HRO and RE: a pragmatic perspective

HRO and RE: a pragmatic perspective

This is a paper by Torgeir Haavik, Stian Antonsen, Ragnar Rosness, and Andrew Hale where they examine the difference between High Reliability Organizational theory and approaches and that of Resilience Engineering.

On the surface it may seem a bit meta, a bit of the sort of thing that we’d want to avoid importing from academia. Perhaps it seems pedantic, what’s the point? That’s actually very close to the view the authors take. They look at the difference between Resilience Engineering and High Reliability Organizations from, as they say, a pragmatic standpoint in order to avoid getting lost in pointless discussions and instead discover useful information.

Just as with other tools, understanding the difference between these two, will allow you to decide which one is right for where your organization or team is at a given time and what you’re trying to accomplish. I’ve also seen a recent uptick in teams that are pursuing resilience or similar goals and better understanding how these two approaches differ can help such a team with future study as well.

The authors explain “the main objective of this paper is to explore the differences that make a difference between HRO and RE.” And they do a pretty good job at that. The authors start by taking a question (since it is not much controversy between the two) and digging into it to discover a fruitful conclusion. This approach was inspired by Bruno Latour. Latour’s approach, the authors explain was to:

“…feed off controversies. Studying science in the making means to study subjects while they are still in the making when it is still possible to identify all the heterogeneous elements and process that go into their fabrication”


The “controversy” that they look at is one of safety, is safety a dynamic non-event or a dynamic event. This issue arises because the RE view is one of “symmetry,” that learning should take place from not just failures but also success, perhaps more so at success. As Erik Hollnagel says:

“Even if the probability of failure is as high as 104, there are still 9999 successes for every failure, hence a much better basis for learning”

HRO has a “preoccupation with failure”. However Kathleen Sutcliffe and Karl Weick say that this completely compatible with other approaches, like Safety II, because they consider safety as a “dynamic ‘non-event’”.

Hollnagel disagrees. He acknowledges the cleverness of the definition but instead calls safety a “dynamic event”. This might just feel like a semantic game, but it helps to explain how the two approaches see differently.

Hollnagel objects because if safety is a non-event, then how do you detect if it’s present? In order to fix this, he takes the approach that safety is a dynamic event. Now the event can be counted and the non-event (failures in this view) can be counted as well as normal.

Though on the surface this was a matter of wording, the authors have taken this difference and found that the two approaches treat what is visible and “the practice of seeing” differently.

One of the most important conclusions that they draw when comparing and contrasting approaches is that they differ in the “visible, invisible, and the culture of seeing.” I really like the way they summarize the issue. That each prospective is valuable but looks at things differently.

These raise the core questions:

“Do HRO and RE see the same? Do they look for the same?”

As soon as the question of something being invisible or visible is raised the authors rightly point out that the correct next question is visible or invisible for whom? The way someone works is often invisible to everyone else in an organization who does not perform that same work. Of what researchers see, Erik Hollnagel has said what you look for is what you find.

As a result the authors conclude that:

“the glasses of HRO are more suitable for accounting for structural and formal aspects of organisations than the RE glasses – which are particularly suitable for capturing situated work.”

“Situated work” being examining work in the context in which it is performed as opposed to independent of that context. This difference in which “glasses” are suitable for what may be influenced by the research backgrounds from which they emerged. This also influences what level of the organization tends to show interest in each approach.

HRO emerged from a background of political science, organization theory, and business schools. Whereas RE came from engineering backgrounds like human machine interaction and human factors. As a result, the authors noticed, that RE tends to garner interest in the level of the organization typically interested in engineering type endeavors, whereas HRO appeals to management portions more familiar with “abstract indicators of work.”

In practice

In order to ground this in reality and make the differences between these approaches more clear, the authors examine how each approach has been used in healthcare. I don’t think it matters all that much that they chose healthcare vs some other field, simply that they picked a field where both have been applied so we can examine the results. We can draw on those results to observe the contrast and also see where our organizations may face the same challenges.

The authors are explicit that their purpose in doing this examination is to:

“demonstrate that HRO and RE do not in general provide contradictory answers to the question of how safety is achieved, but they produce and cultivate different (although overlapping) discourses of safety, and provide new knowledge within those discourses.”


Two different groups of attempts have been made to apply HRO to health care. One focuses on the design of the organization and it’s structure, where as the other uses HRO as a way of analyzing how organizations deal with patient safety.

