Taking Things in One’s Stride: Cognitive Features of Two Resilient Performances
“If you can keep your head when all about you
Are losing theirs and blaming it on you…”
Rudyard Kipling, 1909
Resilience is a feature of some systems that allows them to respond to sudden, unanticipated demands for performance and then to return to their normal operating condition quickly and with a minimum decrement in their performance.
Our approach to resilience comes from the perspective of practitioner researchers, that is, as people engaged directly with sharp-end work and sharp end workers. Regarding human performance in complex systems, we take two things for granted:
• The many-to-many mapping that characterizes a system’s goal – means hierarchy relates the possible to the necessary (Rasmussen, Pejtersen & Goodstein, 1994).
• Human operator performance in these systems is encouraged, constrained, guided, frustrated, and enacted within this hierarchy (Rasmussen, 1986).
The location of the system’s operating point is the result of a balance of forces (Cook & Rasmussen, 2005) and tracing the movement of that point over time maps the history of that system’s safety.
This chapter describes some of the cognitive features of resilience. We seek to prepare the ground for discussions of resilience and its characteristics using two examples. While the following accounts of two events are both drawn from the domain of healthcare, they have few overlapping features. One occurs over a very short period of time, is a highly technical problem, and has a few actors. The other spans a half a day, encompasses a large socio-technical system, and involves hundreds of people. Despite their differences, the two events demonstrate resilience. To be more precise, each is a resilient performance.
Example 1: Handling a ‘Soft’ Emergency
Setting. The Department of Anesthesia (DA) in a major urban teaching hospital is responsible for management of the Surgical and Critical Care unit (SACC). The unit includes six activities in the hospital complex that require anesthesia services. Cases are scheduled for the outpatient clinic (SurgiCenter), the inpatient operating rooms (IOR) or four other services, such as radiology. The most acute cases are treated in the IOR. Cases vary from cardiology to neurology to orthopedics, transplants and general surgery. In each case, staff, equipment, facilities, and supplies must be matched to meet patient needs and surgeon requirements.
The DA employs a master schedule to funnel cases into a manageable set of arrangements. Completed each day by an anesthesia coordinator (AC), the schedule lists each case along with all needed information. The finished master schedule, which is distributed throughout the organization at 15:30 each day, is the department’s best match between demand and available resources. One hard copy is posted at the coordinator station that is located at the entryway of a 16-operating room SACC unit. There, the AC and Nurse Coordinator (NC) monitor the status and progress of activity on the unit as the day progresses. Surgeons reserve OR rooms by phone, fax, e-mail, and inperson at the coordinator station. This flow of bookings is continuous and varies according to patient needs and surgeon demands. Because the flow of demand for care is on-going, surgeons continue to add cases through the day. These are hand written onto the master schedule ‘add-on list.’
Cases are either scheduled procedures or emergencies. Each surgeon estimates case severity as either routine, urgent or emergency. If a room is not available, a case is pre-empted (‘bumped’) to make resources available. While the bumping arrangement works, the staffs prefer not to use it if possible. That is because bumping causes significant changes to the way staff, rooms and patients are assigned, which the unit would prefer to avoid. Officially, the potential loss of a patient’s life or vital organ would be an emergency; and its declaration requires that resources be freed up and made available for that case. In reality, there are very fine gradations of what constitutes an emergency. Because of that ambiguity, it is possible for a surgeon to declare a case ‘emergent’ in order to free up resources for his or her case. In the instance of cases that are declared emergent, it is up to the AC to evaluate the case. Anesthesia and nursing staff members use the term ‘soft emergency’ to those cases that are declared yet have no apparent high risk. It is up to the AC to ‘read’ each declaration to ensure that there is an equitable match between demand and resource.
It is Friday at mid-day and surgeons are attempting to clear out cases before the weekend. The master schedule add-on list is 2.5 pages long. One page is more routine. The chief anesthesiology resident speculates that there may be a backlog because admitting was suspended on Thursday due to no available beds (the queue ‘went solid’). At around 11:30 the Anesthesiology Coordinator fields a phone call to the AC. The following sequence of events occurs over about ten minutes from 11:30 to 11:40. Portions of the dialog (Nemeth, 2003) that occur at the Coordinator station provide a sense of what is occurring.
