SafetyInsights.org

This study compared 22 conventional root cause analysis (RCA) style investigations against 22 Learning Teams (LT). I couldn’t see any additional info on what methodology they used in the RCAs. The investigation reports were compared and interviews conducted with staff familiar with both approaches.

LT are defined as a “facilitated conversation between those that do the work and those that design the work, to share operational intelligence between the two groups and improve system design.” (p1801).

In their version of LTs, two facilitated workshops are undertaken by a team of staff. The attendees are directly involved in the processes under investigation but weren’t necessarily involved with the specific incident. It’s said a focus is placed more on the process as a whole rather than any specific incident or person so as to avoid blame and get buy-in for more meaningful change.

Results

LT investigations resulted in a median of 7.5 actions compared to 3.5 for RCAs. Further, 57% of LT actions were system focused versus 30% for RCAs.

Differences between LTs and RCAs were observed in personnel selection, content of discussions, cultural items identified/discussed, challenges and outcomes. Some of these factors will now be covered.

1. Personnel involved

Interviewees identified that LTs drive different outcomes because they were “designed from the outset

to be inclusive events involving all people involved with the processes under discussion, generating better representation of the whole team, not just a focus on those team members directly involved in the incident” (p1802).

Interviewees also believed that people perceived to be lower in the organisational hierarchy were less often invited to RCA investigations; whereas their expertise is more highly valued under the LT framework.

2. Content of discussions

What people discussed during LTs was consistently seen to be different to RCAs, with more focus on normal work and less reliance on details of the specific incident.

In contrast, language in RCA investigations were seen to be more “focused around a single “specific incident” or “event” with significant time spent establishing exactly what went wrong.” (p1802). It was also felt that RCAs focused more on the “last barrier to fail”, whereas LTs focused more holistically on other barriers and factors that influence operational work.

RCAs were also seen to focus less on “why” particular events occurred in the context of the setting, context and environment. LTs were seen to provide more flexibility for discussions (e.g. more complex themes and multiple narratives), compared to RCAs which were seen to have a tendency to distil many factors into a string of few root causes; risking oversimplification.

The greater flexibility in LTs was seen to allow more transferability of the lessons to other areas and better facilitate more system-focused actions.

3. Culture

A difference in the “culture” between LTs and RCAs was observed by interviewees. LTs were felt to be more supportive and open to learning, with less fear of interpersonal risk. LTs also favoured less technical themes, thereby allowing people with less technical prowess to participate freely.

In contrast, it’s believed that RCA investigations foster more language around blame and mistake at the individual level. People used words like fear and guilt and described an environment where they were less likely to speak truth to power.

4. Challenges

While people were positive about LTs, importantly it’s said that they’re not suitable for every occasion; rather, one tool with another range of options.

RCAs are well established and familiar to people compared to more of the “unknown” with LTs. Also, people lower in the hierarchy found it harder to get the time to attend LTs, and a potentially higher time investment was necessary with LTs. On the plus, people that attended the first session were often highly engaged and interested in attending the second LT workshop.

LTs were seen to challenge status quo more often than RCA and seek more innovation solutions. Further, LTs were seen to focus more at the coal face compared to RCAs which were seen to reflect more from senior management.

The findings are then discussed. In this sample, LTs delivered more system-focused actions of a higher quality. They also focused on more broader issues compared to RCAs which focused more on singular incidents.

Moving back to which technique is suitable for what occasion, it’s again recognised that LTs are not more suitable for all situations. According to the paper, RCA investigations may be more suitable for a specific incident that is less likely to recur outside of the specific context of that event.

They also believe it’s of limited benefit to repeat an LT for a repeat event if the first LT was comprehensive. Rather, in this circumstance, an RCA may be more efficient to capture any changes since the LT.

They conclude that “There may not be one “right” investigative methods, and rather incidents lie on a spectrum, and therefore, judgment needs to be applied to establish which investigative methods to choose” (p1804).

Some limitations were present. One is that LTs aren’t randomly allocated to events but selected based on their perceived utility in that specific event; this may introduce a bias. The methodology was also retrospective rather than prospective. Also, the semi-structured interviews may also have some bias, since they involved only those people involved with chairing LTs (but all of those people also had extensive experience in RCA investigations). Nevertheless, a potential bias cannot be eliminated.

Finally, no follow-up in the future was possible to see if either method resulted in better long-term change.

Overall, they summarise some key findings as:

• LTs involve all staff (even those not involved in the incident)
• LTs more often focus on systems problems compared to RCA (which focus more on staff performance problems)
• LTs were believed to better empower front-line staff to identify solutions

Authors: Tim Robbins, Stephen Tipper, Justin King, Satya Krishna Ramachandran, Jaideep J Pandit, Meghana Pandit, 2021, Journal of Patient Safety, Volume 17, Number 8.

Study link: https://doi.org/10.1097/pts.0000000000000641  

Link to the LinkedIn article: https://www.linkedin.com/pulse/evaluation-learning-teams-versus-root-cause-analysis-large-ben