Flashcards
Workplace Design and Alarm Systems
Cards in this deck (32)
- 1Workplaces vs. AnthropometricsPhysiological side of things - workplace may not just mean "desk job" - different environments have different issues.
Workplaces = desks, computers, workstations, truck cabins, etc.
Anthropometrics deal with reach distances, height, weight, width leg length, etc.
Population data is used to ensure the general population can be comfortable at work without suffering.
Issues can come from the fact that they are modelled for the average individual, those who are outside of the norms may suffer, and biomechanics - when a worker has to move around the workplace sometimes you need to ignore anthropometric data - minimising the reach minimises the risk of an RSI, but does not fit with "average population data". There needs to be a balance between what the worker is actually doing, and that the average worker can comfortably perform tasks. - 2Principles of Workplace DesignClearance requirement of users (can people make their way easily into and out of a workspace?)
Reach requirements (can workers reach their equipment/ aren't cramped?)
Adjustability (can workers adjust their workspace/ can the workplace be adjusted for differently abled workers?)
Visibility (can visual displays be read easily regardless of environmental conditions?)
Component arrangement (displays and controls laid out in an easily reachable and perceptual way) - 3Consequences of Ignoring AnthropometricsCumulative trauma disorder - carpal tunnel syndrome/ RSI
Absences & sick pay
Reduced productivity
Potential for error increased - 4Principles of Display Design(Visual display design)
1. Perception
2. Mental models
3. Attention
4. Memory - 5Kegworth Air Disaster (1989)British Midland Flight 92 (new design 737-400) from Heathrow to Belfast crashed near the M1 near Kegworth. As the aircraft was climbing the outer panel of the No 1 (left) engine detached, damaging the engine. This resulted in shuddering and smoke and fumes on to the flight deck. The crew erroneously believed that the No 2 engine was the one which had suffered damage. Coincidentally the juddering from engine No 1 ceased during that operation, so they then proceeded to shut No 2 engine down. A short distance from the runway No1 engine lost power. They attempted to make an emergency landing and failed to do so, crashing into the M1.
47 deaths. - 6Identified Potential Causes for the Kegworth Air DisasterSmoke in cabin indicates that the engine from which bleed air is taken will have smoke in it. The pilot thought the bleed air was taken from the right engine – but in the new design 737-400 bleed air was drawn from both engines. The smoke is no longer giving you an indication of which engine it is coming from, merely that one is in trouble. Memory failures - cockpit crew used to old design and based actions on memories of previous model.
No visibility of engines - perception: if you can't see what's going on you have to draw all conclusions from the instrumentation in front of them.
The vibration sensors were tiny and had a new digital display style (perception).
On the 737 the vibration sensors were previously known to be inaccurate, but had now improved, but the crew ignored them based on old plane. (memory).
There was no alarm system based on engine vibrations (alarms).
The instrumentation was split into primary and secondary instruments rather than split into left and right - the instruments didn't mirror the arrangement of the aircraft (mental model). - 7Design Issues with the PlaneThe dials in the new plane were much smaller, and so harder to read. This is an issue of LEGIBILITY.
The dials were split into primary and secondary instruments, which was counterintuitive, they would have been more easily interpreted if they were split left to right. Issue of 'pictorial realism' - left right association was broken. - 8PerceptionLimitations:
Physiological issues:
Near / far sightedness
Fatigue / eye strain (close work / computer monitor)
Illumination
Accomodation (change in focus)
Colour blindness
Top down processing:
Prior expectations guide perception - experience dictates how we expect a workplace to operate.
Difference threshold:
Smallest level of difference required to perceive two stimuli as different
Signal detection theory - 9Preventing Perceptual IssuesMake displays legible:
If components, objects or sounds are not discernible, the operator cannot make use of them.
Legibility is necessary for information processing:
Consider legibility under different environmental conditions, e.g. at night, in bright lights, etc.
Avoid absolute judgement limits:
Operators should not be asked to judge a variable based on a single parameter, it can be difficult to perceive differences - make as much contrast between signals as possible.
