Root Cause Analysis: Turning a needless maternal death into better care for all

November 18th, 2010 by avatar

On the morning of July 5, 2006, a 16-year-old patient came to St. Mary’s Hospital in Madison, Wisconsin, to deliver her baby. During the process of her care, an infusion intended exclusively for the epidural route was connected to the patient’s peripheral IV line and infused by pump. Within minutes, the patient experienced cardiovascular collapse. A cesarean section resulted in the delivery of a healthy infant, but the medical team was unable to resuscitate the mother. The medication error and its consequences were devastating for the patient’s family, the nurse who made the error, and the medical team that labored to save the patient’s life.

This is the real story of a tragic and unnecessary maternal death that occurred not in a mud hut in a third world country, nor in a backwater rural health clinic—but in a fully licensed and accredited 440-bed community teaching hospital that delivers more than 3,500 babies annually and serves as a regional referral center for all of south-central Wisconsin. In a highly unusual and commendable move, senior management at St. Mary’s requested an outside independent investigation of this event and published their findings in an effort to share painful lessons learned with the medical community and the public.

What happens when an unanticipated maternal death occurs? If the event occurred in a hospital accredited by the Joint Commission on Accreditation of Healthcare Organizations (“Joint Commission”), the hospital must complete a root cause analysis (RCA) as a first step. Since 1996, a total of 84 cases of maternal death have been reported to The Joint Commission. The lessons learned from these most extreme patient care outcomes, also called “sentinel events,” have widespread implications for everyone involved in maternal and infant care. As William M. Callaghan, M.D., M.P.H., senior scientist in the Division of Reproductive Health at the Centers for Disease Control and Prevention remarked, “Maternal deaths are the tip of the iceberg, for they are a signal that there are likely bigger problems beneath – some of which are preventable,” says Dr. Callaghan. “It is important to consider the women who get very, very sick and do not die, because for every woman who dies, there are 50 who are very ill, suffering significant complications of pregnancy, labor and delivery.”

What is a root cause analysis?

Root cause analysis, or RCA, is “a process for identifying the basic or causal factors that underlie variation in performance, including the occurrence or possible occurrence of a sentinel event.” The RCA seeks to answer these questions: What happened? Why did it happen? What will we do to prevent this from happening again? The RCA is not about assigning blame, but rather identifying the direct and indirect contributing factors –latent system errors—that create the “perfect storm” in which the event occurred.

The RCA process might seem deceptively simple. We may be tempted to approach the RCA in the following manner:

What happened? The patient mistakenly received an IV infusion of epidural medication.

Why did it happen? The nurse hung the wrong IV bag.

What we will do to prevent this from happening again? Fire the nurse.

Indeed, this overly simplistic and ineffective “shame and blame” approach is the one that many hospitals take in conducting internal investigations of adverse medical events. A more thorough and credible RCA digs at the underlying factors and causes by asking a series of “Why?” questions, which might look something like this:

What happened?

The patient mistakenly received an IV infusion of epidural medication.

Why did it happen?

The nurse hung the wrong IV bag.

Why did the nurse hang the wrong IV bag?

Because she confused the epidural bag with the IV penicillin bag which were next to each other on the counter.

Why were the bags next to each on the counter?

Because the work flow process included having epidural medications and supplies set up and ready in the room ahead of time.

Why did the work flow process include having analgesia medication in the room ahead of time?

Because anesthesia had in the past expressed dissatisfaction with nursing staff over patients’ state of readiness for epidurals.


Why did the nurse hang the wrong IV bag?

Because she confused the epidural bag with the IV penicillin bag.

Why did the nurse get confused?

Because she was tired.

Why was the nurse tired?

Because she had worked two consecutive eight-hour shifts the day before, then slept in the hospital before coming on duty again the following morning.

Why did she work consecutive shifts?

Because she was covering for another colleague and her departure would have left the unit inadequately staffed.

As we repeatedly ask “Why?” we start to see groups of factors emerge, and these groups can help us to organize our thinking and later, to identify remedies. These groups might include: patient characteristics, task factors, individual staff factors, team factors, work environment, and organizational and management factors. We can map these factors and groups onto a fishbone diagram, a commonly used RCA visual aid:
(Click on graphic for improved viewing)

Now you try it!
Although root cause analyses are most commonly performed in cases of serious permanent physical or psychological harm, we can apply these same principles to “near-miss” events and instances of suboptimal, although not lethal, care. 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 I will include them in a root cause analysis of Karen’s case in my next blog post.


