I have used task analysis for many years and have always found it very useful for a wide range of human factors applications.  However, in the last year or so people have started to show more interest in the analyses and have been asking more probing questions about the method, presentation and application.  This has led me to develop a more comprehensive template for recording the findings of task analyses.

Download my Task Analysis Template issued Christmas 2013

The purpose of this paper is to present my new template and to answer the questions that I am being asked (and some that I have asked myself over the years).  I hope that by sharing this template more people will start to see the potential value of carrying out formal analyses of their most safety critical tasks.  Also, I would like to hear from anyone willing to share ideas about how the template could be improved further.  I am acutely aware that analyses are now taking longer to complete, and that there has to be a balance between the level of detail vs. the time and effort involved.

I have not changed the basic task analysis techniques that I use.  I still find that Hierarchical Task Analysis is a very good way of developing structured and systematic descriptions of task methods; and HAZOP style prompt words are effective when identifying potential human failures.  But there are additional activities that can be performed that lead to a deeper understanding of how a task can contribute to accidents and the methods used to control risks.  This is proving to be particularly effective in the process safety field where avoidance of major accidents is the main aim.

1      Questions and Answers

I have developed the template to answer the questions that have been asked about task analysis over the years.  Not all of these questions relate directly to features of the template, but I hope they give you a good idea of how all the key elements of a task analysis fit together.

Which tasks should be analysed?

Completing a task analysis to the level of detail required by this template takes time.  This means it is particularly important to focus on tasks where there is likely to be the most benefit for this effort.  This typically means that there are hazards that can result in major accidents and the task has a degree of complexity or other features that make it potentially vulnerable to human error.  There are different ways of identifying these tasks. I described my normal approach in a previous paper[1], which advocates the scoring method presented in HSE report OTO 1999/092[2].  I know there are other methods of prioritising tasks for analysis.  However, what is important is that the approach taken is systematic and focusses on both hazard and the potential for human error.

How does the analysis start?

Agreeing the task title and any assumptions or preconditions are the first stages of any analysis.  But it is also useful to discuss and note potential major accidents that may be associated with the task.  Any relevant safety case or report should be referred to.

One point to note: experience shows that safety cases/reports do not always cover every potential major accident.  This seems to occur because of a focus on technical rather than human failures when identifying scenarios.  Whilst a potential can of worms, this is a clear indication of how out task analysis and can contribute to the wider process safety topic work scope and should be an integral part of the development of a safety case/report.

How do you analyse the task?

I suggest you use Hierarchical Task Analysis to map out the task method and a list of prompt words for identifying potential human errors (e.g. a task HAZOP).  These methods are discussed in my previous paper.

Section 3 of the template is used to record the findings of the analysis in a tabular format.  It includes the columns that I find sufficient to record the necessary details.  I know there are other templates with more columns, suggesting that there may be more information that can be recorded but I personally feel that this table is sufficient.  Also, you rarely gain much by using additional columns but end up spending more time trying to decide where text needs to be recorded.

Should you record every possible error or just the ‘important’ ones?

There are two schools of thought about whether you should record errors for every task step or just the ones that may have serious safety or environmental consequences.  I tend to record errors for every step, even if the consequence is of less importance (e.g. financial or commercial) for a number of reasons, including:

  • It really does not take long to record the less important errors;
  • It demonstrates that every step of the task has been examined.  This is difficult to do if you don’t record errors for all steps as it is not clear if an error was not recorded because it was considered to be unimportant or because the step had been overlooked;
  • Identifying potential financial and commercial consequences can be useful to the company.  Showing these additional spin-off benefits can help to increase ‘buy in’ to task analysis.

How do you link the task with potential major accidents?

Although I have always identified potential major accident consequences when carrying out task analysis, I have not always managed to create a clear link with the scenarios identified in safety cases/reports.  This has meant it was sometimes difficult to provide the full picture of how the human factors risks of major accidents had been identified, assessed, and were being managed.

The very simple solution I have developed to improve these links has been to add a standard code against any consequences identified in the human error analysis that is considered to be a potential major accident -  I use the abbreviation ‘MAH’.  This makes it very easy to look through a task analysis in order to pick out the steps of most interest from a process safety perspective.

How do you link Performance Influencing factors to a task?

I use the HSE’s list of Performance Influencing Factors (PIFs)[3] to carry out my assessments, but with 25 on the list it is impractical to review every PIF for every step of a task.

My solution is to review the PIFs for each main section or sub-task (not step).  I use the PIF column in the main assessment table (Section 3 of the template) to simply identify the ones that are likely to be most relevant.  I use a simple number code to refer to the PIFs from the list, and add a couple of words of explanation.

Even when only reviewing the main sub-tasks I find there is a lot of repetition, which is not really a surprise as most PIFs have a fairly wide influence.  I tend not to record duplicates.  As I work through the sub-tasks I just add any PIFs not already mentioned above.  Having been through the sub-tasks I do have a quick scan through all the task steps just to check whether any have specific issues that have not been identified at the sub-task level.

