Safety | Accident investigation
Lessons learned1 November 2012
Operational experience feedback can be a valuable tool for preventing recurrence of similar events at nuclear plants. However, human and organizational factors can significantly hamper accident investigations. By Stanislovas Ziedelis and Benoit Zerger
Nuclear safety management is founded to a considerable extent on learning from experience. However, there is growing consensus among experts that safety improvement based on operational experience may have already reached their limits in high-risk industries, since no decreasing trend in number of events has been observed . This is a problem of fundamental importance for nuclear industry. It is caused by insufficient effectiveness of operational experience feedback (OEF) systems to prevent or decrease the frequency of recurring events.
The main insights about the effectiveness of OEF process have been derived from several topical studies performed by the European Clearinghouse on Operational Experience [2-5]. These studies involved extracting, organising and analysing information from event reports stored in various incident-reporting databases, including IRS of IAEA/NEA, LER (US NRC), SAPIDE of IRSN (France) and VERA of GRS (Germany). The common methodology for the studies included identifying, highlighting and examining the circumstances, precursors, direct/root causes and safety significance of said events, as well as the most important corrective actions, lessons learned and recommendations . Repetitive shortcomings were observed during analysis of event reports:
- Real root causes of events are often not properly identified
- Issues related to management deficiencies and human & organizational factors (HOF) are typically not well-addressed
- Even if identified, HOF issues are typically understood as human performance errors at the individual worker level.
In order to reveal the latent causes of these shortcomings, the European Clearinghouse on Operational Experience conducted a survey-based study in 2010-2011. The study provided an in-depth analysis of the event investigation practices used in nuclear power plants and the constraints limiting the effectiveness of the operational experience feedback (OEF) process . The survey included 47 questions, aiming to unveil any human/organizational factors preventing the implementation of effective corrective actions. Data from nuclear industry and regulatory bodies of 12 European countries were collected and analysed.
The main weaknesses of current operational experience feedback systems seem to be related to procedural problems, human and organizational factors (HOF), safety culture and/or knowledge management.
Procedural weaknesses include a lack of standards for event investigation methods/ tools, an absence of agreed terms and definitions, as well as undefined thresholds for determining specific levels of investigation.
Barriers in the event investigation process originating from human and organizational factors (HOF) and inadequate quality and/or safety culture of personnel at all levels, including management, appear to be most serious (Figure 1).
The first barrier is the event review/screening process, where the level of investigation is determined. Since thresholds for initiating an event investigation at various levels (troubleshooting, apparent cause investigation or root cause analysis) are not standardised and frequently not even clearly defined, some events important for safety or containing valuable OE information are not adequately analysed.
The second potential barrier lies in the preparation of the event investigation. An event investigation process should be established and implemented within the management system. Investigation teams should be adequately staffed, protected from undue pressures and equipped with appropriate resources. Failure to fulfill these requirements may lead to the appointment of a weak investigation team and/or creation of unfavourable conditions for investigations.
The third potential barrier that could hamper the outcome of an investigation lies in its execution. Here are numerous potential pitfalls, sometimes aggravated by insufficient support from management, including limited access to information, exposure of the investigation team to external or internal pressures, and/or exposure to a punitive culture. A key factor is that persons and organizations may be unwilling to be self-critical because of potential repercussions (internal and/or external). The management may also influence the investigation in a particular direction or restrict the investigation to identifying individual failures of shop-floor workers or junior engineering personnel. These influences are especially evident when the underlying causes for the incidents are ultimately supervisory and/or managerial deficiencies; managers are reluctant to dig deeply enough into incidents in case the outcome of an analysis could reflect badly on the arrangements that they have put in place. In doing this, the tendency is generally to treat the symptom rather than the underlying fundamental problem. This finding was frequently noted among other analyses of event reports [2-5]. Events with a clear indication of HOF causes were still analysed and described as almost entirely caused by technical factors. The role of the senior management and culturally-related organizational factors were rarely considered. HOF issues were typically understood as human performance errors at the individual worker level.
The last serious barrier to a successful event investigation lies at the stage of generating, approving and implementing the corrective actions. Completion of a good root cause analysis does not in itself solve the problem. Investigation teams could make recommendations that may not work, or they could make good recommendations that do not get approved by management. Good recommendations could be approved, but not implemented. Indeed, at this stage, improper prioritisation of corrective actions could hinder latent root causes from being fixed.
Accident investigations may also suffer due to the poor management of the vast amounts of information related to event causal factors and latent root causes. Typically, raw information is not properly screened, classified, catalogued, and formatted in a user-friendly manner because of a lack of resources or motivation. This means that the relevant data is difficult to find, retrieve and use. Another important factor that impedes the utilisation of event-related information is the restrictions placed on potential users of numerous databases.
The results of the survey confirmed initial assumptions about the origins of repetitive deficiencies in nuclear event reports. For example, event investigation teams are typically composed of technical specialists from the plant departments concerned, professional event investigators and representatives of management. Psychologists, human factor specialists and external experts or consultants are rarely included in the investigation teams. Such practices lead to the possibility of misdirecting the investigation from deeply-rooted issues related to management, human and organizational factors to traditional technical/procedural issues. The lack of any external consultants in the event investigation teams prevents discussion of the independence of event investigation and creates favourable conditions for purposeful management impact.
The majority of event investigators declared having received a formal training in the event investigation/root cause analysis methods. However, the prevailing form of training was found to be participation in short-term workshops with a duration of between two and five days. Such training could be scarcely considered adequate. The education system for event investigators could be improved by organising adequate training, elaborating simpler manuals and guidelines, and providing more practical examples.
The final effectiveness of the operational experience feedback process mostly depends on the safety culture of the organization’s personnel, and especially its senior management. Questionable management decisions concerning event investigations may be driven by focus on costs or schedule, fear of blame, career considerations, arrogance, ignorance, overconfidence, pervasive belief in entitlement and/or autocratic decision-making. The potential improvements of OEF should be oriented to eliminating the potential bottlenecks and emerging pitfalls in the root cause analysis-based nuclear events investigation process. Upgrades in safety culture, re-focusing the investigations from technical causal factors to deeper and latent root causes, such as human-, organizational- and management-related factors contributing to the event should play the key role. Further methodological refinements of root cause analysis tools and techniques, increasing the competencies, capabilities and level of independence of event investigation teams, and ensuring a positive approach, adequate support and impartiality of management could also facilitate improving the quality of the event investigations.
Stanislovas Ziedelis (Stanislovas.Ziedelis@ec.europa.eu) and Benoit Zerger, European Clearinghouse on Operational Experience Feedback for Nuclear Power Plants, Institute for Energy and Transport, Joint Research Centre, European Commission, P.O. Box 2, 1755 ZG Petten, The Netherlands.
This article was originally published in the October 2012 issue of Nuclear Engineering International
 Dien Y., Dechy N., Guillaume E. Accident investigation: From searching direct causes to finding in-depth causes â€“ Problem of analysis or/and of analyst? Safety Science, 2012, v. 50, p. 1398-1407.