Limit the levels of differentiation to a maximum of 7 - make it as simple as possible to remember. - 10PerceptionTop–down processing:
Expectancy can guide perception. If a signal is to be presented that is contrary to expectation that signal must be enhanced - if only one switch should be on it could be a different colour/ size.
Redundancy gain:
Multiple presentations increase the likelihood of the signal being interpreted correctly - colour and location in traffic lights allows colourblind people to observe changes in signal.
Use discriminable elements:
Similarity can cause confusion. Unnecessary similar features should be removed and dissimilar features highlighted. - 11Mental ModelHow the workplace works, what we're doing, and the effect of our actions.
‘A mental model is a representation of some situation or domain that supports understanding, reasoning and prediction’ Gentner, 2002.
The ability to understand how an object works, or can be used, plus the ability to predict the outcome of an action in relation to that object.
Mental models are derived from experience and familiarity with an object.
Models need not be technically accurate but they must be functional – most of us can operate a telephone but have only basic knowledge of how the telephone network operates (Norman, 1983). - 12Mental ModelMental models can be incomplete (workers don't need to understand everything that's going on, merely their own tasks and consequences), unstable (subject to change, with change needed to accommodate them), and incorrect (roadrunner physics - if a ball rolls off a table, the ball should fall to the floor, participants suggested other options, such as the balls momentum propelling it for a second).
Mental models are not always intuitive or correct - it is important in the design of the workplace to guide individuals to formulate accurate and functional mental models. - 13Mental Model - Pictorial Realism and Moving PartsPictorial realism:
A display should look like the variable that it represents and match expectations (fuel gauge moves to right when full). If the display contains multiple elements they should be configured so that they match how those elements are in the environment - instruments should match movements.
Moving parts:
The moving elements of a display should mimic how the represented element actually moves in the environment. Increasing altitude should move element in altimeter upwards. - 14AttentionAttention - a limited resource (a process of selection and selective attention, Wood, Cox & Cheng, 2006).
Selective attention:
Allows the perceiver to choose the required information sources for a task and attend to those. Can result in reduced awareness of unattended stimuli. Difficulty in switching attention from one source to another in rapid succession.
Focused attention:
Allows the sources to be perceived without distraction from competing sources. High arousal (stress) can lead to a perceptual narrowing – leading to peripheral information being ignored.
Divided attention:
Allows parallel processing of two or more sources if required. Executing tasks simultaneously is possible until attentional limits are exceeded, at that point resources may be devoted to one task at the expense of the other. - 15AttentionMinimising information access cost:
Frequently accessed components should be readily available (same display, central location to user).
Route of access should be simple (e.g. right mouse click for menu).
Proximity compatibility principle:
Where a task requires two sources to be accessed these sources should be in close proximity, or linked through similar colour or shape (Gestault again). This will aid attention switching.
Multiple resources
Multiple sources of information can be processed more easily across different modalities (visual and auditory). - 16When Focused Attention Goes WrongInattentional blindness: failure by an individual to notice a stimulus that is in plain sight due to focus on another task.
Heads Up Display (HUD) in aviation: display projects visual information in front of pilots eyes or onto a helmet mounted display.
Aim was to allow pilots to continue to look through windshield with no need to look down to check instruments.
Simulator studies found that HUD’s increased the likelihood of a runway collision (failure to notice aircraft in far domain). - 17Memory - STM and LTM LimitationsShort-term memory limitations:
Order errors (items recalled but in incorrect sequence)
Intrusion error (item recalled that was not present due to us remembering previous tasks)
Magical number 7 (max amount we can remember), plus or minus 2 (individual variability).
Long term memory limitations:
Recall: information provided separate from an event by a lengthy period of time is not always recalled.
Salience: information must be presented in a salient way to prompt recall.
Consequences of not considering memory limitations: Officers forgot (due to overly complex instructions) how they should interpret certain events prior to Pearl Harbour (1941) – resulting in a submarine sighting going un-reported. Naval forces then unprepared for attack that followed. Instructions too complex can leave elements undone. - 18Memory - Consistence and Predictive AidingConsistency:
Long-term memory can trigger actions the user considers appropriate (actions they have always done).
