Root Cause Analysis types

The types of data required for incident reporting and root cause analysis systems are specified. Data Collection practices in the CPI are described, and a detailed specification of the types of information needed for causal analyses is provided. 

In the case of root cause analysis systems, more comprehensive evaluations of PIFs will normally be carried out as part of a full-scale human factors audit. This could make use of the types of comprehensive PIF evaluation methods described in Chapter 2 (see Section 2.7.7 and Figure 2.12). 

One approach is to mesh all investigation and root cause analysis activities under one management system for investigation. Such a system must address all four business drivers (1) process and personnel safety, (2) environmental responsibility, (3) quality, and (4) profitability. This approach works well since techniques used for data collection, causal factor analysis, and root cause analysis can be the same regardless of the type of incident. Many companies realize that root causes of a quality or reliability incident may become the root cause of a safety or process safety incident in the future and vice versa. 

Lagging indicators by themselves do not provide much explicit guidance to management as to what needs to be done to keep improving safety. The events themselves have to be analyzed using some type of root cause analysis. Also, lagging indicators tend to react quite slowly to system changes. 

Once the facts have been established and an understanding of the event has been established, a root cause analysis can be carried out in order to apply lessons learned to a broader set of circumstances. There are four types of root cause analysis ... 

Once the root cause analysis is complete, management will have to decide on the action to be taken. Three types of response can be considered (Wilson, 2013). Problems and corrective actions are given in Table 11.7. 

The fourth type of root cause analysis discussed at the start of this chapter is Systems Analysis. Two methods are described here to illustrate this approach Why Tree Analysis and Fault Tree Analysis. 

Tables 4.1 and 4.2 summarize the main types of studies of errors and adverse events, and their respective advantages and limitations. Thomas and Petersen s original source version has been separated into two separate tables and the content has been adjusted in particular, a section on case analysis has been added. Case analyses, usually referred to as root cause analysis or systems analysis, share some of the features of morbidity and mortality meetings, but are generally more focused and follow a particular method of analysis (Vincent, 2003) (Chapter 8).

The location-specific model assumes that different types of shocks occur at particular rates /Xj, for j = 1,..., i. To quantify the parameters of the location-specific model, the analyst needs to classify the failure events according to the type of failure mechanism underlying the event. This may be accomplished by performing root cause analysis. The data available and relevant to the specific k components can be written in the form... 

The natural consequence of the causality credo, combined with the Domino model, is the assumption that there is a basic or first cause, which can be found if the systematic search is continued until it can go no further. This is often called the root cause, although definitions differ. In the Domino model, the root cause was the ancestry and social environment, which led to undesirable traits of character. Since this was the fifth domino, it was not possible to continue the analysis any further. Other approaches, particular if they subscribe to some form of abstraction hierarchy, suffer from the same limitation. The type of analysis (which actually is a family of methods) that tries to find the root cause is unsurprisingly called Root Cause Analysis (RCS). 

The reporting and analysis of all types of incidents is a responsibility of employees and management. The Health and Safety Department can assist in managing safety in several ways. These professionals have special training and expertise in the area of root cause analysis. Their services are very valuable in identifying potential hazards and contributing causes, which you may not recognize. 

Let us consider one type of root cause analysis before leaving the topic of root cause analysis completely. Most workplaces are dynamic and subject to change. Thus, the use of change analysis is often an appropriate root cause analysis to apply. 

The overall objective of the system is to map from three types of numeric input process data into, generally, one to three root causes out of the possible 300. The data available include numeric information from sensors, product-specific numeric information such as molecular weight and area under peak from gel permeation chromatography (GPC) analysis of the product, and additional information from the GPC in the form of variances in expected shapes of traces. The plant also uses univariate statistical methods for data analysis of numeric product information. 

Another type of logic tree, the event tree, is an inductive technique. Event Tree Analysis (ETA) also provides a structured method to aid in understanding and determining the causes of an incident.(i) While the fault tree starts at the undesired event and works backward to identify root causes, the event tree looks forward to display the progression of various combinations of equipment failures and human errors that result in the incident graphically. 

