Recent Blog Posts

Safety Metrics: Are They Measuring Up?

More

Learning to Deal with Drift

More

Enacting resilience in everyday practice: steps along the journey

More

The Learning Review -- Looking up the River

More

Practical Resilience: Misapplication of an Important Concept

More

Topic

Board Education Governance Health


Archive

2017 (4)
June (1)
April (1)
February (1)
January (1)
2016 (3)
September (1)
June (1)
April (1)
2015 (3)
May (1)
March (1)
January (1)
2014 (5)
October (1)
September (1)
June (1)
May (1)
January (1)
2013 (4)
November (2)
October (1)
September (1)

Resilient Procedures: Oxymoron or Innovation?

Posted by Robert L Wears on Tuesday January 6th, 2015

 Resilient Procedures – Oxymoron or Innovation?

Robert L Wears, MD, MS, PhD

Discussions about resilience often fetch up on the shoal of procedures[1] – prescriptive, feed-forward guidance. This is unfortunate because, resilience is about a great deal more than procedures. In addition, the issue of procedures is often contested and acts as a ‘black hole’ sucking in all our attention and energy, while leaving little for other important areas of concern. This essay will discuss procedures in four parts: what baggage tends to accompany procedures; what is bad about procedures; what is good about them; and how procedures might be designed to support or even enhance resilience, instead of degrading it.

Contested nature of procedures. Procedures often come with excess baggage; issues that strictly speaking have little to do with the procedures themselves, but are social and political. Often, these ‘extraneous’ issues are dispositive, major contributors to the success or failure of attempts at proceduralisation.

Procedures are often the field upon which struggles for control play out. The popular 1969 movie Butch Cassidy and the Sundance Kid opens with a contest between Butch and Harvey for the leadership of the Hole-in-the-Wall gang, but the struggle doesn’t actually begin until the protagonists start to argue over the rules to be applied to their knife fight. Less fancifully, sociologists studying evidence-based medicine (a proceduralizing movement in health care) have characterized it as “… a move in the chess game of countervailing powers vying for dominance in the healthcare marketplace, the professional answer to government agencies’ and insurance companies’ demands to render medicine more efficient and cost-effective” (Timmermans, 2010).

In addition, procedures may be produced for reasons other than providing guidance. They can be performances for external audiences, demonstrations that system leaders have adequate control of any hazards, and so used to counter demands for external oversight. This commonly leads to idealized procedures that cannot possibly be followed. Clarke has called such procedures ‘fantasy documents’ (Clarke, 1999).

Procedures can also be useful foils for shifting responsibility for adverse events from the powerful (e.g., managers) to the less powerful (e.g., front line workers). “Failure to follow procedure” is often noted as a cause of accidents, while “poorly designed procedures” is almost never cited.

Finally, some of the appeal of procedures has been criticized as the technocratic wish, a reincarnation of scientism, “… modernity’s rationalist dream that science can produce the knowledge required to emancipate us from scarcity, ignorance, and error” (Goldenberg, 2006), an objectivist presumption that the world and its technologies are ultimately knowable, and that epistemic ambiguity and uncertainty can be swept away (Kenny, 2014).

Problems with procedures. Social and political issues aside, there are more intrinsic problems with procedures. Procedures can be seen as a specific case of standardization (Wears, 2014), and so they inherit all its associated advantages and disadvantages.

Procedures are necessarily decontextualized; abstractions of what generally is done, but not necessarily what specifically should be done in every situation. Complex organisations always run in mildly degraded states – someone is always absent, some device is inevitably out of service, some supplies are missing; workers normally must adapt procedures that presume ideal conditions to the actual conditions they face. If this context sensitivity is important to success, then slavish adherence to decontextualized procedures would be risky.

