Anxiety and Psychological Fatigue
The data presented in the previous section suggest that physiological fatigue is experienced by pilots in both acute and chronic states and that adequate recovery is a significant factor in both conditions. Psychological fatigue is shaped by our physiological state but is manifested through emotional and motivational responses to our environment. We also saw that the stress mechanism was triggered by both fear and anxiety. Whereas fear was an immediate reaction to an apparent existential threat, anxiety was a response to an intangible stimulus. The category of stress triggers associated with anxiety is often labelled ‘life stress events’. They are typically problems that we might not know how to deal with or what effect they will have on our lives. The common feature of life stress events is that they represent change and uncertainty, as the following example shows:
‘My marriage had been falling apart for a couple of years and it got to the stage where I would get up early to be out of the house before my wife woke up. On this particular day I did not manage to get away early enough and my wife and 1 had a very angry argument, which left me seething with rage as I drove to the airport. I remained in this state throughout flight planning and right up to the incident. I had been warned of a vehicle in the undershoot. I had been given radar vectors to get in between 2 faster aircraft and АТС requested that I land short in order to clear by an early taxiway. Just before touchdown there was a solid bump as I clipped the vehicle. A couple of days later I landed on the same runway and the vehicle was in the same place. It could not have been more obvious but on the day in question I just did not see it.
I realised afterwards that my total loss of concentration was caused by the fact that my mind was full of the row I had had with my wife. We later divorced and I understood for the first time what a strain I had been under. At the time I had been oblivious to the problem and if anyone had suggested that I needed help I would have thought that they were joking’.
(Source: UK Confidential human factors reporting programme - CHIRP)
The need to balance the demands of work and home life is reflected in the work/ life balance (WLB) concept. The pilots in my survey were asked to assess their own experience of dealing with the demands of home life using the work-home negative interaction scale of the SWING tool (Geurts et al., 2005). This scale assesses the degree to which work has a negative impact on home life and has been hypothesised to reflect the individual’s fatigue state. Figure 4.4 shows the percentage distribution of pilots in the top and bottom quartiles on the scale (higher scores reflect better WLB). Chronic fatigue and negative work-home interaction show a strong association. Demerouti et al. (2001) make the claim that burnout negatively relates
FIGURE 4.4 Negative work-home interaction (percentage of respondents in each quartile).
to happiness and generalises outside of work, impacting on overall life and that of families, too. This relationship is suggested by the survey data.
Hockey identifies three sets of demands we all have to manage: things we must do, such as eat and drink; things we should do, such as comply w'ith instructions; things we want to do, like spend time w'ith friends and family. At any moment we typically have multiple demands to satisfy. He proposes that psychological fatigue shapes the way we cope w'ith multiple demands. In Chapter 3, we saw' that HollnageTs (2009) efficiency/thoroughness trade-off principle proposed that performance can be evaluated in terms of the time invested in completing the task and the degree of thoroughness, or quality, with which the task is accomplished. Demerouti and her colleagues (2019) also looked at the relationship between burnout and performance in the flight simulator during recurrent training. No effect was found, illustrating the complex relationship between fatigue and performance. What the authors did find was that pilots showing high levels of burnout were unwilling to seek additional challenges in training that would improve their skills. Requesting to do more than it is necessary for training is a discretionary task that is happily shed by tired pilots. As well as affecting the way we manage out-of-work demands, psychological fatigue affects the way we cope w'ith the job itself. For the group of pilots in my study, when the data for the disengagement scale were examined, all of the burnout risk group were in the upper quartile on the scale (i.e. less engaged), whereas all of the ‘rested’ group were in the lowest quartile (most engaged). The data suggest that pilots will shed tasks or elect not to participate in discretionary action when suffering excessive psychological fatigue.
Work stressors include such things as your role in the organisation, relationships at work, concerns about career development and organisational structure and climate. In aviation, the lack of control over one’s personal life is often cited as a significant cause of stress. This loss of autonomy sets up a dissonance, contrasting the supposed high level of professionalism associated with the role w'ith apparently inflexible work schedules that conflict w'ith out-of-work demands. In an operational context, constant changes in schedules, alterations to the in-flight service, poor organisational communication, the feeling of too much to do before departure and the behaviour of some passengers are just some of the anxiety-inducing conditions encountered.
Data captured during a LOSA give us some idea of the rate of operational disturbance encountered by the crew. A ‘threat’, according to LOSA, is something external to the crew that has potential to cause a need to make a change to a plan or be prepared to take appropriate action at some point in the future. A threat can also be a constraint requiring the crew to be aware that their original plan might not be feasible. Threats, then, are perturbations that disturb the normal flow of work, and it is this induced variability or departure from the expected task that adds complexity and increases the workload. Threats can be considered as triggers for anxiety in that they represent problems that have to be dealt with as part of daily life, the implications and consequences of which are not always clear.
