3: Because she had an undetected serious illness such as sepsis.
  • Hypothesis 4: Because there was a malfunction in the medical equipment.
  • Hypothesis 5: Because there was an error in diagnosis (e.g., we thought she was in labor when she was really having a heart attack).
  • Hypothesis 6: Because there was an error made in medical treatment.

As you might imagine, not all of the hypotheses are true and some can be ruled out fairly quickly. The point is that, whether or not the hypotheses turn out to be true, by asking How could?, the Logic Tree method systematically seeks out all the possibilities (not only the most likely) and then uses evidence to back up what did and did not occur.

We keep asking How could? and follow the cause-and-effect chain back down the Tree until we get to the Root Causes (circled boxes). There are three progressively in-depth levels of Root Causes: Physical, Human and Latent.  Physical Root Causes (purple boxes) are the immediate, observable consequences of a poor decision (e.g., patient received wrong medication).  Human Root Causes (blue boxes) are decision-making errors (e.g., nurse picked up the wrong medication bag).  Deepest are the Latent Root Causes (brown boxes) which are flawed organizational errors that feed bad information to the decision-makers (e.g., inadequate supervision and training).

According to Mr. Latino, the main advantages of the Logic Tree are that it:

  • Supports tightly-coupled cause-and effect relationships,
  • Requires evidence to back-up proposed Hypotheses, and
  • Requires in-depth understanding to identify system flaws that contribute to human errors as a result of poor decision-making.


The RCA is only the first step to remedying real and potential problems in our maternity care delivery system. Mr. Latino has graciously offered to extend his expert analytical skills to a project I proposed in my last blog post, which I will repeat here. Neither Mr. Latino nor I are birthing professionals, so here is your chance to weigh in! Read Rima Jolivet’s thought-provoking allegorical tale of two births As you compare the two women’s stories, consider the factors that contributed to Karen’s negative birth experience. Even if the causes were not stated explicitly in the article, draw upon your own experience as a birthing professional and fill in the gaps. Think about:

  1. Patient characteristics: Are there pre-existing or co-morbid medical conditions, physical limitations, language and communication barriers, cultural issues, social support needs that play a role?
  2. Task factors: What protocols and procedures are in place for labor and delivery, for use of analgesia, for dystocia, for C-sections? Are they safe? Are they practical? Are they effective? Are they consistently applied?
  3. Individual staff: How did the knowledge, skills, training, motivation, and health of Karen’s providers affect her care?
  4. Team factors: How well do the various health care professionals involved in Karen’s care work together? What is the nature of the communication? Are there hierarchies? What is the responsiveness of nursing supervisors or attending physicians? How easily can a team member ask for help or clarification?
  5. Work environment: Is the labor and delivery unit adequately staffed? What is the workload? What happens when the census fluctuates unexpectedly? What is the staffing level of experience, functionality of the equipment, quality of administrative support?
  6. Organizational and management factors: How do the values of the hospital translate into clinical practice? Do their standards and policies focus more on patient safety and quality of care, or volume and speed? Are management’s priorities patient- or provider-centered? Does senior leadership foster a culture of teamwork and safety or blame and shame?

Add your comments below, and we will include it in the next blog post!

Posted by:  Tricia Pil, MD

Do No Harm, Patient Advocacy, Uncategorized , , , , , , , , , ,