Attempts to compensate for this problem are ultimately self-defeating, because procedures’ abstraction is part of their utility. A map is useful because of what it leaves out as much as by what is shows; a procedure covering everything would not be able to accomplish anything. Indeed, a strong critique has developed arguing that some industries are over-proceduralized (Bieder & Bourier, 2013). A proliferation of procedures in ever-greater detail becomes both internally contradictory and impossible to grasp.

This problem points out an important precondition for procedures to be successful. Because they are decontextualized and abstracted, they depend heavily on the ‘learned intuition’ of the workers using them; the experiential knowledge of how things work and what is intended in context. Because all of us know how to open a jar, a procedure for making a peanut butter sandwich can elide the multiple steps involved in opening a jar. But this common grounding is seldom explicitly acknowledged or even recognized by procedure writers or users. (For example, the first step in reaching forward to press an elevator button while standing has nothing to do with the arms; it is a tensioning of the gastrocnemius muscles of the legs to counteract the coming imbalance caused by moving an arm (roughly 8 pounds) forward of the body’s center of gravity).

Although it could theoretically be specifically included as an element of procedures (as noted below), many procedures in practice tend to deny the existence of tradeoffs among goals, placing workers in a classic double bind based on the outcome of events, not on the adequacy of their judgment in negotiating tradeoffs. Any tradeoff conflicts that procedure writers are not able to solve are pushed onto the front line worker.

Of course in open systems, there will always be examples of situations in which following procedures led to disaster that deviation might have avoided. Vicente reports a simulation exercise in a nuclear power plant where the operators realized that the formal procedures for addressing a fault would eventually put the plant back into the original fault condition; so they deviated from the procedures in order to stabilize the plant safely. They were instructed that this was wrong, and they should instead adhere to procedures. When they did this in the next simulation, the plant was caught in an endless loop (i.e., the ostensibly corrective procedures returned it to its original fault condition); the operators were then cited for “malicious compliance” (Vicente, 1999).

Benefits of procedures. Procedures are generally viewed by their critics as intrinsically restraining, but there is no reason they need to be. The musical directives in Figure 1 illustrate this point. The musician is directed by the ‘ff’ to play very loud, by the ‘dim.’ (for diminuendo) to gradually get softer, but also by the ‘ad libitum’ to take liberties with the score – playing the notes exactly as written would be to play them incorrectly, because here, the composer wants performance variability.

 

 Figure 1. Procedural directives in music

Procedures have some intrinsic advantages. At their best, they can be viewed as distillations of experience, expertise, and vernacular knowledge accumulated over many years by many workers in many situations. This can be particularly useful in seldom-encountered situations, particularly those situations in which the best action may be counter-intuitive.

Good procedures can serve as resources for situated actions. Even in off-design-base situations, procedures can provide ‘a place to start’; a way to make sense of an unfolding situation by taking actions and observing the responses. By automating, in a sense, some bits of work, they allow workers to direct scarce energy and attentional resources to its more important or more demanding aspects. In this sense they are particularly valuable as guides for novices.

When procedures are well-adapted to context, well-understood and generally observed, they can promote coordination at a distance and reducing the burden of articulation work, by providing multiple actors an understanding of what their counterparts are likely to be doing. Similarly, by providing common ground, they promote efficiency in communication.

Schulman’s study of the Diablo Canyon nuclear power plant provided an example of the effective use of a high degree of proceduralization in promoting safe and effective operations (Schulman, 1993). Interestingly the high degree of procedural formality was not accompanied by centralization of authority, but rather by the reverse. By requiring a myriad of technical subunits to have formalized procedures, and by giving each unit effective veto power over others’ procedures, the plant increased the variety of viewpoints engaged in producing procedures, decreased structural secrecy, and diminished the gap between work as imagined and work as done. By requiring frequent re-writing of procedures (often in one unit occasioned by another unit’s procedural change) they were able to keep pace with technological and environmental change.

Better Procedures as a Substrate for Resilience.

Given this potential for procedures to be either good or bad, how might we craft better procedures that would support or even enhance, rather than degrade, resilience? This will require attention to two aspects of procedures: process and content.