Observations. Operating room number 4 is the day’s designated bump room. The sequence begins as the AC explains to Surgeon 2, whose next case is scheduled in the bump room, that an emergency declared by Surgeon 1 is bumping him and, therefore, that his next case may not be done today.
AC: [on phone with Surgeon 2] “I don’t think we can do it today. We can book it and see if we can fit it in.” [pause]
“You’re in the bump room, number 1 [in line to be bumped].” [pause]
“He [Surgeon 1] wants to do it.” [pause]
“I don’t make those decisions.” [pause]
“You’ll have to talk to him [Surgeon 1].”
Nurse1 [from Room number 4] “Are we being bumped?”
AC: “Yes, you’re being bumped.”
Nurse: “How do they decide which rooms get bumped?”
AC: “The bump list.”
Nurse: “Doesn’t it make sense to bump by the severity of the case?”
AC: “It depends on the surgeons to make those decisions.”
AC: [to NC] “Clean it up and let’s set it up. Ask the person bumping to talk to the surgeon who is being bumped.”
--break--
NC: “[O.R. number] Six is coming out.”
AC: [to NC] “We can put [Surgeon 1] in there.”
[Calls on the phone and waits, no answer.]
[to NC] “We can put [Surgeon 2] in [OR number] 4 and [Surgeon 1] in [OR number] 6.”
“Let me page [Surgeon 1].”
--break--
[to NC] “You’re planning to put something in [OR number] 6.”
[points to a case on the add-on list]
[pages Surgeon 1]
[calls Surgeon 2]
--break--
Nurse1: [while walking past station] “Are we being bumped? Is this for sure now?”
AC: [to Nurse1] “Do the scheduled case.”
--break--
AC: [to Surgeon 1] “Surgeon 1? Room 6 is coming out.” [pause]
--break—
AC: [to Surgeon 2] “Surgeon 2? I just talked to Surgeon 1 and he doesn’t mind waiting 45 minutes [for OR number 6 to be setup and the emergency case to start].”
The surgeon had called to declare an emergency in order to have his skin graft case (an add-on procedure) assigned to an operating room when none were open. Opening a room would require a scheduled case to be bumped. The AC negotiated an arrangement in which the surgeon accepted a 45-minute delay until a room became available. The arrangement avoided bumping a procedure. The AC reflected on the process shortly afterwards:
When they declare emergency I don’t have anything to do with it. We let them work it out. I struck a deal by getting [Surgeon 1] to wait 45 minutes and was able to put him into a room that was coming open. That avoided having to bump a procedure.
This particular AC has a reputation for many capabilities, notably for his expertise in coping effectively with demands for resources. In this case, he confronts a situation that formally requires a particular course of action – the use of the bump room. Although he is prepared to follow this disruptive course, he does more than the formal system requires. He manages the problem expectantly, discovers an opportunity, and presents it to the parties for consideration, all the while reminding all the parties of the formal requirements of the situation. He manages to avoid confrontation by avoiding a controlling role. Instead, he invites surgeons to work out the resource conflict between them. By offering the bumping surgeon a viable alternative and inviting him to accept it, he relieves the tension between resource and demand. He averts a confrontation, making good use of available resources. He understands the surgeons’ personalities, the nature of the kinds of cases at issue, the pace of work in the operating room, and the types of approaches that will be viable. He evaluated the skin graft as a ‘soft emergency’, that is, a case that could be delayed but not put off indefinitely.
The evolution of responses that the AC gave through the brief period indicates how elastic the boundaries are between what exists and what might exist. Circumstances changed from a demand for unavailable resources, to the potential for an arrangement, to the reality of an arrangement. As he saw that evolving, he adjusted his responses in order to set actions in motion so that resources were ready for use when needed (e.g., “Clean it up and let’s set it up,” and “Do the scheduled case”).
