Accidents dont just happen they are caused and the answer to a successful accident investigation is to unearth the cause. The gateway into ascertaining the cause of an accident is to explore the sequence of events leading up to it. Understanding this sequence is the aim of most accident-causation models/theories (Oakley 2003).
Investigating an accident within an organisation is not pleasant for any safety professional. Outcomes may vary from minor injuries to fatalities making the investigation a very delicate matter. However, the process is a necessary one and is a key part of the occupational safety process. The investigation procedure involves the gathering of facts and information, finding the cause and the recommendation of corrective measures.
The comprehension of accident causation is essential to its success and to highlight and assist in accident complexity, authors have invented and developed a multitude of models and theories. Heinrich's Domino Theory led the way until the 1980's with a simple linear model until it became apparent that there were many more factors that influences the likelihood of an accident other than the direct cause. Following this, accident causation theories increased and have became more complex (SHP 2010).
The appeal of accident causation models/theories is understandable following a succession of large scale incidents such as the Harold of Free Enterprise ,Piper Alpha, Clapham rail crash etc and the need for answers. However, with such a selection of models and theories available to organisations and safety practitioners, and with little evidence of their worth (Stoop 2004), it is difficult to make a selection of the best to employ. As individual models use different approaches, the information can be inconsistent and therefore, to provide a more dependable platform for the investigation process, a combination of models may be more reliable (Katsakiori et al., 2009).
The purpose of this paper with the use of peer review research and academic literature is to evaluate the effectiveness of accident causation theory in investigating small scale occupational accidents and larger scale industrial incidents. This paper will primarily concentrate on two theories; Integrated Safety Investigation Methodology - ISIM and Work Accidents Investigation Technique - WAIT, looking at their place within large and small organisations, identifying their qualities and therefore comparing and contrasting their effectiveness.
Causation Theories
Accidents are defined as unplanned events with injuries, fatalities, loss of production or damage to property and assets as the possible outcome. The preventing of accidents is a complex matter especially with a lack of an understanding as to the cause (Raouf 2004).
Statistics published by the HSE for 2011/12 show that there were a total of 173 employee fatalities during this period (HSE 2012). Even though this number has significantly reduced over the past 10 years, the amount of deaths within UK workplaces is still not acceptable and proves that further work in this area is necessary. This is a huge challenge however for research, government, employers and safety professionals. Every accident is individual and presents a different scenario and if one tries to pigeonhole all accident the same, fundamental information may be lost in the process (Carrillo, Onieva 2012).
Accident investigation methods can be very specific and are designed as a practical tool to assist the user to complete the task of accident analysis in a particular environment. An accident investigation method is not always related to a particular accident theory and ISIM and WAIT are tools within their own right. Accident investigation methods fall into one or a number of categories: Tree techniques, Systemic, Human Factors or Sequential, and both WAIT and ISIM are based on systemic and the human Factors.
Multi-causality of accidents was introduced by Reason in the 1980's. He suggests that the accident causation process is an interaction between latent and active failures and for this interaction to be avoided, a proactive approach to safety management is vital (Reason 1990). As the active failure of an accident is usually easily identified, latent failures are more difficult as they are often hidden within the systems of an organisation i.e. lack of training, supervision etc,. James Reasons 'Accident Causation Model', commonly known as the Swiss Cheese Model is a theoretical model based on organisation hierarchy and human error (1990). Research suggests that human and organisational factors are behind most accidents (Johnson 2002) and the Accident Causation Model claims that an accident occurs due to a number of human errors within an organisation that line up in such a way that the accident was inevitable (1990).
Over the past twenty years, many accident investigation methods have been developed and many follow Reasons model. A competent investigation model should be applicable to the specific circumstances of the accident and this is why there are methods designed for different industries such as Norske Statesbaner (NSB) for railway sectors, accident evolution and barrier function (AEB) for the nuclear industry, Health and Safety Executive (HSE245) and WAIT for occupational accidents and ISIM for the aviation sector and other general transportation systems.
