Tier Analysis

Tier diagramming is a root cause analysis technique that focuses on those levels of management that have the responsibility to correct potential problems. It explicitly forces investigators to consider organisational factors as the initial root causes of many failures.

Each row in one of these diagrams refers to a different level of management within an organisation. The rows are intended to represent levels of organisational responsibility that range from the operator up to senior management. The columns in a tier diagram list the causal factors that are derived from the Causal Factor analysis together with any higher-level Root Causes that may or may not be identified as shown in Table 1. It is important to note, however, that Table 1 is a generic template that must be tailored to reflect the organisations that are involved in a particular incident.

Each causal factor is assigned to a tier of management responsibility. This is intended to help identify any common links between causal factors that relate to particular levels in an organisation. For instance, a failure in supervision would be exposed by a number of causal factors that cluster around this level in the tier diagram. This is intended to offer a number of benefits to any incident investigation

• It helps to focus any root cause analysis on the deeper organisational causes of failure.
• The tabular format helps to structure an investigation around concepts, or groups, that have a clear organisational meaning for those involved in an incident. This is important because many incident reports often talk in vague terms about a 'failure in safety culture' without grounding these observations in the activities of particular organisations and groups.
• Responsibility is explicitly assigned for each root cause and causal factor. These judgments provide a focus for subsequent discussion and can, ultimately, help to form the recommendations for future practice.

Different tier diagrams are drawn up for each of the organisations that is involved in an incident. Tier diagramming, typically, begins with the organisation that is most closely involved in the incident.

Stages of Developing a Tier Diagram

1. Develop the tier diagram. Create a tier diagram that reflects the management structure of the organisation being considered.
2. Identify direct causes. Examine the cause-context summaries to identify any catalytic events that cannot be directly associated with operators or management activities. Enter these along the direct cause row. Repeat this process for any conditions that are associated with these causal events in an ECF chart. Initially, this tier might contain events that describe the failure of process components or problems due to the contamination of raw materials. As analysis progresses, however, it is likely that most of these direct causes will be associated with other tiers in the diagram. For instance, component failures may be due to a managerial failure to ensure an adequate maintenance regime. Similarly, the conatmination of raw materials can be associated with acquisitions and screen policies.
3. Identify worker actions. For each causal factor in the cause-context summary, ask whether or not they stemmed directly from 'worker actions'. Develop guidelines to direct this stage of analysis. For instance, ask whether or not the worker's knowledge, skills and abilities were adequate to perform the job safely. Also ask whether the worker understood the work that was to be performed. As with direct causes, these actions often raise questions about the performance of other groups in a tier diagram. The worker's lack of understanding may be due to an inadequate training regime. Investigators must, therefore, ask whether or not the worker was solely responsible for the causal factor. If the answer is no then investigators must move the event to a higher tier in the diagram.
4. Analyse remaining tiers. The analysis progresses in a similar fashion for each tier. The intention is to place each causal factor as high up the diagram as possible. Ultimately, all incidents can be associated with regulatory problems or a failure in oversight. It is important,however, to balance this observation about ultimate responsibility against the need to identify those levels in an organisation that are most directly responsible for certain causal factors. This is most often done by developing analystical guidelines. These guidelines help investigators to assess whether or not a causal factor can be associated with a particular tier in the diagram. They are, in turn, typically derived from the safety cases that justify the operation of an application process. For instance, if middle management has an identified responsibility to ensure the operation of an incident reporting system then it is possible to place any causal factor that relates to the failure of such a system at this level in a tier diagram.
5. Identify links. After all of the causal factors and associated conditions have been entered into a tier diagram, investigators can begin to look for common factors. The success of this activity depends upon the skill and expertise of the investigator. Investigators can use different colors or symbols to denote those causes that are considered to be linked.
6. Identify root causes. Compare each of the causal factors in the tier diagram against the definition of a root cause.
   A root cause is distinguished as follows:
   • If A and B are states (conditions) or events, then A is a necessary causal factor of B if and only if it is the case that if A had not occurred then B would not have occurred either.
   • If A is corrected, this would prevent the recurrence of the same or similar incidents; it also solves deeper line management, oversight and management system deficiences that could cause or contribute to future mishaps
   • Root causes must address a class of deficiencies, rather than single problems or faults.
   If a causal factor meets these criteria then an additional entry can be made to denote this finding in the Root Cause column of the tier diagram. Investigators must, therefore, compose a root cause 'statement' to summarise each of the causal factors groupings that were identified in the previous stage of analysis.
   Root cause analysis can reveal events and conditions that were not represented on ECF charts, or cause-context summaries. These must be added to ensure consistency between these various products of a root cause analysis. It should also be noted that one tier diagram may provide input for another. For instance, if the upper management of a contractor was responsible for a particular root cause then the regulator and supervisory organisation may share responsibility for that particular root cause if there is a deficiency in the directives given by those organisations.

Difficulties in Developing a Tier Diagram

• Breakdown of management structures. The different teams and individuals who are associated with different levels in a tier diagram may change as organisations attempt to adapt to the pressures that are created by many projects. Often staff are unable to reply to questions such as 'who is in charge?' or 'who is the project manager?'. Possible solutions:
  • Develop a number of tier diagrams to represent each change
  • Use a relatively abstract classification of organisational structures, and then provide more detailed to support the interpretation of those categories at particular stages of the incident.
• Restricted access to information. It may not be possible for investigators from one organisation to gain access to detailed information about the management of another organisation. Investigators may be forced to use more generic tiers.
• Interpretation depends on the investigator. This illustrates the subjective nature of tier analysis, and the output depends on the skill, expertise and viewpoint of the investigator. ECF charts, cause-context summaries, tier analysis are all artifacts that help to document the path towards a causal analysis. They do not replace the skill and expertise of the investigators nor do they 'automate' key stages of the analysis.
• Undue emphasis on higher levels within an organisation. Tier analysis tends to associate root causes with the higher levels of management. This is a natural consequence of the iterative process that is used to analyse each causal factor; the intention is to place each causal factor as high up the diagram as possible. This is an important strength of the technique. However, the investigator's attention is drawn away from individual instances of operator error, and undue emphasis may be placed on individuals at higher levels. This is inappropriate if operational responsibility is devolved to lower levels within the management structure.
• Tier analysis is strongly influenced by the use of ECF charts during the initial stages of an investigation. This technique encourages analyst to focus on particular events rather than on the organisational factors that created the ocnditions for an incident. This initial focus can be broadened by barrier and change analysis. Both these techniques help to ensure that causal factor analysis looks beyond the immediate events that contribute to an incident.

Practice drawing Tier Diagrams

• Using the incident of a fall from a wheelchair, draw Tier Diagrams for the following: RCA Case Practice
  • the new nurse assisting the wheelchair patient to the smoking area, then leaving him there
  • the broken wheelchair awaiting repair being left in the ward
  • using wax on the footpath outside the hospital precinct
    • within the hospital hierarchy for managing outsourcing
    • within the cleaning company
• Apply your findings to your ECF diagram(s) for the incident