The episode is subtle, small, and embedded in the ordinary flow of work in ways that make it almost invisible to outsiders. The AC employed complex social interactions and various kinds of knowledge to reorder priorities. The event’s significance lies in the absence of disruption that the AC’s efforts made possible in the face of a sudden demand for resources.
Example 2: Response to a Bus Bombing
Setting. At 07:10 on Thursday morning, 21 November, a suicide bombing on #20 Egged bus occurred while the bus was in the Kiryat Menahem neighborhood of Jerusalem. The bomb was detonated by a 22 year old Palestinian from Bethlehem who had just boarded the bus and was standing near the middle of the bus at the time of detonation. The explosive device included a variety of shrapnel objects; typically these are ball-bearings, small bolts or screws, or nails. Early reports from the scene said that seven people had been killed. Two persons evacuated from the scene died in hospital on the first day. Forty-eight other people were wounded; twelve people had died of their wounds by the end of the day (eleven plus the bomber).
Bus attacks by suicide bombers have fairly monotonous features. They occur during the morning rush hour because ridership is high at that time. Bombers board busses near the end of their routes in order to maximize the number of people in the bus at the time of detonation. They preferentially board at the middle doors in order to be centered in the midst of the passengers. They detonate shortly after boarding the bus because of concern that they will be discovered, restrained, and prevented from detonating. They stand as they detonate in order to provide a direct, injurious path for shrapnel. Head and chest injuries are common among seated passengers. The injured are usually those some distance away from the bomber; those nearby are killed outright, those at the ends of the bus may escape with minor injuries. The primary mechanism of injury of those not killed outright by the blast is impaling by shrapnel. Shrapnel is sometimes soaked in poison, e.g. organophosphate crop insecticides, to increase lethality.
Traveling to the Hospital. A senior Israeli physician (DY) heard the detonation of the bomb at 07:10 and drove with one of the authors (RC) to the Hadassah Ein Karem hospital. It is routine for physicians and nurses to go to their respective hospitals whenever it is clear that an attack has taken place. Within 15 minutes of the detonation, FM radio broadcasts included the bus route name and location at which the explosion had occurred and provided early estimates of casualties. DY did not use his cell phone to call the hospital for information, explaining that there was no reason to do so, that he could not participate effectively by telephone, and that the senior staff person at the hospital would already be managing the response. DY explained that this was a ‘small’ attack and that the fact that the operating room day had not commenced at the time of the detonation meant that there were ample staff and facilities to handle whatever casualties would appear.
Ambulance drivers are trained to go to the site of a mass casualty without direction from the dispatch center. The dispatch center triages ambulance hospital assignments once they are on scene and loaded with patients. Emergency personnel describe this approach as ‘grab and go.’ This leads to rapid delivery of patients to hospitals where definitive care may be obtained. It decompresses the scene so that police forensic and recovery operations can take place unhindered by the presence of injured people. It also reduces the risk of further injuries from secondary attacks.
At the Hospital. As the largest modern hospital in the Mid-East, Haddassah Ein Karem could be expected to be busy and crowded in the morning of a regular workday but it was especially so at this time. Although the detonation had been less than an hour before, people were already arriving to search for family members. The hospital had already dedicated a large meeting room for the families and friends, and hand-lettered signs were posted indicating the location of this room. In the room were trained social workers and physicians. These personnel had donned colored plastic vests with English and Hebrew print indicating that they were social workers. The room itself was supplied with tables and chairs, a computer terminal connected to the internet, and water and juice. This room had been selected, provisioned, and manned within 30 minutes of the detonation and in advance of the first families arriving at the hospital.
Staffing for the social work activities is flexible and it is possible to recruit additional support personnel from the hospital clerical and administrative staff. The goal of the social work staffing is to provide sufficient numbers of staff that there is one staff person for each individual or family group. Once attached to the individual or family, the social worker focuses narrowly on him or them for the duration of the emergency.
