SWIFT (structured what if technique) is a brainstorming method used in safety-critical industries such as healthcare, radioactive waste management, offshore installations, control of major accident hazards plants and medical devices. As the name suggests, it uses a structured approach (using guide words) to generate what if? questions to assess the safety of a method or process. It can be applied to proposed or existing scenarios, and gives more reassurance that all the right questions have been asked than the method often used in risk assessments, where the questions asked rely solely on the competence of the person carrying out the assessment.
Consider the traditional process. A safety and health professional is asked to produce a work at height assessment. Consultation with the workforce identifies that the activities involve stepladders with six or eight steps. Any work that requires cherry pickers or scaffolding is carried out by contractors and would require a separate assessment. It seems clear that all work on a ladder can be covered by a single assessment. The assessment identifies controls, such as regular documented ladder inspections by a competent member of the team, plus pre-use checks by users applying the JOLLY method (assessing risks posed by job, others, the ladder, the location and yourself).
A short two-step ladder is used to make climbing over pipes in a congested plant room safer but has not been identified in the risk assessment. One day an electrician decides to use the short ladder to change a lightbulb in the middle of an office. He overreaches and falls, dislocating his shoulder. Following the accident, the assessment and the training are updated to make it clear that this is not an appropriate use for the short ladder.
The hazard and operability analysis (HAZOP) approach (see box on p 50 for more details) used in the process industries might have picked up the omission. But HAZOP requires the system to be examined component by component which can take days and involve pages of process and instrumentation drawings.
An approach that most risk assessors use, even if they don’t label it as such, is an unstructured what if? process. Perhaps as they work through the stages in moving, setting up and using a ladder they ask, “What if the ladder was damaged?” or “What if you had a lot of tools to carry?”
However, the questions rely on the experience – and the imagination – of the risk assessor. There is no structure to ensure that, so far as is reasonably practicable, the key what if? questions are asked. Without structure people tend to focus on what they know has already happened.
Making risk assessments a team effort will provide a greater pool of experience, particularly when it involves those carrying out the operation as well as those supervising, aided by safety specialists with a broader view. However, the tendency of a group to make weaker decisions because harmony in the group becomes more important than making the right decision is a recognised phenomenon, known as groupthink. A supervisor says “that could never happen” and the operators are too frightened or too polite to disagree.
SWIFT can overcome some of these problems. In a similar fashion to a HAZOP analysis, the SWIFT leader applies guide words to push the team to consider questions which might otherwise not have been asked, but the guide words apply to parts of the process or to the stakeholders rather than to individual components. In our ladder example, the guide words “too long” or “too short” would have led to the questions, “What if the ladder is too short?” and “In what situations could the ladder be too short?” Such questions might have revealed the existence of the shorter ladder, and the possibility that someone would consider using it for something other than its intended purpose. SWIFT could overcome the self-censorship that occurs in an unstructured group – allowing members of the group to think the unthinkable.
Any team member can pose a what if? question under set question categories based on the hazards of the task. The analysis uses category-specific checklists at the conclusion of each question category. The results of the meeting are recorded on log sheets in columns with the headings: What if?, Consequences, Existing safeguards and Recommendations. An example of a log sheet is at bit.ly/2zO1WSC
The advantages claimed for SWIFT over HAZOP are that it takes less time to apply (around one-third of the time by some accounts), it can be used at any stage in the development of a procedure, process or system, and it can be applied outside the process industries. Because SWIFT does not require detailed design drawings, it can be used earlier than HAZOP to eliminate problems during the preliminary design process.
Mark Collins, a chartered engineer specialising in safety-critical systems assurance and assessment, has used SWIFT to assess and adapt a procedure for the one-off deployment of some heavy and awkward equipment from the side of a small coastal vessel into the North Sea. As he explains, the subject matter experts he gathered for the analysis were not “table people”, so were more comfortable out on the water than in a meeting room.
For risk assessors used to applying five-by-five risk matrices, something might seem to be missing from this process
A hazard and operability analysis (HAZOP) provides a highly structured and rigorous approach to identifying deviations from normal or expected operations. It was developed for use in the process industry, but variations have been developed for other areas – such as FMEA (failure modes and effects analysis) and an escape, evacuation and rescue HAZOP for the offshore industry (bit.ly/2yTVoox).
In the original version, process characteristics such as temperature, flow and pressure are combined with deviations such as “more”, “less” and “reverse” to create questions to prompt the team to consider possibilities they might not otherwise consider. Every guideword combination is applied to every component to ask about the consequences of “too much” or “too little” flow through each valve, or “too late” or “too early” an alarm signal or “too low” or “too high” concentration of oxygen. If working without a structure, the issue of “too early” an alarm might not be considered, but it could raise concerns which otherwise would not have been voiced.
Collins prepared storyboards, using photographs of equipment to be used as well as computer-aided design drawings, to provide a model of the proposed process for discussion. He developed his own guide words, from which the what if? questions were generated. For example, “too wet” prompted the question, “What if the weather changes and there is too much rain to carry on safely? How will we recognise what ‘too wet’ looks like?” The guide phrase “too cold” prompted the question, “What should we do if it’s too cold? Will gloves be sufficient?”
Collins points out that the SWIFT team should not be constrained by the idea that all questions should be what if?; it is equally important to ask how, where, why, when and who questions.