I do recognise that this approach does have some potential weakness because it may not link specific PIFs to task steps.  However, I feel it provides a suitable balance between effort required and benefits obtained.

How do you evaluate PIFs?

The section above describes how I identify PIFs relevant to a task.  It appears to me that lots of people finish their assessments at that point, which I feel is of limited value.  In order to create more value from the PIF identification I have developed a PIF evaluation, which I record in Section 4 of the template.

I complete the evaluation by reviewing the PIFs identified in the task analysis (see above).  I use the ‘key points’ column to explain why the PIF is considered relevant to the task, which then forms the basis of a site visit and task walkthrough.  The ‘site assessment’ column is used to record the evaluation, noting good and bad features.  The ‘action’ column is used to record recommendations for improvement.  In some cases there may be a number of key points associated with a single PIF, whilst other PIFs may not be considered relevant to the task.

How do you support your PIF evaluation?

I think it is inevitable that a PIF evaluation will always be quite subjective, which means it will always be open to challenge.  I have concluded that the best way of supporting the evaluation is to take photos and collect any other relevant images (e.g. control graphic print-out).  I use Section 4.2 of the template for this.  A short description is included, along with notes about whether the PIF was considered good or bad.

What do you do once the assessment is complete?

Like all assessments, there is no point carrying out task analyses unless you do something with the findings.  The action column in both the task analysis table (Section 3 of the template) and PIF evaluation (Section 4) are used to record recommendations that emerge as the analysis progresses.  Also, I believe it is useful to take a step back at the end and review what you have learnt.  I use this to form the main summary, which is presented in Section 2 of the template.  I feel this is really the output from the analysis, whilst the remainder of the information recorded on the template is the data used to make the assessment.

I use the Task Criticality Overview (Section 2.1 of the template) to determine whether the initial assessment of the task criticality has changed now that a thorough task analysis has been completed (i.e. it is possible that additional hazards or potential errors have been identified, or risk controls may be less effective than assumed).  As I use a scoring system to assign criticality when prioritising the tasks to be analysed it is quite easy to compare the before and after results.  Experience has shown that, whilst individual scores have changed, the overall criticality (i.e. high, medium or low) remains valid.

A second important part of the task summary is recorded under Major Accident Potential (Section 2.2 of the template).  I complete this by going through the task analysis and picking out the steps where the consequence was marked with ‘MAH.’  I aim to write a couple of sentences for all of the potential scenarios to explain how human errors can contribute to major accidents, the perceived risks and existing controls.  Reference can be made here to any relevant improvement actions.

How do you link task analyses to the company’s risk assessments?

Most companies have their own risk assessment methods, with many using matrices to determine overall risk based on potential consequence and likelihood.  A common request is for the task analyses to link in with these methods.  Section 2.3 of the assessment has been provided for this, although it needs to be tailored to the company’s method.  I typically list the major accident scenarios summarised in Section 2.2, although other health and safety considerations can be included.

To be honest, I find this adds little value to the task analysis, but it is relatively quick and easy to do and helps some people put the findings into context.

How do you demonstrate risks are As Low As Reasonably Practicable?

Ultimately, the objective of any risk assessment is to demonstrate that risks have been reduced as low as reasonably practicable (ALARP).  The test for this is to identify what else can be done that could potentially reduce risks further and justifying why they have not been (or will not be) implemented.  Part of this involves demonstrating that a suitable hierarchy of risk control has been implemented.

I have addressed this requirement by including a Risk Control Statement (Section 2.4 of the template).  This presents a high level set of potential risk control strategies (can the hazard be eliminated or reduced, can engineering or administrative controls be implemented?).  A column is provided to record current arrangements for each, and another for a discussion of the options available to reduce risk further.  Finally, a statement is made about the strategy in place and whether ALARP has been achieved or further action is required.

This is usually the last thing to do when performing a task analysis.  It is proving to be a very useful final review of the findings, and quite often further improvement actions are identified as a result.

 

What happens to the analysis?

An immediate requirement having completed a task analysis is to put a plan into place to address the recommended actions.  A table is provided (Section 2.5 of the template) to summarise all actions generated during the task and error analysis, PIF evaluation and in the summing up stages (e.g. when completing the risk control statement).  An important element of this is assigning ownership.  Unfortunately, it is still very common to find that improvement actions resulting from task analyses are not being fulfilled in a timely and effective manner.

The longer term requirement is to keep analyses up to date.  My view is that they can be viewed like any other risk assessment.  This means they should be reviewed on a defined frequency or as the result of change.  For sites dealing with major accident hazards and required to produce safety reports/cases, I suggest they link their task analyses to these documents.  I would suggest the time between reviews should be no more than five years.

 


[1] http://abrisk.co.uk/papers/Task_Risk_Management-practical_guide02.pdf

[2] http://www.hse.gov.uk/research/otopdf/1999/oto99092.pdf

[3] http://www.hse.gov.uk/humanfactors/topics/pifs.pdf