In a representative sample of 20 sectors, the average threat rate was 7, with a range of 2-17 threats per sector. The pre-departure phase of flight had the highest threat rate (46.9% of all threats). The need to deal with extreme weather represented 30.61% of the recorded threats of which ‘thunderstorms and/or turbulence’ was the single largest subcategory. Air traffic control (АТС) comprised 14.28% of the threats, over half of which were АТС clarity. Threats will vary according to season and destination.
Threats are of interest because they influence the work process. For example, the need to repeat АТС transmissions consume time and can lead to rushed actions. Airport issues made up 14.28% of the sample of which three-quarters of the observations were to do with closed runways and taxiways. In terms of crew behaviour, these ‘threats’ usually require the crew to replan their routing or consider a different arrival or departure. A failure to effectively cope with threats increases the risk of an error being committed by the crew. We will look at error in more detail in a later chapter but, broadly speaking, 20% of errors committed by the crew in a sample of 177 flights were associated with a threat. I should stress that it would be wrong to assume a causal relationship between a threat and an error. Threats absorb atten- tional capacity, thereby increasing the risk of an error.
In an attempt to quantify the relationship between the interruptions, a commonplace disruption of workflow, and the propensity for error, LOSA observers were asked to note how many times the crew were disturbed during the period from their boarding the aircraft to closing the passenger doors. An interruption was defined as an occasion that required one or both crew members to pay attention to, or respond to, a request from a third party on the flight deck, be it cabin crew, engineers or some other agency. The observers were also asked to note whether the crew subsequently committed an error after the interruption (typically missing an item on a checklist or failing to complete a checklist). In the sample of 177 flights, 80% of departures had an interruption (range 1-22). Crew performance was not affected by interruption rates of up to 5 per departure. From six interruptions onwards, error rates increased and by ten interruptions the probability of an error was 100%. From 11 to 22 interruptions - the maximum observed - the error rate returned to zero: something had now gone wrong, and so the need for care was apparent. These disruptions to the work process, in all probability, represent a low-level anxiety trigger, akin to a moderate surprise. In short, an unexpected, novel event triggers a requirement to reconcile the difference between the planned and the actual situations. Interruptions and distractions occurring at time-critical phases, such as preparing to depart, consume time already allocated to fixed tasks. These low intensity ‘surprises’ can trigger some of the physiological components of the stress mechanism. In a study of short-haul pilots, Aeschbach et al. (2017) looked at what they called ‘hassles’, which equate to ‘threats’ in LOSA terminology. Although ‘hassles’ had no effect on fatigue, they did have a significant effect on the pilots’ subjective assessment of workload as measured by the NASA task load index. Threats, then, increase the workload, and this elevated level of demand increases the risk of error. The following report from a B-777 captain beautifully illustrates how the job of work is, almost by definition, subject to disruptions.
‘The NOTAMs included a reference to work in progress on a runway that needed a code to be inserted in the take off weight calculation page of the computer. In addition, the Maintenance Log noted an outstanding item, “Semi Gear Lever Sys”, which also required a code to be inserted. At this point, I offered our S/O the opportunity to set up the FMC, conduct a preflight flow and brief. I listened and followed along from the middle jump seat while in liaison with the ground personnel regarding cargo loading, refuelling and passenger boarding. The inbound flight arrived significantly late, and we wanted to reduce the outbound delay as much as possible. During the request to pushback we received instructions to expect runway 24L. Takeoff weight was within the parameters to utilise 24R (much shorter taxi), we obtained a new ACARS RTOW and amended the FMC with the new runway. During cruise we noticed that we had accidentally omitted the insertion of the code for the gear issue during the ‘amended’ final preparation procedure’.
This brief narrative is actually very complex. First, the two codes referred to are changes to the normal workflow, communicated to the crew by different means. The decision to let the SO participate in the preparation for departure was a discretionary act on the part of the captain and is a departure from the prescribed process. The delay, and the subsequent desire to reduce further delay, induced pressure in that the time available for the quantity of work that needed to be done was compressed. Of significance here is that, in all probability, the system was set up correctly in the first instance, but the change in the runway, and by now all the crew were back in their proper seats, would have caused the ‘gear’ code to be deleted and therefore need reinserting. This is a software design issue. That fact was missed by the captain and the FO as they managed the final checklists during the shorter taxi to the new runway: another time constraint. The ‘threats’ in the scenario include the changes arising from the two codes, the late inbound flight, the runway change and the interruptions during the final preparation for departure. Most crew reading this example would probably consider it to be a (nearly) normal day at work, and it illustrates the fragmented, disrupted nature of work.
From a system perspective, the boundary is the point at which the aircraft must be correctly configured for take-off. The margin is the space between the crew commencing their flight deck set-up activities to the point at which the aircraft is cleared to enter the active runway. Buffering is the system’s capacity to cope with the normal procedures contained in company manuals, the additional work associated with the set-up procedure not necessarily described in procedures, those permitted discretionary acts such as including the SO, contingencies such as operational delays and runway changes and also unanticipated variability, such as interruptions. We saw earlier that interruptions have a measurable impact. In this example, we see fairly innocuous, routine events overwhelming available capacity without the crew being aware, and the system boundary was breached. The implication of the failure was that, on rotation, a change to the geometry of the main undercarriage resulted in reduced ground clearance and an increased risk of striking the runway with the rear of the fuselage. Failure, or tolerance, was graceful. The system still functioned.