Integrated Safety Investigation Methodology - (ISIM)
(Investigation of Large scale incidents)
ISIM was originally developed by the Transport Safety Board (TSB) of Canada in 1998. This method of accident investigation came about when the Canadian public began to raise concerns as to the appropriateness of the systems in place for the investigation of accidents and initiated appropriate changes (TSB 2000). ISIM was the outcome of this reform, providing an accident investigator with the structure and techniques required to be able to dig further for details beyond the immediate cause. ISIM also encourages investigators to be better armed in the assessment of highest risks environments (Ayeko 2002). Ayeko quotes:
"ISIM embraces the 'defence in depth' philosophy which seeks multiple and diverse lines of defence to migrate the risks of normal human errors"
To minimise risks in transportation, ISIM concentrates on the identification of safety deficiencies and the assessment of risk associated with these deficiencies. The process comprises of eight systematic steps (Appendix 1) beginning with an initial assessment of the accident ('to investigate or not to investigate') and concluding with the effective communication of the identified risks to the people who will be responsible for making the necessary changes happen.
The investigation begins with the collection of information involving staff, tasks, equipment and environment conditions involved in the occurrence. Applying a systematic approach is vital at this stage to make certain a complete examination of the facts is possible. Once the "how's and why's" of the accident have been analysed, the sequence of events before and after the incident can be determined and the underlying factors (latent) can be identified. The next step in this methods is to assess the risk levels associated with such underlying factors and study the physical and administrative barriers to identify those that are ineffective and require attention. The aim of the ISIM method is to make sure that accident investigation and safety deficiency analysis are incorporated to generate recommendations to improve safety.
During the research for this paper there has been much praise for the ISIM method; it's easy and practical (Katsakiori et al., 2009), it will help the safety of life, property and the environment around the world (Johnson 2002). However, it has been difficult to find literature that highlights its limitations apart from the recognition that to ensure its ongoing effectiveness, there is a requirement for follow-up on the job training and sustained reinforcement of the job.
An example of a large scale incident where this accident investigation method may have been applied is the Tenerife Airport Disaster (detailed below) with a total of 583 fatalities and classed as the deadliest accident in aviation history. The ISIM approach would assist in the collation of relevant information regarding staff, tasks, equipment and the environment, analyse the sequence of events before and after the incident and determine the underlying contributing factors. As a consequence of this incident, substantial changes were made to airline regulations throughout the world which has resulted in improved aviation safety today.
Example of a large scale incident:
Tenerife airport disaster March 27th 1977
Total fatalities 583
Total survivors 61
The investigation concluded that the fundamental cause of the accident was that of Captain Veldhuyzen van Zanten who took off without take off clearance. It was suggested that the reason for his mistake might have been a desire to leave as soon as possible in order to comply with KLM's duty-time regulations and before the weather conditions deteriorated further.
Other major contributing factors were:
The sudden fog greatly limited visibility. The control tower and the crews of both planes were unable to see each other.
Simultaneous radio transmissions, with the result that neither message could be heard.
The following factors were considered contributing facts but not critical:
Use of ambiguous non-standard phrases by the KLM co-pilot and Tenerife control tower.
The Pan Am aircraft had not exited the runway at C-3.
The airport was (due to the rerouting from the bomb threat) forced to accommodate a great number of large aircrafts, resulting in disruption of normal use of taxiways.
Concluding a thorough investigation, KLM accepted responsibility for the accident due to the cause being human and organisational factors. The investigation identified safety deficiencies and the risks associated with them. As a result of the accident, major changes were introduced to international airline regulations and to aircraft and to changes in the cockpit. Hierarchical dealings were discouraged amongst crew and there was an importance placed on team work and mutual decision making and shared.
Work Accidents Investigation Technique - ( WAIT)
(Investigating Small scale occupational accidents)
WAIT was developed by Jacinto and Aspinwall in 2003 and was again based on a theoretical foundation originally developed by Reason (1997) and Hollnagel (2002). WAIT offers a comprehensive process that provides organisations with the tools they need to fully investigate occupational accident and near misses. As discussed earlier in this paper, Reasons model of 'Accident Causation' looks at accident sequence, active failures and latent conditions and WAIT follows this model and comprises of two sequential stages (Appendix 1).