Process issues involve promoting conditions that provide the stage for, but do not necessarily ensure, better procedures. Increasing the variety of skills and viewpoints involved in procedure development, particularly if accompanied by flattening the organisation’s hierarchy and shifting towards distributed, polycentric control.

Careful monitoring of the gap between work-as-imagined in the procedures and work-as-done “at the coal face”, coupled with a stance towards understanding an increasing gap as a sign that revision, not enforcement, is required, can help keep procedures practicable and abreast of changes in the work and the environment. This requires a commitment, not just lip service, to the Principle of Local Rationality: that people do not do things that do not make sense to them – thus if they are not following the procedures, they must not make sense to them, given their locally understood resources, constraints, and goals.

On the content side, it would be useful to attend more to the principle of equifinality: in complex systems, there are often many paths to the same goal. This is the anti-thesis of the Taylorist belief that there must always be “one best way” to manage a work situation. Particularly under conditions of uncertainty, resource scarcity, or time pressure, organisations cannot be certain that any single path will be available to them, so they would do well to maintain multiple possible paths in a sort of ‘portfolio strategy’ to hedge against uncertainty.

Hale and Swuste have usefully advanced thinking about how to craft procedures wisely in their conception of goal, process, and action rules (Hale & Swuste, 1998). Goal-oriented procedures identify the end state desired, but leave the selection of means unspecified. They provide useful guidance to resolving goal conflicts, particularly in intractable systems (Hollnagel, 2014) where the object is not invariant performance, but rather the successful management of fluctuations. A simple example in healthcare comes from the principle when treating the two patients (mother and fetus) in a pregnancy, that “the mother’s life comes first.” This is not to say that mothers should never be exposed to risk to benefit the baby, but that when the two come directly into conflict, the best way to save the baby is to save the mother first.

Process-oriented procedures do not specify actions, but rather how actions should be chosen; who should do the choosing, and what factors should decision makers consider (and conversely, which should they ignore). These procedures are useful in situations that still require flexibility in response, or as a means of coordination in polycentric control.

The most detailed and specific procedures are action procedures, which list in detail, which specific actions should be taken, by whom, in what order. Action procedures are typically what comes to mind when we think of procedures; Hale and Swuste’s contribution is significant in broadening our conception of what constitutes a procedure and thereby increasing procedure’s utility. Clearly, action rules require both tractable systems (Hollnagel, 2014) and stable environments to function properly.

Gudela Grote has noted that in high risk systems, the most appropriate rules are not those that are most restrictive and therefore uncertainty reducing, but rather are flexible rules that promote adaptive action by providing more degrees of freedom to decision makers thus increasing, rather than decreasing, their uncertainty (Grote, 2014).

There are some examples of flexible procedures in practice. One of the best is the German army’s Auftragssystem from Franco-Prussian War through World War II, which emphasized the primacy of the mission (goals) over orders (procedures) (Finkel, 2011). “Far from being a horde of rigid and inflexible robots (… one of the most insidious of the Allied misconceptions about the Germans), the Wehrmacht … had a far greater ability to react or regain the initiative – especially in a moment of reverse – than was possessed particularly by the British Army of 1944” (Horne & Montgomery, 1994). This system provided explicitly that a subordinate commander should be able to pursue the tactical goals of his superiors with or without orders, based on his knowledge of the local context (Reason, 1997). This of course presupposes a considerable degree of initiative and tactical understanding on the part of these junior leaders, which points out that supportive procedures are only one component of a resilient system; another is developing the capability for responsive adaptation in front line leaders. The success of these flexible procedures showed on the battlefield; whether attacking or defending, the German army inflicted substantially more causalities than it suffered in these conflicts.

