Desktop version

Home arrow Management

  • Increase font
  • Decrease font

<<   CONTENTS   >>

Lapses – Forgetting as a part of work

Chou and Funk (1990) found that 39% of task management errors related to failing (forgetting) to initiate task elements at the appropriate point in the sequence. In a study of prospective memory (that is, the memory function that keeps track of future intentions and the failure of which results in a lapse), Marsh and Hicks (1998) found that the difficulty of one task increased the probability of forgetting another task, the value increasing from 28% under conditions of low workload to 57% under high workload. Lapses are the result of forgetting to do something.

The Helios aircraft had undergone maintenance the night before the flight, and one task undertaken by the technicians was a check of the aircraft pressurisation. When they had finished their work, it was the view of the investigation board that the engineers had left the pressurisation mode selector in the ‘manual’ position (although this was contested by the technician involved). If it was, indeed, an oversight, then it is an example of a lapse: they forgot to do the appropriate action for some reason. In February, and again in March of 2005, the same problem was reported to the ASRS. The first involved an MD-88 and the second a Boeing 737-300. The completion of the maintenance task should have triggered the behaviours necessary to return the flight deck to its normal state. Such actions may sometimes be prescribed in procedures but, more often, ‘work’ comprises procedural elements embedded in a generic process, usually referred to as ‘common sense’. This episode is indicative of the true nature of the complex w'ork undertaken by skilled practitioners. Because we simply cannot specify every single action to be taken at every stage throughout the day, we rely on ‘expertise’ to fill the gaps. Increased expertise allows for greater autonomy in the way work gets done. However, autonomy also means variability and increased risk of a breakdow n in control. Control and autonomy are traded off in hazardous systems in order to achieve flexibility, efficiency and safety.

The use of checklists to manage activity is variable across the aviation industry. Although aircraft and equipment manufacturers develop checklists as part of the airworthiness certification process, the complexity of checklists will vary according to the nature of the aircraft. For smaller aircraft, like the Cessna Caravan we looked at in Chapter 1, it is not uncommon for pilots to work completely from memory once they have mastered the aircraft’s - admittedly succinct - checklist. For more complex aircrafts, checklists are used in a variety of ways. In some cases, they are used in a ‘challenge and response’ sense in that one pilot calls an item off the list, and another verifies that the action is complete. In other cases, the pilots complete the checklist actions from memory and then use the document to verify that actions have been completed. In some multi-crew' aircraft, each pilot is responsible for completing their assigned portions of the checklist, which they do in a ‘silent’ manner. More modern aircraft have electronic checklists that incorporate locks or warnings that protect against incorrect or incomplete checklist action. The checklist process can also differ between normal and abnormal or emergency situations. These different approaches to the same task are an embodiment of decisions about job design and task management. A common response to a previous failure is to add additional steps to the checklist in an attempt to prevent a repetition.

We saw in the previous chapter that work comprises bundles of interwoven tasks often involving interaction w'ith other agencies. Concurrent tasks compete for attention, and this can lead to lapses. In February 2005, an MD-88 was climbing after take-off, having been initially cleared to 6000ft. The captain, w'ho was the pilot monitoring (PM), noticed that the aircraft w'as accumulating rime ice and so he turned on the w'ing anti-icing system. At this point, the crew received a further clearance to climb to 13,000ft. Soon after recommencing the climb, the pilots received a report from the cabin of a squealing noise coming from the rear galley, quickly followed by the cabin altitude aural warning. By now, the aircraft was at 10,000ft, and the crew' realised that the cabin altitude control selector w'as in the manual position, something that should have been picked up in the after-take-off checklist. At that point, the captain remembered that, because of having to deal with the icing conditions, the checklist had not been completed. Coincidentally, the selector was in the manual position because the maintenance personnel had conducted a pressurisation check during the previous night, exactly the same task that had been conducted on the Helios aircraft.

Because, as we have just seen, complex control requires several tasks to be running concurrently, the nature of work often requires us to interrupt tasks while we attend to another, more important, task. The fact that we work in a social context also means that other team members who need our attention, momentarily, as part of their work process sometimes interrupt us. This is the essence of prioritisation. When we interrupt a routine, we run the risk of failing to complete the action sequence or of resuming the process at the wrong point in the sequence. In December 2000, a B-737 crew were doing their cockpit checks prior to departing from Cork in the Irish Republic (AAIU, 2001). The captain noticed that his passengers were attempting to board an empty aircraft parked next to theirs. He tried to alert ground operations to the problem, but a line technician luckily intervened. During the engine start sequence, the crew encountered a problem, which needed the support of an engineer. Next, while completing the after-start checklist, the crew were interrupted by the cabin supervisor who informed them that the front left cabin door was making a strange gurgling noise. The FO went to investigate and, on his return, the crew agreed that it was probably water in the seal. At that point, the forward toilet fire alarm was triggered by smoke from the galley ovens where the cabin crew were heating the passengers’ breakfasts. Once they finally got airborne, it is not clear if the crew completed the after-start checklist, but what we do know is that they forgot to set the pressurisation. We will return to this crew later but, for now, theirs is a classic illustration of the consequences of being interrupted.

The crew of the Cork aircraft were dealing with external factors, but we also have to recognise that tasks can be deliberately interrupted or modified. In January 2003, the FO on a Boeing 737-300 was giving a briefing to airport fire service personnel, which included a discussion of how to configure the aircraft for an evacuation. During the demonstration, the FO showed the visitors how to turn the pressurisation to manual. After his guests had departed, the FO forgot to reset the panel. An event reminiscent of the Helios tragedy, it shows how quickly task requirements - resetting the panel - can be forgotten. Interruptions and alterations to work routines represent sources of variability.

Skill-based behaviour, the precursor to slips and lapses, involves the execution of action routines associated with the configuration and control. Errors arise from the manner in which control is exercised. Of course, in complex manipulative tasks, like hand-flying an aircraft, individual competence and experience will have an effect on the sophistication of the control being exercised. Specific consequences associated with skill-based behaviour are:

  • • Failure to start a sequence
  • • Failure to complete a sequence
  • • Omitting steps in a sequence
  • • Incorrect sequencing of steps
  • • Inserting fragments from other sequences
  • • Doing the wrong sequence
  • • Inadequate manual dexterity

Slips and lapses flow from the execution of skill-based behaviour. Because skill- based behaviours have become highly automated through extensive rehearsal, attention is allocated in proportion to the perceived load. Ordinarily, as demand increases so should the degree of attention allocated to the task but the irony of being highly skilled is that attention management can lag behind demand.

The behaviours we have been looking at capture the flow of events that we see on the flight deck under normal conditions. Well-rehearsed work routines are executed according to the specific demands of the task. When discussing task management in Chapter 5,1 proposed that decision-making was a part of the normal management of tasks. Occasionally, though, there was a need to consider the suitability of planned goals. Changed circumstances dictated a need to adopt a new goal or even to create an innovative solution to a problem. And that exposes the operation to the risk of mistakes being made.

<<   CONTENTS   >>

Related topics