Jolivet, R. “Two Birth Stories: An Allegory to Compare Experiences in Current and Envisioned Maternity Care Systems.” Childbirth Connection, 2010. http://www.childbirthconnection.org/pdfs/allegory_illustrating_vision.pdf

“PS104: Root Cause and Systems Analysis.” Institute for Healthcare Improvement Open School for Health Professions. http://www.ihi.org/IHI/Programs/IHIOpenSchool/.

“Sentinel Event Alert: Preventing Maternal Death.” The Joint Commission. Issue 44, January 26, 2010. http://www.jointcommission.org/SentinelEvents/SentinelEventAlert/sea_44.htm

“Sentinel Event Policy and Procedures.” The Joint Commission. July 2007. http://www.jointcommission.org/NR/rdonlyres/F84F9DC6-A5DA-490F-A91F-A9FCE26347C4/0/SE_chapter_july07.pdf

Smetzer J, Baker C, Byrne FD, Cohen MR. “Shaping Systems for Better Behavioral Choices: Lessons Learned from a Fatal Medication Error.” Jt Comm J Qual Patient Saf. 2010 Apr; 36(4):152-63. http://psnet.ahrq.gov/public/Smetzer-JCJQPS-2010-s4.pdf

Uncategorized , , ,

  1. November 18th, 2010 at 10:27 | #1

    OMG!! I am actually in tears reading this because I understand the need for this type of analysis of sentinel events and I know the scenario that you paint personally and vividly (a tired nurse, annoyed anesthesia docs, short staffed) It happens in hospitals all over the country and while they fortunately don’t always result in death there are an awful lot of near misses. Medication errors happen all the time even with good prudent nursing care and we have to start really looking at the root of the problems so we can make hospital care safer for everyone. As nurses this is definitely a fear of severe disciplinary action including loss of job so many things are not even reported.

    Just a couple of weeks ago I was in a chaotic twin delivery. The doctor yelled out an order that did not sound right or safe but the RN was moving to carry out the order. I leaned over and said… “I don’t think you want to give that med right now because …. ” and it was like I woke the nurse up out of a trance. Later she came and hugged me and thanked me for intervening. She had been called in to help out. She was “thrown” into the OR for the birth with little report/information. The OR was loud and the MD was being impatient. It was a near miss but it could have easily turned fatal. We have to move to a place where we are doing more of these RCAs and using them to learn and teach not to punish. Thanks for sharing and for doing this type of work/reasearch.

  2. November 18th, 2010 at 16:05 | #2

    Absolutely we need more RCAs! In my previous life I had a corporate job and whenever we had a mistake involving a client, we had a RCA meeting. They often got heated as blame was passed, but in the end we would truly find the root cause. Once we brought that root cause to the client with an apology and an action plan, things were almost always smoothed over. The mistake wasn’t made again because we had active communication in the office.

    This can translate well to any industry, medicine included. RCAs need to be done and then the information needs to be passed to all the hospital staff who would be affected by the results. Maybe even passed to other hospitals, in a central database (in keeping with the HIPPA rules)

  3. November 21st, 2010 at 17:13 | #3

    Thank you for this excellent analysis. In Australia we have no national agreement on a definition of maternal death and RANZCOG estimates maternal death to be around 30% under-reported. While the doctors fill our headlines with lies about homebirth, women are dying unnoticed by anyone but their families in our hospitals. Chilling. There was recently a highly publicised epidural accident in a major hospital here too. The woman hasn’t died but she will probably never recover and is completely unable to parent her baby. Accountablility? What’s that? “Routine” epidurals have risks, these are only some of them.

  4. November 21st, 2010 at 17:49 | #4

    I’ve been involved in several RCAs, and can say that they are always both painful and enlightening. As you have correctly pointed out, it is always easy to point to a simple reason an error occurred, when in reality it almost always takes a long series of errors and misadventures to hurt a patient. We have so many systems of checks and balances that when major errors occur, we are often surprised how all of our defenses failed. These analyses can be painful, but they usually lead to systemic improvement.