Stage one involves a questionnaire that assists in the collection of relevant data/information and permits the identification of factors that influence 'active' failures (unsafe acts). When the active failures have been identified, stage two of the process involves a more detailed in-depth investigation enabling the investigator to identify connections between these unsafe acts and other factors such as human factors, employee duties and tasks and also organisational and management deficiencies hidden within the organisation (latent conditions).
Its method, as does the original model, draws attention to three areas, the organisation the workplace and the person. It offers a sequential and easy process (Katsakiori 2009) allowing investigators to examine all contributing facts in connection with these three areas of concern, known as; influencing factors, individual factors and job factors (Jacinto & Aspinwall 2003). The in-depth stage of this investigation process progresses further than legal obligations, it also offers organisations a functional tool that could subsequently improve company policies and safe working practices (2003). One of the main strengths of this model is that it offers companies with or without management systems in place and regardless of size or sector, the ability to identify weaknesses and as a consequence of this will help improve operating procedures (OHSAS1999). As with ISIM it has been difficult to address the effectiveness of WAIT. However, trials of this method carried out in industry by Jacinto and Aspinwall showed it to be promising with further experimental work required (2003).
An example of a smaller occupational accident where this accident investigation method may have be applied is the warehouse fire at BASF, Wilton, Teaside (detailed below). The WAIT approach allows a first stage simplified investigation to gather general data regarding the incident followed by a second in-depth investigation taking into account human factors and management/organisational conditions. The investigation into this incident identified many 'active' failures within the organisation and also many management failures. As a result of this investigation, the WAIT method would instigates the search for 'positive influencing factors' which would potentially allow lessons to be learnt and the organisation to improve preventative and protective strategies (Jacinto & Aspinwall 2003).
Example of a small scale occupational accident:
BASF, Witon, Teaside, warehose fire. October 9th, 1995
Total fatalities nil
Total injuries nil
No direct root causes for the fire was determined. However, the result of the investigation by BASF and the Cleveland County Fire Brigade suggests the probable cause was a fluorescent light fitting overheating causing the ignition of its Perspex refectory which dropped flaming molten plastic onto stored products below. The warehouse lighting was in continual use.
Other major contributing factors were:
The warehouse facility met the building regulations and was equipped with a range of fire safety features including fire doors operated by fusible links and smoke detectors which failed to close during the fire. No cause was established for this, however it may have been attributed to the fact that the warehouse did not become completely smoke logged as smoke vented through the roof. Hence, the smoke failed to activate the detectors which would have closed the doors.
The perceived risk was low and therefore no formal risk assessments for dealing with a major fire was undertaken.
The investigation process into this incident identified many failures. Because the warehouse was situated in the middle of the ICI complex there was the potential for escalation into a much more serious event. Also, prior to this incident there had been a number of major fires in which light fittings have been the source of ignition or in which they had contributed to the spread of the fire. Also identified was the value of having a well defined emergency plan and procedures in place as well as trained personnel to execute it.
Even though we have acknowledged that accident causation models and accident investigation methods could both benefit from further research and exploration (Lowalgi &Saleh 2012), (Molinero 2008), the advantages of ISIM and WAIT can be understood as both offer a structured method for the investigation process. It can be argued however, that it does not necessarily matter which model or method is used as its the application and experience of the investigator that makes the difference; this point is a justified one, however providing the investigator with a tool that assists the process adapted specifically for a particular sector, should have its benefits even to the most experienced investigator.