Icons, actions and procedures can transfer across displays (workstations) if they remain consistent. Try to keep elements between displays consistent to aide learning. Microsoft Windows maintains consistency across display - tasks and processes are different, but usability is still retained.
Predictive aiding:
Anticipate the information required to execute a task.
Remove the most complex elements and replace it with perceptual task. - 19MemoryReplace memory with visual information:
Good displays require the user to use a mixture of long- term, short-term memory and visual information to complete a task.
Surgery checklists (Soar et al., 2009):
Adverse event rate of 3 – 17% in UK hospitals.
Surgery is the most common location of an adverse event (129, 419 events in UK, 2007).
Checklists were designed to improve communication within the surgical team, prevent reliance on memory (for surgical site, tools used etc) and ensure standard procedures.
Three stages: prior to anaesthesia, prior to first incision, prior to patient’s departure from surgery.
Deaths reduced by 47% after checklist in-place in select NHS hospitals. Subject to the Hawthorne effect - the presence of observers of the new system could have prompted behavioural changes. - 20AlarmsAlarms
“an unexpected change in system state, a means of signalling state changes, a means of attracting attention, a means of arousing the operator and a change in the operator’s mental state.“
Environments where alarms are important:
Intensive care
Nuclear power plant
At home (fire alarm)
Everywhere......
Response to human limitations:
Warn of change in process or system
Draw attention to relevant area
Prompt immediate action in an emergency
Modality:
Visual (usually part of visual display)
Auditory
Combined - 21Model of Alarm ResponseRoute taken influenced by:
• Perception of alarm (or not)
• Alarm uncertainty
• Choice of actions
• Situation context (critical or
routine)
Route 1: Ignore
Route 2: Monitor for further change (will we need to take action)
Route 3: Superficial investigation Route (get rid of alarm)
4: Further investigation Route 5: Seek cause
We may not have heard it/ know what to do/ understand what it signals, etc, etc. - 22Texaco Milford Haven Refinery (1994)Severe electrical storm caused plant disturbances 5 hours after storm, led to fires and eventual explosion.
26 people injured, £48 million in damages
Alarm issues:
Multiple alarms, undifferentiated in action and indication of alarms.
Control room displays unclear.
During final 11 mins the operators had to recognise, acknowledge and act on 275 alarms.
Due to the faulty alarm system, the fire remained undetected and led to an explosion. - 23Common Alarm ProblemsWorkers tend to ignore alarms as a whole and fail to recognise when a major alarm sounds.
- 24Visual Alarm Design (Carvalho et al., 2011)Conducted investigations into power plant workers for 30 hours in simulation - were asked to verbalise all actions.
Red associated with alarm or failure:
Must be used consistently
Undermined if red also associated with function.
In the system studied, red was also associated with normal function, it ceases in importance.
Alarm should be informative
Lack of information can confuse the operator
Text should be legible
Abbreviations or unclear text can increase operator response time.
The operators of the power plant did not speak the language displayed on the alarms, and new workers were unfamiliar with abbreviations, and due to alarms being poorly differentiated, workers became confused.
Importance of colour:
Many bright, saturated colours displayed together can fail to draw attention to critical states. The selective use of colour can draw attention to abnormal states. Red should indicate serious problems, with another colour representing normal function. - 25Auditory Alarms (Patterson, 1990)Auditory alarms must be loud enough to be heard over background noise (15 dB above threshold of hearing above noise level within the working environment).
The alarm should not be overly startling or abrupt - if a worker is startled they may make errors.
The alarm should not disrupt perception of other signals or impair speech communication - they need to share information and coordinate.
The alarm should be informative – indicating to the listener the type of emergency and guiding the listener towards taking the appropriate action.
The alarm should be distinct to avoid confusion with other alarm signals. - 26False and Nuisance AlarmsFalse alarms (aviation):
If alarms are too sensitive they may activate frequently when conditions do not warrant an alarm it causes distrust, reduced alarm response frequency, slower response times and poor task performance (cry-wolf effect).
