Causal Factors

In addition to these formal studies of human error in the CPI, almost all the major accident investigations in recent years, for example, Texas City, Piper Alpha, Phillips 66, Feyzin, Mexico City, have shown human error as a significant causal factors in design, operations, maintenance or the management of the process. Figures 4.4-1 and 4.4-2 show the effects of human error on nuclear plant operation. 

The intention of this section is to provide a selection of case studies of varying complexity and from different stages of chemical process plant operation. The purpose of these case studies is to indicate that human error occurs at all stages of plant operation, and to emphasize the need to get at root causes. The case studies are grouped under a number of headings to illustrate some of the commonly recurring causal factors. Many of these factors will be discussed in later chapters. 

A range of techniques is described for analyzing the structure of incidents and the causal factors involved. 

There will be strong emphasis on the collection of data on possible causal factors that could have contributed to an accident. The specific data that are collected may be based on an error model such as that shown in Figure 6.2. However, this model will usually be modified depending upon the extent to which it fits the data collected over a period of time. The systems approach is therefore dynamic rather than static. 

It should be emphasized that it is usually necessary to develop the data collection specification on an incremental basis and to utilize feedback from the system to modify the initial model relating causal factors to error types. This dynamic approach provides the best answer to the problem that no predefined error model will be applicable to every situation. 

In the following sections, a number of methodologies for accident analysis will be presented. These focus primarily on the sequence and structure of an accident and the external causal factors involved. These methods provide valuable information for the interpretation process and the development of remedial measures. Because most of these techniques include a procedure for delineating the structure of an incident, and are therefore likely to be time consuming, they will usually be applied in the root cause analysis of incidents with severe consequences. 

An extension of the tree of causes, called variation diagrams (Leplat and Rasmussen, 1984) was developed to answer some of these criticisms. In this method, the Rasmussen stepladder model of human error (see Chapter 2) is applied to analyze causal factors at each node of the tree. A detailed example of the use of this technique is provided in Chapter 7 (Case Study 1). 

Analyze Barriers and Potential Human Performance Difficulties During this phase of the analysis process, the barriers that have been breached by the accident are identified. TTiese barriers could include existing safety systems, guards, containment, etc. This analysis is called barrier analysis. The causal factors from SORTM are also applied in more detail. 

Using the ECFC representation of the incident, a series of defailed questions which address specific causal factors (e.g., poor procedures), are applied to evaluate direct and indirect root causes. These detailed questions are contained in a series of HPIP modules. 

With regard to evaluating these factors, it is recommended that structured checklists be used, such as those provided by the HFAM method described in Chapter 2. These checklists provide an explicit link between the direct causal factors and management policies. Figure 2.12 shows how these checklists could be used to investigate possible procedures deficiencies, and the policies that led to the deficiencies, as part of the incident investigation. Similar checklists can be used to investigate possible culture problems (e.g., inappropriate trade-offs between safety and production) that could have been implicated in an accident. 

In the second case study, variation tree analysis and the events and causal factors chart/root cause analysis method are applied to an incident in a resin plant. This case study illustrates the application of retrospective analysis methods to identify the imderlying causes of an incident and to prescribe remedial actions. This approach is one of the recommended strategies in the overall error management framework described in Chapter 8. 

The events and causal factors chart for this incident is shown in Figure 7.9. The primary sequence of events is shown horizontally in bold boxes. Secondary events are shown in the other boxes, and conditions are in ovals. From the diagram three causal factors were identified and carried forward to the Root Cause Coding to establish the root causes of the causal factors. 

Causal Factor 1 Operator A Connects Pump to 21A Pipe Not 12A Pipe Root cause coding identified the following root causes ... 

Causal Factor 2 Operator A Failed to Close 21A Blender Valve... 

The human factors audit was part of a hazard analysis which was used to recommend the degree of automation required in blowdown situations. The results of the human factors audit were mainly in terms of major errors which could affect blowdown success likelihood, and causal factors such as procedures, training, control room design, team communications, and aspects of hardware equipment. The major emphasis of the study was on improving the human interaction with the blowdown system, whether manual or automatic. Two specific platform scenarios were investigated. One was a significant gas release in the molecular sieve module (MSM) on a relatively new platform, and the other a release in the separator module (SM) on an older generation platform. 

Contemporary forest declines were initiated about 1950-1960, virtually simultaneously throughout the industrial world at the same time as damage to aquatic systems and structures became apparent. A broad array of natural and anthropogenic stresses have been identified as components of a complex web of primary causal factors that vary in time and space, interact among each other, affect various plant growth and development systems and may result in the death of trees in mountainous ecosystems. As these ecosystems decline, the alterations in forest ecology, independent of the initial causal complex, become themselves additional stress factor complexes leading to further alterations. 

Obviously, one looks for causes. That declines in one or another species have natural factor etiologies is unequivocal. The demise of American elms and of the chestnut were due to natural factors. Insect infestations, bacterial and fungal diseases, hurricanes, floods, freezes, droughts and many other stresses can cause extensive tree death (5). But in such declines typically only a single species is affected or climatic events caused decline in a delimited area. In almost all declines caused by natural events, the causal factors can be identified we know their precise etiologies. Natural events are always part of the natural environment and must be factored in when evaluating forest declines (Table I). 

The initial causes of forest decline - natural and anthropogenic - have resulted in a forest decline syndrome that is, per se, a new series of causal factors whose consequences are themselves new causes for ecosystem alteration. 

Although seen only occasional during the first half of this century, winter injury of first-year red spruce needles has become an annual event in the coniferous montane forest area, resulting in the formation of red-brown first year needles that subsequently desiccate and are shed 20, 21). The loss of foliage reduces photosynthesis and the obligatory accumulation of carbohydrate in the twigs and root systems. There is some evidence that this phenomenon involves both natural and anthropogenic causal factors. 

National initiatives in North America and Europe are designed to reduce pollution emissions from both stationary and mobile sources. Independently of whether they succeed in reducing pollutant loadings, the available evidence indicates that alterations in affected forests will continue. Obviously, no one knows what affected forests will be like in 50 years. There is little doubt that they will be different. And the sooner the anthropogenic causal factors - all of them - are reduced qualitatively and quantitatively, the better are the chances of retaining or regenerating forests that will have meaning and value for those who will want to use them. 

Retrospective Study—A type of cohort study based on a group of persons known to have been exposed at some time in the past. Data are collected from routinely recorded events, up to the time the study is undertaken. Retrospective studies are limited to causal factors that can be ascertained from existing records and/or examining survivors of the cohort. 

If excessive noradrenergic transmission is a causal factor in anxiety, then it would be predicted that a lesion of central noradrenergic neurons would have an anti-anxiety effect in behavioural models of this condition. Unfortunately, the behavioural effects of such lesions are notoriously inconsistent and there are many reports of negative findings (e.g. Salmon, Tsaltas and Gray 1989). One study has even shown that a lesion of central noradrenergic neurons, induced by the selective neurotoxin, DSP-4, abolishes the anti-anxiety effects of tricyclic antidepressants and MAO inhibitors, but not those of the benzodiazepine, alprazolam, or the barbiturate, phenobarbitone (Fontana,... 

---