The implementation of an investigation method does come with a cost and in today's economic climate the question has to be asked; would SME's firstly have the recourses to implement such methods and secondly have the understanding and knowledge required to appreciate the benefits of such a method. Based on HSE 2011/12 statistics, large organisations and SME's continue to have their fair share of accident and incidents but it's the smaller companies that many studies suggest have higher accidents rates. Interestingly, a report by the HSE in 2003 estimated that it costs SME's approximately nine times more per employee to achieve compliance in health and safety than per employee of a larger organisation (HSE 2003). Therefore, even though the smaller organisations would benefit greatly from accident causation models or an investigation method, it can be assumed that in most cases the costs in their opinion may outweigh the benefits.
Conclusion
We have established that accidents and their control continue to be a challenge for organisations large and small. We also understand that an employer has a legal duty to provide a safe workplace for their employees and therefore, the accident investigation process is necessary and a key part of the occupational safety process.
Accident Causation Models were first introduced to investigate major incidents within larger high risk industries and were not necessarily suitable for smaller occupational circumstances. However, accident investigation has developed greatly over the years from Heinrich's Domino Theory, a single linear model to James Reason's approach suggesting that all accidents are a multi layered interaction between active and latent failures that occur due to a number of organizational and human errors.
Accident prevention hinges to a great extent on previous accident investigations and lessons learned from them. For this to be possible, there needs to be an understanding of the immediate cause (active) followed by identification of the other underlying factors (latent) that have contributed to the accident happening. Though it can be said that if an investigator focuses to strongly on the cause and not the 'sequence of events leading to the cause' the investigation could go off in a direction of blame, as many large scale investigations often do!
Even though root cause is the drive for all accident investigations, different models are based on different approaches and the key is to fit an investigation theory, not only to the organization and environment but also to the investigator. ISIM and WAIT are both based on 'Reasons Accident Causation Model'. ISIM has been specifically designed for use in transportation systems such as aviation, railway etc and WAIT designed for use in smaller occupational accidents across a broad range of industries. It has been said that ISIM is easy and practical and this is a very important feature. If a theory is too complex it may be too difficult to administer as the process will need to be communicated to both employers and employees. WAIT offers organizations a tool to fully investigate occupational accidents and near misses in the workplace, looking at the sequence of events and active and latent failures. As a result, WAIT should instigate the search for 'positive influencing factors' potentially allowing for lessons to be learnt, resulting in an overall safer workplace for the company.
Nevertheless, it has been noted that firstly, investigation theories in general require ongoing research and exploration, especially into human factors. As investigations often recognize human error, they fail to delve any further into the area of human behavior. Secondly, there is little evidence of their worth and it can be said that the selection of a model or theory is not easy. For a more dependable platform for an accident investigation, a combination of models is suggested and may be the way forward.
Acknowledging that larger organisations tend to have more resources necessary to manage occupational accidents, it is interesting that statistics tell us that SME's continue to have higher accident rates. This may be due to the fact that many SME's are struggling in today's economic climate to stay in business and the additional cost of implementing an accident investigation method in conjunction with their management system may not be considered an option for them. Knowledge and understanding of health and safety in general within smaller organisations continue to be a foremost issue for the UK and a criticism can be given of the help on offered to them. Recognition of the relationship between good safety practice (which includes accident investigation) and financial performance is vital. This matter needs to be tackled first before lessons can be learned and any further headway can be made in the reduction of accidents within many smaller companies.
Appendix 1
Integrated Safety Investigation Methodology - (ISIM)
C:\Users\Jules\Documents\My Docs\aMSc\Msc H&S\ISIM fig.JPG
Appendix 2
WORK ACCIDENTS INVESTIGATION TECHNIQUE (WAIT)
STAGE 1
Collect information through direct observation and by interviewing all people involved in the occurrence, whether or not they were injured. These people will provide their own description of the accident sequence and their part in the events. At the end of the interview, distribute to each person a list of standard questions, which may help to draw attention to other less obvious facts, and disclose further relevant information. Standard questionnaire is provided.
Step 1
Collecting information
Step 2
Identifying all active
failures
Decide which events constituted "active failures" by systematically searching through all the following possible categories: HUM (humans), E&B (equip. and buildings), HAZ (hazards), LOR (living organisms), and NAT (natural phenomena).