Cause analysis is usually divided into three types (1) direct causes (2) contributing causes and (3) root causes. The direct cause of an incident is the immediate event or condition that caused the incident. Contributing causes are events or conditions that collectively increase the likelihood of the direct cause but that are not the main factors causing the incident. Root causes are the events or conditions underlying the root cause. Corrective measures for root causes will prevent the recurrence of the incident. In simple cases, root causes include materials or equipment deficiencies or their inappropriate handling. More complex examples are management failures, inadequate competencies, omissions, nonadherence to procedures, and inadequate communication. Root causes can be typically attributed to an action or lack of action by a group or individual. 

The Type A investigation of a sodium potassium (NaK) accident that occurred at the Y-12 plant on December 8, 1999, identified a lack of understanding of the hazard from NaK and its reactive byproducts as one of the root causes of the accident. The investigation found that personnel involved in planning the task, the safety documentation for the facility, the procedure for the task, and the procedures supporting hazard identification and analysis did not address the complete NaK hazard. The investigation also determined that detailed hazard identification data supported by accident analysis and appropriate control information was readily available. 

Before the interview starts, it must be made clear to the interviewee that the purpose of the investigation is simply to establish the facts associated with incident under investigation. It is not the intent of the interview to establish root causes or to conduct any other type of analysis. 

A top-down approach is used when management wants to improve overall reliability and/or does not know what the principal causes of problems may be. If a facility manager notes that production losses through unanticipated downtime are increasing, or maintenance costs for the whole facility have increased, he will probably call for a top-down analysis. He may also authorize this type of analysis when he suspects that there is a pervasive root cause problem, such as inadequate operator training, that is affecting many facility subsystems. 

Common cause failures (CCFs) are an important part of reliability analysis and engineers have been aware of these types of failures since the mid-seventies by Fleming (1974). Today there exist numerous models which explain this concept and attempt to model the impact such CCFs have on different systems. Even though this topic has been given much attention, it is still considered to be difficult and of a complex nature. CCFs are difficult to quantify correctly, i.e. it is difficult to know if a component fails due to a common root cause that affects several components, or if it fails because it is old and worn out. Usually, not much feedback data exist, so modelling this properly has proven difficult. In addition, different systems have different properties meaning that a model that may work for one system does not necessarily work for another. 

The causes of the specific types of accidents/incidents that have occurred within your workplace must be assessed. The approach you wish to use in the assessment and analysis process depends greatly on your familiarity with and the types of occurrences that have transpired in your workplace. Analyses covered in this book are root cause, causal factor, change, and barrier analysis. 

Initiating event In hazard analysis, an event could be the occurrence of a deviation which may lead to an accident. So, the initiating events are the causes for which there is the process deviation. The initiating events may be or may not be the most basic underlying root-causes, but are the results of the root causes. According to CCPS there are three types of initiating events or causes ... 

All too often, the description of a failure is limited to the most obvious cause or a simple statement that the component has failed . Without a proper identification of the root cause, it is difficult to select the most efficient defense measure. For example, in the review of reported failures, we see that failed detectors are often just replaced by a similar type without further analysis. We also see that valves being stuck during a functional test may be lubricated and then operated successfully. Lubrication has a short term effect on the valve performance, and actual root cause is seldom revealed. Such strategies are often insufficient for the purpose of selecting the most effective defense measures. 

We will later apply the accident-analysis framework in a review of different types of methods used in the collection and analysis of data of accident risks. We will start at the output side of the model by reviewing the different types of classification systems used to document the consequences of accidents and different measures of loss. We will then continue by looking into the classification systems used to document incidents and deviations. Finally, we will review the different classification systems for contributing factors and root causes. Our aims will be twofold first, to be complete, i.e. by presenting all alternative means of measuring and classification, and second, to give specific advice on the preferred method. The reader will find recommended alternatives in shaded tables and checklists. 

---