    Sharing an RCA with the family of an affected patient is a great idea, and hopefully is the policy followed by hospitals. It is a difficult thing however, as one side of that coin is transparency and openness, while the other is giving a paint by numbers for a plaintiff’s attorney. Often though, an RCA case is one that would not be defended if a suit were brought – like the case you mention.

    The one other comment I feel I have to make is that the idea that you are going to publish your own RCA of this case is a little troubling to me. It is completely impossible that you are actually going to do an RCA, as you do not have access to the medical records and people involved in the case. An RCA based on hearsay is not an RCA, it just editorialization of information already made public. Editorialize away, but please don’t claim to be doing an RCA. Its been done already by the only people who could.

  5. avatar
    Tricia Pil
    November 21st, 2010 at 22:06 | #5

    @Nicholas Fogelson, MD

    Thank you, Dr. Fogelson, for your feedback and comments.

    The idea of sharing RCAs with injured patients and families is a new and fairly controversial one, and I would venture to say not the policy followed by most hospitals at this point. On the one hand, involving the patient and members of the healthcare team in an RCA can lend an invaluable firsthand perspective of the events that occurred, as well as offer patients and clinicians the opportunity for healing and reconciliation and tangible proof of organizational transparency and openness to change. On the other hand, for some patients and families, particularly those who suffered severe injury or death, RCA participation can be emotionally overwhelming and retraumatizing.

    Although fear of litigation is often cited by providers as a major reason to exclude families from participating an RCA, studies have shown that most patients file suit for non-economics reasons such as recognized cover-up, desire for more information about the event, vengeance, desire to protect others from experiencing similar harm in the future. (e.g., “Factors that Prompted Families to File Medical Malpractice Claims Following Perinatal Injuries”, http://jama.ama-assn.org/cgi/content/abstract/267/10/1359).

    For an excellent review of the pros and cons of patient participation in RCA, see this article from the Institute for Safe Medication Practices

    As for your comment that I am planning to publish my own RCA, I am confused. If you are referring to the St. Mary’s case, then yes, the RCA was done already and published, courageously so, in the public domain. Everything in my blogpost about the St. Mary’s case was drawn directly from that article; I’m not claiming to have done anything novel here. If you are referring to my future blogpost plans for an RCA of Rima Jolivet’s allegorial “Tale of Two Births,” Karen’s case is a hypothetical one. Even so, Karen’s story is not unlike thousands of other real suboptimal and psychologically harmful, though not deadly, birthing experiences that occur each year due to reasons some of which are preventable. And who better to identify those reasons and factors that would go into such an RCA than the midwives, doulas, L & D nurses –and obstetricians too– who labor (no pun intended) in the birthing trenches?

  6. November 22nd, 2010 at 07:41 | #6

    I am a career RCA analyst as well as an author of several texts on related topics in Risk Management. I have taken the liberty of constructing a Logic Tree of this case per the written text. I want to communicate this graphical expression of cause-and-effect logic to those that have expresed interest to this article including the author. I have found that communicating RCA results in this fashion makes it much easier to convey the true story of such an unfortunate incident to others unrelated to the event itself. However I do not see where I can attach a .pdf for review. I will try this route and see if you are able to contact me and request the .pdf and then provide feedback as to how to improve the expression. My email is blatino@reliability.com.

  7. November 22nd, 2010 at 08:25 | #7

    In follow-up to Dr. Fogelson’s concern about Dr. Pil’s plan to “publish your own RCA of this case” I would just like to reiterate Tricia’s intention: what she plans to publish, via this blog community, is the responses from commenters who chose to try, and submit here, their own attempt at an RCA, based on the fictional (but representative) case of “Karen’s” birth. (Thus, the descriptor, allegorical.)

    Regarding Nick’s mention of including families in RCAs–this is something we may explore in a future post, as I think it brings up an additional, important, component to the RCA process. Thank you, Dr. Fogelson, for highlighting this.

    Thank you, as always, to everyone’s interest and contribution to the dialog here on Science & Sensibility.

  1. December 17th, 2010 at 11:26 | #1
  2. February 11th, 2011 at 08:15 | #2