To help with the words, you may find it useful to use the checklists provided (classification schemes for all 5 categories).
Display all active failures in a table (column 1 of the table) in
chronological order - register each single event in one row.
Step 3
Establishing the applicable
influencing factors
For each active failure, search for possible "influencing factors", which might have facilitated or triggered the failure under consideration". "To do this, use the answers to the questionnaire in step 1, combined with the classification lists provided.
If more than one factor is encountered, subdivide the particular row - displaying the findings in column 2.
Repeat the search for each row of column 1 (each active failure).
Review the analysis and gather more information if necessary
Step 4
Comparing findings
with relevant Risk
Assessment(s) - RA
Compare all findings of columns 1+2 with relevant risk assessments. Risk Assessment is a legal and fundamental duty. Check if the hazards, human failures, and risks involved in that particular occurrence had been actually considered in the risk assessment(s). If RA exists and all risks were considered, ask why it failed to prevent that particular case. Establish whether or not the applicable RA is good enough or needs improvement.
The record of the risk assessment itself may draw your attention to other possible problems, either active failures or influences, which were not mentioned in the previous steps.
Other failures or factors could be identified as
probable?
Step 4 will help the investigator to determine if the previous ones are complete and whether the relationships encountered are logical, coherent and consistent. If no more relationships are found, this is the END of the basic investigation, and an in-depth analysis can be performed whenever necessary. In-depth analysis goes beyond official reporting duties and companies should have a criterion for deciding which cases need a full or in-depth investigation.
YES
NO
WORK ACCIDENTS INVESTIGATION TECHNIQUE (WAIT)
STAGE 2 - IN DEPTH INVESTIGATION
For each new row of column 2, search for human factors - within the individual(s) and the job, or working system - which are believed to have contributed to the active failures and their influencing factor. Note: the number of initial rows will probably have increased in column 2, by adding the context in which active failures occurred. Consider each one of them. To help the search, use the classification scheme provided for individual and job factors.
If more than one factor is found, then, subdivide each row again, and display the results in column 3.
Step 5
Analysing individuals &
job factors
Include new factor(s) in
column 2 - and review
the analysis from step 4
Step 6
Analyzing organizational &
management conditions
For each new row of column 3, search for organisational and management factors or conditions which may have facilitated, or may explain, why the previous events occurred. Use the classification scheme provided, to help identify such weaknesses in a systematic manner.
Display the results in column 4.
More Influencing factors were identified?
For each new row obtained in column 4, verify if any other influencing factor is necessary to help explain cause-effect relationships. The in-depth analysis only STOPS when no more cause-effect relationships can be established. The table is now complete.
YES
On a separate form, link the general management problems encountered to your own H&S Management System". Notice that a particular problem, such as, "inadequate management of contractors, may have to be linked to different elements of the system (e.g.: planning or implementation). If your company does not have a formal system, this is a good opportunity for prioritising needs in terms of implementing one. In WAIT, a standard OH&S Management System (the OHSAS 18001:1999) will be used as a model for establishing the links.
NO
Step 7
Linking findings to H&S
Management System
Based on the results of the analysis (basic + in-depth), make a list of recommendations and propose a plan of action. Whenever possible, include the following information: Recommended action / responsibility for executing / time expected for completion / rough estimation of cost / expected benefits/ priority.
Step 8
Making Recommendations
Re-analyse the case from a different point of view - this time searching for the existence of "positive influencing factors". To do this, re-analyse all information and, if necessary, re-interview people under this new perspective. In addition, highlight the benefits of "good practice" (if it was present) so that other co-workers can appreciate successful behaviours and realise their importance.
Step 9
Searching for positive
influencing factors
Any positive
Influencing factors were identified?
If positive influences or circumstances are found, establish whether they are merely a "random" coincidence, or if they are of a "controllable" nature - thus providing clues for new or better preventive / protective measures. If so, review step 8 and include them in the recommendations.
End No (or only random-not possible to control)