One study analyzed an organizational redesign of a pediatric ICU. In this redesign, nurses were given more authority and more training. The result showed a decrease in mortality rates while also having a low rate of turnover for the nurses and some other care staff. During this it took “vigilant efforts from unit leaders to buffer this fragile new organisational design” and ultimately the unit reverted back to its old practices.

This is unsurprising when examining early HRO research with focused on organizations that already had it’s own “natural buffers,” like nuclear submarines and aircraft carriers.

Later HRO research which has focused on things like “organisational mindfulness” and cultural parts of an organization. For example a framework to improve the culture in ICUs. The framework addressed many levels of the organization from the front line staff to the top management. It focused on “contextual factors that can exert lasting influences over organisational behaviour”. This, like other HRO applications are normative. Their goal is to find and implement “measurable reliability interventions.” They seek to understand “what works” and then apply that elsewhere.

This drive creates a stronger reliance on quantitative and experimental approaches, which differs from the observational approaches in early HRO research. This runs into a bit of trouble as some examinations of this trend has found that empirical research on high reliability cultures in health care was “geared towards developing better survey instruments rather than developing interventions and tracking changes over time.”

It’s possible that this shift is itself an adaptation to the organizations themselves and the economic pressures present in the domain. As the approach sought to influence the organization, so too did the organization influence the approach.


The authors begin their examination of resilience engineering in health care by pointing out criticisms of attempts to turn health care organizations into HRO. Christopher Nemeth and Richard Cook along with others who have argued that health care “is characterised by high variability, diversity, partition between workers and managers, and production pressure”. Or as they sum it up “a hospital is not an aircraft carrier” and as a result advocate for a focus on resilience and not reliability. (For more on the difference see Four concepts for resilience)

The authors summarize RE approaches as:

“to study work as actually done, through studying everyday work, and to focus on things that go right
(emphasis theirs)

This has persisted though even more recent research on resilience in health care, though some of the theory and vocabulary has developed some. The authors characterize much of RE’s contribution to health care as being a focus on “everyday practices of practitioners” such as nurses and surgeons. This can lead to the idea that RE only focuses on individuals and not organizations, which is not the case.

For example some recent research by Jeffery Braithwaite and Jennifer Plumb shows organizations as “networks” that lead “attention primarily to mundane actions and only subsidiary fomral arragnements” and they are quick to point out that this does not mean only focusing on lower levels of the organization:

“The discrete surgeon, or anesthetist or nurse, although important, is not the optimal focus of attention in safety studies. Resilience reside in and merges from collective not individual activities.”

RE looks at functional dependencies in every day work, HRO focuses on creating safety by building more, lasting organizational structure. As a result, HRO is much more normative than RE. HRO continues to expand its list of desirable characteristics for organizations. Whereas RE looks at understanding what trade-offs and adaptations are necessary and how to organize around that.

The authors summarize by saying:

“HRO studies how organisations work, while RE studies how ‘work works’. They study the same empirical phenomena, but they represent different cultures of seeing – one particularly tuned for organisations and organisational constraints, the other particularly tuned for sociotechnical systems, the same and different entities at the same time – they see and describe different kinds of events”

I think that’s a great and useful way to think of both methods and matches much of what we see in software, that we have a number of tools, many of which have some overlap, but each might be “particularly tuned” to one thing or another, which influences which we choose to use for a given task.


  • “HRO studies how organisations work, while RE studies how ‘work works’”
  • While RE and HRO take different views, they are both useful.
  • Understanding the difference between these two approaches can help you decide which one to use at a given time.
  • RE and HRO tend to appeal to different levels in an organization.
    • It’s possible this is due to where each originated, from more traditional organizational disciplines or engineering backgrounds.
  • HRO and RE primarily differ in what they pay attention to.
  • Early HRO studies were environments that were somewhat shielded from larger organizational pressures.
    • Making organizations without such a shielding into an HRO will likely take considerably more effort.
  • HRO is typically much more normative than RE approaches
  • At first resilience engineering can seem to focus on only individuals, but that’s not the case.
← Incident Reviews in High-Hazard Industries: Sense Making and Learning Under Ambiguity and Accountability
Safe operation as a social construct →

Subscribe to Resilience Roundup

Subscribe to the newsletter.