It is important to remember that individual differences, such as competence, prior experience and stress tolerance (neuroticism) will shape how we deal with disturbances. The subjective nature of stressors was explored by Vine et al. (2015) using a method that drew on the job demands-resources (JD-R) model illustrated in Figure 4.5. First, pilots were asked to evaluate a series of scenarios according to how demanding they thought the situation would be in real life (demand estimation) and how well they thought they would be able to cope with the situation if they ever had to deal with it (resource estimation). Subjects were categorised based on whether they saw the scenarios as challenges to be met or threats to be dealt with. The pilots then flew an exercise in a simulator that included an engine failure just after take-off, while wearing eye-tracking technology. Pilots who evaluated the initial scenarios as challenges tended to perform better in the simulator (as measured by accuracy of flying and assessment by instructor pilots) while those who fell into the ‘threat’ group demonstrated reduced ability to inhibit distractions, a more random scanning of displays and a higher search rate (less structured scan). Coping with the disruptive nature of work, then, depends, in part, on our disposition towards variability in life. Those with poor coping skills are more prone to experience anxiety.
The JD-R model (Bakker & Demerouti, 2007) brings together out-of-work and in-work factors as inputs and generates two outputs, exhaustion and cynicism, that then shape performance. Psychological fatigue captures how individuals respond to anxiety-inducing stressors and then act in response to task demands. It is one output
FIGURE 4.5 Job demands-resources (JD-R) model.
from the JD-R model, and at high levels of psychological fatigue, it is manifested in the effort we apply to a task such that efficiency and/or thoroughness are minimised or the task itself is abandoned: we engage in load-shedding.
The factors that trigger anxiety seem to be instrumental in shaping our level of psychological fatigue. Furthermore, physiological fatigue similarly reduces our resilience. However, I said earlier that it is not exactly clear if ‘fatigue’ is a condition or a symptom. The implication of this is that you do not have to ‘feel’ fatigued (that is, experience, or be manifesting the symptoms of, physiological fatigue) to still be demonstrating psychological fatigue-related behaviour. This next example illustrates the point.
A B-777 aircraft was on approach to San Francisco Airport after a 12-hour flight. The captain was the flying pilot. The weather was fine with good visibility. Having disconnected the autopilot, and as they were descending through 300 ft, the crew then heard АТС give clearance for another aircraft to depart on the crossing runway. This is a commonplace event at San Francisco. The crew' of four on the flight deck could see the other aircraft as it lined up on the runway. At 200ft, АТС instructed the B777 to go-around, the FO acknowledged but the captain countermanded the instruction. In effect, the captain has just disobeyed a law'ful command from АТС. At 98 ft the aircraft was given clearance to land. San Francisco АТС filed a complaint. When questioned, the captain said that he had rested well before the flight, had taken adequate rest during the flight and did not think that fatigue played a part in his decision. When asked why he did not comply w'ith the instruction, he said that he was concerned that there might be a risk of a collision between his aircraft and the departing B757 over the airfield. When asked what might have happened if the B757 had rejected the take-off, the captain was shocked: it had never occurred to him that this w'as a possibility. A cursory search of the ASRS database found several instances where, in identical situations to this one, the departing aircraft had, in fact, rejected the take-off causing the landing aircraft to go-around.
There are several aspects of this event that are of interest. In Chapter 2, we talked about risk. In aviation, in trying to avoid one risk, we can often set up a countervailing risk. So, instead of meeting over the airfield, a risk the captain wanted to avoid, the two aircraft could have met on the crossing runways. We saw' earlier that sleep- deprived subjects have a reduced ability to assess risk but, in this case, the captain reported good sleep. Load-shedding seems to be the most likely explanation for this event. Executing a go-around in busy airspace w'ith a very powerful, relatively lightweight aircraft at the end of a 12-hour flight is demanding. Landing the aircraft w'as the simplest thing to do despite the fact that it constituted a violation, resulted in legal action and could have ended in a collision on the runways.
Job demands and coping abilities are not absolute; they are estimations made on the part of each individual. As a result, we can see differences both between and within individuals in the extent to wfiich they believe they can cope with specific events. Many people experience stress when, in fact, the situation is well w'ithin their level of competence. Figure 4.6 illustrates the concept. At low levels of fatigue, personal and task goals will be aligned, but as fatigue increases, those goals become disconnected until a point is reached at which behaviour w'ill be influenced by those actions needed to sustain a belief in one’s ability to cope.
FIGURE 4.6 The relationship between fatigue and stress. (After Phillips, 2014.)