To the Operating Rooms. DY went immediately to the operating room desk for a brief conversation with the four or five anesthesiologists and nurses present there. He then manned the telephone at that location. He checked an internet terminal on the desk for the latest information from the scene and took a telephone call regarding a casualty being transferred in from another hospital. The apparent death toll was seven or eight and there were thought to be forty to fifty wounded. The wounded had already been dispersed to hospitals in the city. This triage of cases is managed by the ambulance drivers and dispatchers in order to distribute the casualties efficiently. In particular, the triage is intended to move the most seriously injured to the main trauma centers while diverting the minor casualties to smaller hospitals. Coordination to free up intensive care unit beds for the incoming casualties had begun.
Two severe casualties had been taken directly to the operating room, bypassing the trauma room. Operating rooms are used as trauma rooms for expediency: cases that will obviously end in the operating room bypass the trauma room and are taken by a team directly to the OR. This speeds the definitive care of these patients and preserves the three trauma room spaces for other casualties; observed activities were considerably faster and involved fewer people than in many US trauma centers.
To the Emergency Room. The emergency room can be reached by driving into a square that is approximately 80 meters across and surrounded on three sides by hospital buildings. Ambulances enter along one side of the square, discharge their patients, and proceed around the square and exit the area. Security personnel had erected a set of bicycle rack barricades to create a lane for the ambulances. This was necessary because the television and newspaper reporters and camera crews were present in force and crowded into the area that is normally reserved for the ambulances.
A hospital spokeswoman was present and patiently repeated the minimal information that was available. Members of the press were intrusive and persistent and were the most visibly keyed-up people observed during the day. The hospital spokeswoman was remarkably patient and quite matter-of-fact in her delivery. The hospital’s director of media relations was standing off to the side, uninvolved in continuous press briefing but monitoring the activities.
Arriving families were directed to the emergency room where they were met by social workers. Triage of casualties to multiple hospitals and the transfer of the dead to the morgue happens before identification. In many instances, casualties and the dead are not readily identifiable by medical personnel. This is particularly true for those who are dead or severely injured because of the effects of blast and fire.
Following a bombing, people begin by attempting contact bus riders by cell phone. If there is no answer or if the rider does not have a cell phone, family members go to the nearest hospital. Social workers at that hospital will gather information about the person being sought. All such information is shared via the internet with the all the other hospitals. Similarly, hospitals collect identifying information about the patients they have received, including gender, approximate age, habitus, and clothing. These two sets of information are then collated by social workers at each hospital. This helps the medical staff identify who is being treated and helps the families determine the location of the person they are seeking. Pooling the information makes it possible quickly to refine a casualty list and identify individuals. Police use this information from the hospital at the morgue formally to identify those who have died. In this case, a complete casualty list was available within about 6 hours of the bombing. The next morning’s newspapers provided names and pictures of all the dead including the suicide bomber.
The Israelis place a high premium on providing family members with access to casualties. The author observed several sets of parents and siblings being brought into the preoperative holding area or the emergency room to be with the casualty. In each case, the family members were accompanied by at least one social worker.
For each person who is injured, as many as four people may be expected to come to the hospital. Parents, grandparents, siblings, and children may arrive simultaneously or in waves. The management of this crowd of people, who are all under intense psychological stress, is a major undertaking that requires substantial resources.
Care of a Severe Head Injury. An ambulance delivered a young female patient with a serious head injury to the Haddassah Ein Karem emergency receiving area. This patient had initially been taken to another hospital and was now transferred to this center where neurosurgeons were located. While the ambulance attendant provided a detailed description of the nature of the suspected injuries, medical staff took the patient to the trauma room performed resuscitation, and took physiological vital signs. The patient was found to have multiple, minor shrapnel wounds but no apparent major trauma other than a penetrating head injury. She had already been intubated and was being mechanically ventilated. She was in a cervical collar and secured on a transport backboard. All the elements of the Advanced Trauma Life Support (ATLS) protocol were observed: a complete examination was made, the cervical spine and chest were x-rayed and the C-spine cleared, and blood samples were sent for evaluation. Verbal exchanges in the trauma room were exclusively technical and the tenor of the conversation was relaxed. A senior radiology physician performed an abdominal ultrasound examination. The blood gases were returned within five minutes and slight adjustments to mechanical ventilation were made. The patient’s vital signs were stable and the patient was mildly hypertensive. A negative pregnancy test result was returned at almost the same time. Narcotics and neuromuscular blocking drugs were administered.