Aviation events: 28% were linked to false alarms where the alarm system activated unnecessarily and directed the flight crew to take inappropriate action or were a distraction. Not only can faulty alarms lead us to overlook problems, but can make us actually cause them.
False / nuisance alarms (healthcare):
Alarms designed to indicate change in physiological function – but change NOT always indicative of crisis.
Alarms usually 80dB (noisy) which can lead to sleep deprivation and stress for patients.
There can be up to 40 alarm sources per patient.
Determining alarm significance difficult due to similarities in pitch and tone, limiting actions of nurses due to them being unable to diagnose urgency.
Suggestion: alarms need to differentiate high and low urgency (at least three levels). - 27Medication AlertsWhen prescribing patients medication, alerts pop up on allergies and conflicting medications. They come up in the form of textboxes, which can easily be clicked to close and ignored.
Phansalker et al. (2014): Assessment of medication alerts for compliance with human factors principles.
Majority clinical decision support alerts over-ridden.
Method:
Developed tool to assess medication alerts: IMeDeSA. Evaluate screenshots of alerts
Results:
Variation across all systems, with issues identified in: alarm philosophy, prioritisation and learnability / confusability. - 28Human Factors Principles - Alarm Philosophy and Learnability and ConfusabilityAlarm philosophy:
Does the system provide a general catalog of unsafe events, correlating the priority level of the alert with the severity of the consequences?
Learnability and confusability:
Are the different severities of alerts easily distinguishable from one another? For example, do major alerts possess visual characteristics that are distinctly different from minor alerts? The use of a signal word to identify the severity of an alert is not considered to be a visual characteristic - 29Human Factors Principles - PrioritisationIs the prioritisation of alerts indicated appropriately by color? (ie, colors such as red and orange imply a high priority compared to colors such as green, blue, and white).
Does the alert use prioritization with colors other than green and red, to take into consideration users who may be color blind?
Are signal words appropriately assigned to each existing level of alert? For example, ‘Warning’ would appropriately be assigned to a level 1 alert and not a level 3 alert. ‘Note’ would appropriately be assigned to a level 3 alert and not a level 1 alert.
Does the alert utilize shapes or icons in order to indicate the priority of the alert? (ie, angular and unstable shapes such as inverted triangles indicate higher levels of priority than regular shapes such as circles).
In the case of multiple alerts, are the alerts placed on the screen in the order of their importance? The highest priority alerts should be visible to the user without having to scroll through the window. - 30Inattentional DeafnessSame as inattentional blindness (e.g. HUD in aeroplane is equivalent)
Giraudet et al. (2015): Impact of cognitive load on perception of auditory alarms (inattentional deafness: auditory alarm not perceived)
Method:
Simulated air-traffic control task - complex
Results:
28.8% of alarms not reported (not perceived at all!) under low cognitive load. 46.2% not reported under high cognitive load. Number of visual notifications observed also decreased.
Workload should therefore be considered when designing alarms to improve perception rates. Complexity and load need to not drown out alarms. - 31Joplin Tornado (Paul & Stimers, 2012)Massive tornado struck southwestern Missouri, killing 162 people. A large number of Joplin residents did not take shelter after the first warning siren:
‘we hear tornado sirens all the time and nobody pays attention to them’. The area is regularly hit by low level tornadoes that elicit the same warning type: there was no differentiation to indicate this tornado was more serious. - 32SummaryIf the four main categories of display design are adhered to the likelihood of error is reduced.
When those principles are disregarded, errors such as the Kegworth air disaster result.
Alarm systems are vital to avoid error in the workplace and guide appropriate worker responses.
Alarms should be designed along HF principles to avoid false alarms, nuisance alarms or alarms that are not understood.
Information
- Description
- Flashcards on Workplace Design and Alarm Systems, created by claire.macaulay on 09/12/2015.
- Created by
- claire.macaulay · Dec 2015
- Updated
- Feb 19, 2016
- Views
- 3
- Rating