While SWIFT provides greater certainty for a safety professional that they have asked the questions that need to be considered, it is not a technique that de-skills the risk analysis process. Every source of information emphasises two components: the skill of the SWIFT team leader, and the expertise of the SWIFT team.
There are limits
Robert Taylor, emeritus research fellow at the Technical University of Denmark, has used SWIFT analyses in the oil and gas industry and in healthcare. His overall finding is that while HAZOP is tested and reliable, “SWIFT is as only as good as the checklist you use and the people using it”.
Taylor’s research scores SWIFT lower than HAZOP for completeness but he sees SWIFT as a useful additional tool, being more flexible and more widely applicable.
A question such as, ‘Are you worried about anything?’ can get those issues out into the open
However, in occupational safety and health, it’s not useful to compare SWIFT with HAZOP. HAZOP is not a proportional method for most of the risks we face, and is not sufficiently flexible for the types of activities we review. From the literature and those with experience of using SWIFT here are ten ideas about how it could enhance your risk assessment process. You might already be doing many of these but if not, see how they could enhance the sufficiency of your risk assessments.
1) Involve the right people. All the references on SWIFT agree that the key to its success is the right team. Clients and high-level project managers are not always the most useful participants. Collins emphasises that the benefits come from the insight and experience of people who know the task, such as the operational staff.
2) Lead well. A good SWIFT leader prepares well. Before meeting the team, draw up your list of guide words and prepare your description of the system to be assessed. Collins used storyboards, but you could use method statements or procedures where they exist, or photographs of a workplace or equipment. Perhaps film someone carrying out a process and discuss each step that way. During the assessment the important skill is to drive the team to work efficiently, balancing the need to consider each possible problem with avoiding procrastination and time wasting.
3) Prepare well. Though the guide words are often generated by the SWIFT leader before the meeting, this could be done collaboratively either at a pre-meeting, or via emails where it is more difficult to get people together. Make use of past incidents, guidance such as the HSE approved codes of practice and the system description.
4) Define the scope. Risk assessment participants have different views of the scope of an exercise. Is it just looking at safety, or should environmental and security risks be considered? Does it include the work done by contractors? What about the preparatory work or clean up? SWIFT demands that a clear process or system is described before the assessment starts, so everyone is assessing the same thing.
5) Check for concerns. Collins recommends that the leader starts the meeting with open questions. If people are sitting in a meeting distracted by a big concern, they are less likely to engage in the process. “A question such as, ‘Are you worried about anything?’ can get those issues out into the open before you start the SWIFT,” he says.
6) Minute the process. As well as a competent leader, the SWIFT technique benefits from having a separate scribe. The note taker needs to know enough about the domain to understand the language being used, but not get tied up with their own point of view. In his 2009 paper Optimising Qualitative Hazard Evaluations for Maximised Brainstorming (bit.ly/2zgQQcn), Bill Bridges of the US-based Process Improvement Institute suggests: “While the scribe is completing the summary of the team’s discussion, the leader can move on to the next topic of discussion.” Otherwise, the team has to wait until the leader finishes their notes, and the leader might be tempted to rush the recording.
7) Meet where the work is done. After an initial round-table meeting, Collins took his team to the quayside, with the equipment laid out on dry land to give everyone an idea of the scale of the problem, and the conditions on the water. If you’re planning a lifting operation, stand in the location where the lift will take place, and examine the equipment to be used.
8) Plan breaks. Concentration can flag in meetings after only half an hour, so plan frequent short breaks (so people know when they are coming) and try to avoid day-long meetings. Two or three shorter meetings over the course of a week might be more productive, with time to write up findings and generate new what if? questions between the sessions.
9) Avoid matrices. Bridges recommends SWIFT meetings don’t try to apply risk matrices, as it inhibits brainstorming. He estimates that eliminating risk matrices “increases the number of scenarios found by 10-20%, and saves 25% in meeting time”.
10) Try variants. An alternative approach to using the full SWIFT process is to have an open brainstorming session first, so that people are not funnelled into one way of thinking, and then use SWIFT guide words and questions as a check of completeness.
How far you take SWIFT depends on how much time you have and what you are trying to achieve. A Health and Safety Executive Offshore Technology Report in 2001 (Marine Risk Assessment, bit.ly/2gTH090) pigeonholes SWIFT as being a method for identifying hazards only. However, the ISO standard on risk management techniques (ISO 31010:2009) classifies SWIFT as “strongly applicable” to all stages of the risk management process.
Collins applied SWIFT to review a proposed method, to make suggested changes to that method, and to review the effectiveness of the controls in the proposed and revised methods. A formal evaluation of the effectiveness of suggested controls is a weakness in many risk assessments, but a well-run SWIFT meeting can make sure that each suggested control is subject to the same rigour as the original process.
For example, if SWIFT was applied to a process where workers need to grip some equipment, and the guideword “too cold” suggested they would suffer if the metal parts of the equipment were frozen, gloves might be proposed. Further SWIFT questions might then raise the issue that wearing gloves could reduce sensitivity when handling other equipment.
For traditional risk assessors used to applying five-by-five risk matrices, something might appear to be missing from this process. The questions asked in SWIFT relate to outcomes and severity of outcomes – what would happen if someone placed their ladder on uneven ground and fell off? The likelihood of the event is less of a priority. “It will never happen” is an assessment of likelihood that is often proved wrong.
A structured and systematic method for identifying those hazards you need to be sure won’t happen is worth a try.