01 May 2009
Jocelyn Dorrell

Early in the morning of 22 May 2006, a senior lifeguard at a Total Fitness gym in the Wirral mistakenly added chlorine granules to an acid tank rather than the chlorine tank. The result was a potentially deadly chlorine leak, which led to the evacuation of the whole building and five people being taken to hospital.

Fining Total Fitness £15,000, the judge said the incident was largely due to "an alarming lapse of concentration" on the part of the lifeguard, who was both trained and experienced. Total Fitness' risk assessment had taken into account the possibility of human error: the tanks were colour-coded, the keys were kept securely, and a written procedure required a second person to be present during dosing.

The judge acknowledged the safety document "was of a high standard". But Total Fitness had failed to ensure the policy was put into practice. Staff and management alike bypassed the correct procedure, creating the conditions in which a momentary lapse could lead to a serious incident.

9 out of 10

Human error is thought to be a factor in as many as 90% of workplace accidents, and has contributed to some of the most high-profile safety disasters of recent years, including the Paddington and Southall rail crashes and the capsizing of the Herald of Free Enterprise (see the Parliamentary Office of Science and Technology's (POST's) report Managing Human Error).

But even if human failings play a part in most accidents, the individual is usually not the (only) problem; rather it's the systems and culture in which they operate. In the Total Fitness prosecution, human error was the trigger cause, but it was management's failure to enforce procedures and monitor staff that created the circumstances in which the accident could happen.

"In certain situations," notes the POST guidance, "human beings will always make mistakes, and there is a limit to what can be done to modify behaviour itself ... In large-scale disasters, the oft-cited cause of 'human error' is usually taken to be synonymous with 'operator error' but a measure of responsibility often lies with the system designers."

Whatever your take on the argument, so long as it's true that mistakes cannot be eliminated, the health and safety manager's focus must be on error management: designing and enforcing systems that minimise the chance of an employee's oversight or lapse resulting in someone getting hurt.

Sally Wearing, director at human factors consultancy System Concepts, has found there are two common misconceptions in this area. "The first one is the belief 'there's not much that we can do'," she says. "I come across this defeatist view less often now, but it's still there.

"The second problem is that people don't look at the reasons behind mistakes. They just say some people are 'accident prone'."

New tricks

Before considering the systems and culture  employees operate in, it's worth eliminating other contributors to risk such as lack of competence, workload and fatigue.

When looking at how to reduce the number of human mistakes in a particular work process or area, the most obvious place to start is often training and competence. If you've recently changed a work process or introduced a new piece of machinery and there's been an increase in accidents, could it be that not all employees have been adequately trained?

"Take a step back and start with task analysis," advises Phoebe Smith, human factors specialist at the HSE's research arm, the Health and Safety Laboratory (HSL). "Get a clear job description and check you've selected the right people for the task. A training needs analysis will then highlight any training requirements."

Blanket training provision is unlikely to fit the bill.

"Tasks change, people change, technology changes," Smith points out, "and some people take longer to get up to speed than others. Focus on individual need: unnecessary training can be very demotivating so it's best to identify just those individuals who need extra training through a performance review process."

It's generally accepted that competence is a mixture of skills and experience, but it's not always the newer or less experienced employees who are most likely to make mistakes. One of the major influences on "skill fade" - the deterioration of competence - is task frequency. So a long-time employee who is redeployed to a different work process or who carries out a task very infrequently may be more prone to errors than a new employee whose induction training is still fresh in their mind.

Smith cites research into how much notice a set of employees took of chemical warning labels: the more experienced employees were less likely to look at the label, preferring instead to rely on their experience of using the chemical over many years, so didn't notice changes in its composition. Younger employees, while more likely to take greater risks overall, paid more attention to the hazard labels.

In high-hazard processes where the consequences of human failure can be very serious, staff will need regular training on tasks performed infrequently.

Recovery time

The effects of fatigue on human performance are well known. Tiredness was a
factor in the incidents at the Three Mile Island and Chernobyl nuclear plants, and in the 2001 Selby rail crash, in which 10 people were killed when a motorist fell asleep at the wheel of his Land Rover and plunged onto the East Coast mainline, colliding with an express train.