  1. A good deal of the literature about resilience has concerned itself with phenomenology -- detailed, thick descriptions of specific manifestations of resilient or brittle performance in demanding circumstances. However, this focus has been limiting because it fails to address the question of where those resilient performances come from; what substrate supports them. There must be something about a system that provides the foundation for resilience when called upon (Fairbanks, Wears, Woods, Hollnagel et al., 2014). Much of a system’s repertoire of responses is hidden in the form of latent behavioral potential, out of awareness and taken for granted until some triggering event calls attention to it (Christianson, Farkas, Sutcliffe, & Weick, 2009). If we want to understand how to engineer resilience into our systems we must find ways of identifying, conserving, and developing that potential.

One way to do that is to learn how to design effective, flexible procedures that support resilience rather than interfere with it. This may require organisations to embrace uncertainty rather than to pursue the will-of-the-wisp of rationalist certainty.

References and further readings

Bieder, C., & Bourier, M. (Eds.). (2013). Trapping Safety into Rules: How Desireable or Avoidable is Proceduralization. Farnham, UK: Ashgate.

Christianson, M. K., Farkas, M. T., Sutcliffe, K. M., & Weick, K. E. (2009). Learning through rare events: significant interruptions at the Baltimore & Ohio Railroad Museum. Organization Science, 20(5), 846 - 860.

Clarke, L. (1999). Mission Improbable: Using Fantasy Documents to Tame Disaster. Chicago, IL: University of Chicago Press.

Fairbanks, R. J., Wears, R. L., Woods, D. D., Hollnagel, E., Plsek, P. E., & Cook, R. I. (2014). Resilience and resilience engineering in healthcare. Joint Commission Journal on Quality and Patient Safety, 40(8), 376 - 383.

Finkel, M. (2011). On Flexibility: Recovery from Technological and Doctrinal Surprise on the Battlefield. Stanford, CA: Stanford University Press.

Goldenberg, M. J. (2006). On evidence and evidence-based medicine: lessons from the philosophy of science. Social Science & Medicine, 62(11), 2621-2632. doi: 10.1016/j.socscimed.2005.11.031

Grote, G. (2014). Promoting safety by increasing uncertainty – Implications for risk management. Safety Science, x(xx), (on line ahead of print). doi: http://dx.doi.org/10.1016/j.ssci.2014.02.010

Hale, A. R., & Swuste, P. (1998). Safety rules: procedural freedom or action constraint? Safety Science, 29(3), 163-177.

Hollnagel, E. (2014). Safety-I and Safety-II: The Past and Future of Safety Management. Farnham, UK: Ashgate.

Horne, A., & Montgomery, D. (1994). The Lonely Leader: Monty 1944-45. London, UK: Macmillan.

Kenny, K. E. (2014). Blaming Deadmen: Causes, Culprits, and Chaos in Accounting for Technological Accidents. Science, Technology & Human Values. doi: 10.1177/0162243914559288

Reason, J. (1997). The human contribution Managing the Risks of Organizational Accidents (pp. 61 - 83). Aldershot, UK: Ashgate Publishing Co.

Schulman, P. R. (1993). The negotiated order of organizational reliability. Administration and Society, 25(3), 353 - 372.

Timmermans, S. (2010). Evidence-based medicine: sociological explorations. In C. E. Bird, P. Conrad & A. M. Fremont (Eds.), Handbook of Medical Sociology (pp. 309-323). Nashville, TN: Vanderbilt University Press.

Vicente, K. J. (1999). Cognitive Work Analysis: Towards Safe, Productive, and Healthy Computer-Based Work. Mahwah, NJ: Lawrence Erlbaum Associates.

Wears, R. L. (2014). Standardisation and its discontents. Cognition, Technology & Work, (online ahead of print). doi: 10.1007/s10111-014-0299-6

 

[1] In this essay, I will use the word ‘procedures’ to refer to a variety of types of feed-forward guidance: rules, regulations, guidelines, checklists, directives, doctrine, etc. While distinctions among these exist, they are not relevant to this discussion.