Much of the paperwork that is completed during routine operations is deferred or abandoned entirely during a mass casualty response. A notable exception is the blood type-and-cross match paperwork, which is conducted using the routine forms and procedures. DY explained that abandoning some documentation was necessary to allow the critical documentation to be completed. Successful blood transfusion depends on accurate documentation.
A computerized tomography (CT) scan was planned and the patient was taken on a gurney to the CT scanner, while connected to a pulse oximeter, electrocardiograph (ECG), and non-invasive blood pressure monitor. The senior anesthesiologist, an anesthesiology resident, and a neurosurgeon accompanied her as a staff security person cleared a path through the hospital corridors. When the team and patient arrived in the CT room, they used a small, compressed gas powered ventilator that is permanently installed in the CT scan room, to continue mechanical ventilation. As the CT scan was made, the chief neurosurgeon arrived and, together with a radiologist, read images at the console as they were processed.
The decision was made to perform a craniotomy, which is surgery that involves opening the skull, and the team transported the patient to the operating room. The neurosurgeon received and answered a cellular telephone call about another patient as he pulled the gurney. While waiting for an elevator, the patient’s parents came to the gurney, accompanied by a social worker. One of the transporting physicians spoke with them about the known injuries and the expected surgical procedure. The parents touched the patient, and looked under the blankets at her legs and feet.
The attending anesthesiologist, a scrub nurse, and a circulating nurse were waiting in the OR for the surgical team and patient to arrive. The anesthesiology resident provided an update on the patient’s condition. After transfer to the operating room table, the patient was positioned, a radial arterial line was started and the head prepared for a craniotomy. The craniotomy began one hour after the detonation.
Return to Normal. Although several operating rooms were occupied with procedures that were related to the bombing, the remainder of the operating rooms had begun cases from the regular daily list. Efforts to resuscitate a trauma victim continued in one room. Several other patients with less serious injuries were in the preoperative holding area. No further direct casualties were received.
The author met briefly with anesthesiologists, surgeons, and nurses in the operating room corridor and then visited the emergency room and the family room. Many of the families had dispersed to other areas in the hospital as the patients received treatment, were admitted to wards, or were being discharged.
DY noted that the team generally holds a brief informal discussion of lessons learned immediately after a mass casualty response. The value of these discussions, he said, has fallen off substantially as the handling of such events has become routine.
We propose the cases as examples of resilient performances and these resilient performances as evidence of the presence of a resilient system. Systems may be resilient but we do not know how to detect resilience short of the direct observation. This is because resilient performances occur in the face of sudden, unanticipated demands. At present, the only strong evidence of resilience that we can identify is the presence of these resilient performances. They are made possible by the configuration of the workplace, incentives, and social structures that surround them. This does not lessen our regard for the critical role that is played by the participants’ attention, knowledge, assessments, and decisions. Confronted with a sudden demand, we observed the practitioners respond purposefully and appropriately. It was clear from their actions and speech that they recognized the demand and its significance. Their reaction responded directly to the demand. Their reaction was also temperate and tempered by other considerations. These resilient performances depend heavily, but not exclusively, on particular individuals and on the particular expert cognitive abilities that they bring into play. Through their experience, intentions, and judgement as events evolve, practitioners invest cognitive effort to maintain successful performance in the face of unexpected challenges. Cognition creates what we observe as resilient performance.