According to the Office of Rail Regulation - which took over rail safety enforcement from the HSE two years ago - causes of fatigue include workload, working environment, shift work, overtime, on-call working, travel and recovery time during periods of duty. Managing these factors will help reduce fatigue-induced errors. But first, cautions Julie Bell at the HSL, you need to make sure fatigue is the real issue.

"Have there been any significant changes recently that could have had an impact on staff morale or the organisational culture, and is it being reflected as general malaise?" she suggests. "Fatigue is usually interlinked with other human factors issues."

If you do believe you have a fatigue problem, says Bell, consider whom it affects: is it the whole workforce, a particular team or certain individuals? If it's a team, you need to make sure there are adequate staffing levels, and that highly skilled, in-demand employees are not overstretched by excessive workloads or extra shifts. If an individual seems to be affected by fatigue, check whether there are personal factors, such as a new baby, that you need to accommodate.

Where staff work shifts, make sure the shift pattern is well designed. "The type of work being undertaken is important to consider," says Bell, "because some tasks, for example vigilance and monitoring tasks, are more susceptible to fatigue than others. Rest breaks and task rotation are particularly important in these situations.

"Other important things to think about are overtime and shift swapping," she adds. "Companies often overlook these elements because employees make the arrangements between themselves, without the knowledge of senior management."

It may be tempting to see travelling time to and from work as being outside your remit, but some companies do take this into account. Bell offers the example of a rail contractor that employs staff on a 12-hour-shift basis. The firm now includes commuting time in the shift to ensure workers get adequate rest time between shifts.

Swiss cheese

To lower the chance of an accident occurring to the lowest reasonable level, a good system will have sufficient layers of defence - controls and checks - to make the chance of every layer failing very small. This idea is illustrated by James Reason's Swiss cheese model, where the holes (failures, defects and unsafe acts) in the various lines of defence must all line up for an accident to occur. The chances of this happening are low if you take steps to plug defensive gaps.

Human error is just one type of failure that may lead to the breakdown of one of your lines of defence, but provided the other defensive measures are solid, it shouldn't be the critical factor leading to an accident.

Examples of system deficiencies that can lead to human errors include an absence of key information at the point it's needed (lack of an indicator display, for example); poor prioritisation of alarms, overwhelming users so they can't diagnose problems quickly; and a poorly designed interface so users take the wrong actions when fatigued. The first deficiency was significant in the Herald of Free Enterprise catastrophe in March 1987: there was no warning light on the bridge indicating whether the bow doors were open or closed, so nobody was alerted when the car ferry sailed with its doors open (see 'The Herald of Free Enterprise' below).

Involving users in system design is fundamentally important.

"If the objective of a design is to enable humans to do something - ie it's not solely an aesthetic exercise - then the design process needs to be 'user centered'," says the HSL's David Riley.

It's important to consider all users, he explains, not just primary ones, so don't neglect maintenance staff and others who'll come into contact with the system.

Involving users will help you build up an understanding of what's required to undertake an activity (room for access, equipment/tools, time, information, other dependent or sub-tasks, the demands on the person, variability in user abilities), as well as identify areas where people may make errors.

User trials are a key part of the development process: these allow the end-user to spot design flaws they might not otherwise foresee, so modifications can be made in good time.

When training employees to use a system, it's worthwhile explaining why certain steps are critical, and how those steps fit into wider operations. Armed with the broader picture, employees are more likely to cope with unforeseen circumstances.

As far as possible, systems should function as users expect, or else they will have to learn to ignore their instincts about what should happen. The more complex a set of steps, the more difficult it will be to remember. Equally, steps should follow on from each other as logically as possible, so users don't forget or skip stages.

Checklists that require user confirmation of completed stages in a process are useful as a back-up to help prevent oversights caused by a lapse in concentration. A good system will prompt users to carry out actions in the correct order by presenting the right options at the right time (see the POST guidance).

But while systems shouldn't be unnecessarily complicated, there are also dangers in going too far the other way.

"You can make a system too simple," warns Sally Wearing. "People can be very competent but they may make errors if they don't have to think and they go into 'unconscious competence' mode. It's like driving to work on autopilot and not remembering the journey when you arrive.

"A key issue is not only minimising the likelihood of an error, but maximising the chance the person who makes the mistake will notice and be able to recover."

Building in safeguards such as those found in software packages can be helpful: a confirmation stage that asks "Are you sure you want to...?", for example, perhaps with the equivalent of an undo button. Other mechanisms include alarms, or a system of work that requires a second person to approve the action.