Add your comment

January 22nd, 2015 by Gary Wong
How might we craft better procedures that would support or even enhance, rather than degrade, resilience?
As Bob noted, one way is to learn how to design effective, flexible procedures. I’ll call that the work-as-imagined route. Another possibility is the work-as-done route.

Since Healthcare is a complex adaptive system, there will always be unexpected surprises to deal with. Safety-II recognizes the varying conditions workers face and the adjustments required to do things right. Effective flexible procedures really need constant updating to document the steps someone carried out in a successful work-as-done. But let’s be honest; it ain’t gonna get done. One, there’s no time. Two, no one would read it anyway.

So let’s focus on the pragmatic side and I'll use the two aspects of procedures, process and content, to explain the work-as-done route.

The process is an AAR (after action report) when a work-as-done event has completed. But we can also report to prevent an incident from occurring (triggering anticipatory awareness.) The worker uses a desktop computer or better, a mobile smart device (iPhone, Android) to enter work-as-done content 24/7/365 into a searchable knowledge management database. We call this a Human Sensor Network where technology augments human intelligence, and not the other way around.

The content is in the form of stories. We are homo narrans and storytelling comes naturally. We were not born to be document writers. Stories are great for complex situations as they allow context to be integrated. Stories also act as ‘strange attractors’ and can help to reveal hidden behaviour patterns. Stories can be recorded as text, voice recordings, and even be photos. We can now convert stories into data points and therefore stories can be indexed, analyzed and visually graphed. See a typical bar chart but when you click on a bar, you can drill down and read the actual stories.

Where is this happening? In Dave Snowden's 2013 State of the Net address he talks about one project in a US medical college (cancer hospital) where patients use an app to keep the continuous diary of their experiences. The primary intent was to understand
the impact of primary care, secondary care, and community care on a minute by minute basis. It also provided early detection of a pattern indicating someone was being discharged too early. The unexpected benefit was the knowledge management database. Doctors and nurses found it more valuable than outdated best practice documents.

Dave's address is on YouTube http://youtu.be/gaFQJaQTXhY
The cancer hospital story begins at 32:04. However, I suggest starting at 29:24 to get the a better concept of the Human Sensor Network.

Submit comment

Dr Wears is an emergency physician, Professor of Emergency Medicine at the University of Florida, and Visiting Professor in the Clinical Safety Research Unit at Imperial College London.  His further training includes a Master’s in computer science, a 1 year research sabbatical focused on psychology and human factors in safety Imperial, after and a PhD in resilience engineering from Mines ParisTech (Ecole Nationale Supérieure des Mines de Paris).  He serves on the board of directors of the Emergency Medicine Patient Safety Foundation, and multiple editorial boards, including Annals of Emergency Medicine, Human Factors and Ergonomics, the Journal of Patient Safety, and the International Journal of Risk and Safety in Medicine.  He has co-edited two books, Patient Safety in Emergency Medicine, and Resilient Health Care.  His research interests include technical work studies, resilience engineering, and patient safety as a social movement. 

 

Dr Wears is an emergency physician, Professor of Emergency Medicine at the University of Florida, and Visiting Professor in the Clinical Safety Research Unit at Imperial College London.  His further training includes a Master’s in computer science, a 1 year research sabbatical focused on psychology and human factors in safety Imperial, after and a PhD in resilience engineering from Mines ParisTech (Ecole Nationale Supérieure des Mines de Paris).  He serves on the board of directors of the Emergency Medicine Patient Safety Foundation, and multiple editorial boards, including Annals of Emergency Medicine, Human Factors and Ergonomics, the Journal of Patient Safety, and the International Journal of Risk and Safety in Medicine.  He has co-edited two books, Patient Safety in Emergency Medicine, and Resilient Health Care.  His research interests include technical work studies, resilience engineering, and patient safety as a social movement. 

 

Events:More...


News:More...

6th Resilient Health Care Meeting

For details about the 6th Resilient Health Care Meeting being hosted in Vanco

More