Unusual but Not Unknown: The Roles of Deep Domain Knowledge
Events that provoke resilient performances are unusual but not entirely novel. The practitioner in the soft emergency case knows that an ‘emergency’ can have different degrees of criticality. The experience of being called on to respond to a sudden demand is common enough that the practitioner has developed expertise in coping with such demands. Similarly, in the bombing case, the practitioners know from the earliest description of the event what sorts of wounds to anticipate, how patients will arrive, and where the bottlenecks in response are likely to be. They also know what resources they can call on and the established processes of dealing with this sort of event. In both examples, the practitioners use their knowledge of how operational tempo plays out, and what pathways can be exploited to respond.
Return to Normal Operations
Resilient performances include a response to a sudden demand and the return to ‘normal’ operations following that response. Concern for routine operations shapes the planning and execution of the response to the sudden demand. In the ‘soft emergency’ case, the response is fashioned in order to have a minimal effect on routine operations. This is a very practical concern. Routine system functions are important and expensive resources such as the OR are usually operated near, or at, full workload.
In the ‘soft emergency’, the anesthesia coordinator is able to foresee an opportunity to meet the new demand within the stream of activity. He knows that emergencies vary in criticality. Some require immediate response but others can tolerate a delay. He queries the surgeon to see whether this case can be delayed. He also examines current operations in order to see whether an operating room will, or could, become available.
In the bus bombing response, the response to sudden demand is continually assessed and normal work resumed even as the casualties from the bombing are being cared for. The resilient system is an inherently valuable resource so that interrupting its normal function generates costs and it is important to resume that normal function as smoothly and quickly as possible.
The concern for routine operations in the midst of sudden demand distinguishes resilience from other types of reactions. In both of these cases the system responds to a need without entirely abandoning routine operations. Normal operations continue in the face of the sudden demand. The response to the demand includes efforts both to gauge its effect on normal operations and to devise the response so as to disrupt those operations as little as possible.
Defining Successful Performance
Performances might be resilient without being successful, at least in the conventional sense. There are ways that these cases could have turned out better. If there were no delay in treating the burn patient, if the head injured patient had been taken directly to the main hospital instead of an outlying one, it is possible that these patients would have done better. Even so, the activities of these practitioners qualify as resilient performances. What, then, is the relationship between resilience and success? Can performances be resilient and unsuccessful? Are successful outcomes evidence of resilience?
Some sudden demands, such as large scale natural disasters, might be so overwhelming that any plan of action is destined to fail. In such situations, just returning to normal in the wake of an event could count as a resilient performance. Indeed, some of the more compelling examples of resilience are those in which the people and processes are restored after terrible outcomes. This is often evident from the small signs of normal activities that resume in the wake of a catastrophe. Abandoning attempts to meet the sudden demand and instead reserving resources to reconstitute the system after the demand has passed could, in our view, qualify as a resilient performance. But in each instance we note that resilient performance is not present because of a certain form nor is it absent if the outcome is undesirable. Instead, the resilient performance itself involves either a tacit or explicit redefinition of what it means to be successful.
Changing the Definition of Success
These examples and other cases of resilient performance involve a shift in goals and redefinition of what constitutes success. To perform resiliently, practitioners necessarily defer or abandon some goals in order to achieve others. Negotiating trade-offs across multiple goals is common in complex systems operations and it should not be surprising to find this type of activity in resilient performance. In these cases, however, the trade-offs occur at relatively high levels within the goal – means hierarchy. Rather than minor trade-offs, the sudden demand presents opportunities for a significant failure. The sudden demand creates a conflict that requires practitioners to refer to high-level goals. The resilience in resilient performance is the degree to which the operators are able to handle the disruption to the usual goal – means features of their work.
Traversing the Goal – Means Hierarchy
Rasmussen’s goal – means hierarchy represents the semantic structure of the domain of practice. During operations, practitioner cognitive performances map a set of transitions through this hierarchy. In routine operational situations, the transitions are likely to map a small space at low levels of abstraction as workers perform ‘routine’ activities and achieve ‘routine’ results. Cognition at these relatively concrete levels of the hierarchy does not mean that the higher-level goals are not active. It is only that worker cognition does not engage high-level goals because the work activities do not produce conflicts.