Safeguards are especially important where systems include non-reversible actions. But such safeguards should be used sparingly, and the mechanism varied, otherwise employees will be less inclined to take heed.

---

The Herald of Free Enterprise

A total of 193 people were killed when the Herald of Free Enterprise capsized on 6 March 1987 after it set sail from the Belgian port of Zeebrugge with its inner and outer bow doors open. The investigation into the incident uncovered a string of management and system failings leading to the fatal human error with the bow doors.

• Crews were under pressure to set sail 15 minutes ahead of schedule. This meant the chief officer, who was responsible for ensuring bow doors were closed, had to be on the bridge before all the cars had been loaded.
• The assistant who should have actually closed the doors was asleep when the ship sailed, having completed a maintenance shift.
• Management had failed to act on repeated requests for bow door warning indicators to be installed on the bridge. The equipment would have cost £400.
  

  ---

Neutral outsider

Errors and accidents are likely to be more common in organisations with a poor safety culture, where there is a lack of effective leadership on health and safety matters, and employees are not involved in influencing the safety management system.

Investigating incidents and near-misses will not only raise awareness of safety and the dangers of ignoring procedures, but also show that management takes safety seriously.

"We encourage our teams to report their own errors as well as those of their colleagues," says Steve Martin, managing director of sign installation company Xmo Strata, "not so we can discipline them, but so we can all learn from each others' experiences. It's very important that we share our learnings, because stopping people from being hurt is far more important than not looking foolish for a minute or two."

Martin offers the example of an on-site installer who inadvertently trapped a power cable between the signware canopy fascia and the structure they were fixing it to.

"We discussed it with those involved, reviewed the procedure and gave the fitters a toolbox talk and training," he says. "A safety bulletin, highlighting our failure, was sent to the customer and also published on our website, so that even our competitors can learn from our mistakes.

"No one was disciplined throughout this whole procedure and those who raised it were recognised and thanked personally by me."

The HSL's Caroline Sugden says she wouldn't ever recommend discipline as a response to human error incidents: "Not only will you not really learn how or why these incidents are happening, but also you won't hear things any more - employees won't tell you when something has happened if they're afraid they'll lose their jobs."

Sugden acknowledges it can be very hard for line managers to distance themselves enough to consider why errors are happening, so believes it's often best for an outsider (such as a human factors specialist) to come in; someone who's not in the line management chain and isn't seen as a threat.

"Line management may have preconceptions," she explains, "but a neutral adviser can ask employees, 'Why did you do that?', 'Does that happen a lot?'."

"Where violations are happening often," says Sally Wearing, "there are often very good reasons." She gives the example of an electronics firm where employees repeatedly removed the interlocks from machines. "After investigating we discovered the real issue was the design of the equipment," she explains. "The guards had been added as an afterthought and there were production issues."

Much will come back to training and competence: whether people understand the consequences of doing something in a certain way. But competence doesn't only apply to the workforce, points out Phoebe Smith. Supervisors and managers have a crucial role to play.

"Supervision is part of competence," says Smith. "It's critical to competence, picking up on substandard performance."

---

Slip, lapse, mistake or violation?

The HSE identifies four types of human failure, as follows.

• Slip: when a straightforward physical action, that is performed regularly, goes wrong: a worker presses button A rather than button B. Slips can be avoided through better design of controls.
• Lapse: a failure in concentration: an employee is interrupted during a task and misses out a stage as a result. Such lapses can be avoided by using checklists to tick off tasks, or through careful supervision.
• Mistake: a mistake results from a lack of understanding or an error in diagnosis planning, for example when an employee without the right information or experience tries to resolve a problem but actually causes another. Mistakes can be prevented by raising competence levels or using special procedures to resolve issues.
• Violation: a deliberate breach of procedures, such as knowingly fitting a wrongly specified part to avoid having to get a replacement. A positive safety culture, where staff follow procedures and learn from incidents, will minimise the chance of workers committing deliberate violations.

Source: www.hse.gov.uk/humanfactors/comah/03humansrisk.pdf

---

 

Categories:
Risk assessment, Risk assessment, Accident reduction, Article, Enforcement (prosecutions), Safe systems of work, Safe systems of work, Accident reduction, Enforcement (prosecutions)