Reference upwards in the goal – means hierarchy occurs when conflicts arise between goals that are located at the same level. The need to refer upwards in the hierarchy is brought about by the presence of conflicting goals. Most operational settings are set-up so that conflicts are rare. (If they were not rare, their operations would be very inefficient!)
Resilient performances are remarkable because the sudden demands that provoke them generate the need for practitioners to engage quite high levels of the goal – means hierarchy. A resilient performance is a cognitive performance that ably traverses the hierarchy in order to resolve a conflict. Within Rasmussen’s goal – means hierarchy, a sudden demand will register as the appearance of a challenge to a goal high up in the hierarchy. The appearance of this challenge leads to the rapid working out of various trade-offs at lower levels as the means to cope with this challenge are sought by workers. This working out appears to observers as a resilient performance.
Taking it in One’s Stride
Our colleagues regard resilience as a desirable and distinct feature of some systems and we agree with them. The skill and expertise that people use to cope with the challenges they face is frequently impressive and sometimes astounding. The robustness of resilient performances is often remarkable. The response to a bus bombing is a notable example because it is so vital and so dramatic. But the handling of a ‘soft emergency’ by the anesthesia coordinator is also remarkable, in part because it is so graceful and subtle.
When they are confronted with potentially disruptive demands, the people involved continue working. They direct their attention to the emerging threat to a high-level goal. They sacrifice lower level goals to meet the challenge but do so in ways that preserve the system. They monitor the escalating involvement with the threat and track the ebb and flow of operational responses. They strike a balance between committing resources to handle immediate needs and reserving resources for future ones. They deliberately accept smaller losses and failures in order to preserve opportunities for more valuable gains and successes. They anticipate the end of the threat and organize the resumption of ‘normal’ activities.
Those who are familiar with the work on cognition in process control settings will recognize that the litany of practitioner actions in a resilient performance is very similar to those of ordinary operations. This leads us to conclude that resilience is not a discrete attribute but rather an expression of some systems’ dynamic functional traits.
Resilience can be large or small scale, depending on how high up in the goal – means hierarchy the sudden demand’s challenge appears. The monotonous, familiar demands of a static world impose no need for resilience. The goal – means hierarchy for that domain may be complex. Practitioner cognitive tasks, though, are firmly bound to only the lowest levels of the hierarchy. Resilience is what we need to cope with the sudden appearance of a demand that puts a high level goal at jeopardy. Observers recognize resilient performances when the threat is abrupt and the goal put at risk is high in the hierarchy. The deeply resonant resilience that is so beautifully portrayed in the system that responded to the bus bombing is a quantitatively, but not qualitatively, different performance from that seen in the ‘soft’ emergency case. In both instances practitioners took the unusual event in their stride without missing a step.
We propose that resilient performance is empirical evidence of resilience. The need for resilient performance is generated by the appearance of a sudden demand for resources – a demand that threatens a high level goal. What distinguishes resilient performance is the fact that practitioners are able to move through the goal – means hierarchy to address the threat. In this formulation, cognition is the critical factor in resilient performances and the central feature of what it means for a system to possess the dynamic functional characteristic of resilience.
If our conclusions are correct, then research on resilience will likely be some combination of three themes. The first is research on cognition – including distributed cognition – in demanding situations. The second is research on the explanation of goal – means hierarchies in naturalistic settings. The third is research on the characteristics of sudden demands for resources and the reactions that they evoke. Understanding resilience is likely to depend almost entirely on assembling research approaches that explore the interactions among these three themes.
There is good reason to be encouraged about the prospects for successful research on resilience. It is certainly difficult to study dramatic events such as the bus bombing we described. But resilience is also found in the everyday operations of complex systems. It is part of the mundane, messy details of real work in the hospital and, we are certain, elsewhere. Although the scope of the threat in these two cases is very different, the mechanisms of resilience are, we believe, the same. The results that we obtained from the study of the work of anesthesia coordinators offer a means to understand the dramatic response to large scale disasters such as bus bombings.