Title: Decision analysis interviews on protective actions in Finland supported by the RODOS system

Authors: R.P. Hämäläinen, K. Sinkko, M. Lindstedt, M. Amman, A. Salo

Status: Finnish Centre for Radiation and Nuclear Safety STUK-A173, 2000, 57 pp.

Keywords: nuclear emergency management, multiattribute risk analysis, decision support, decision analysis interviews

This work was undertaken in order to study the utilisation of decision analysis interviews and of the rodos system when planning protective actions in the case of a nuclear accident. Six decision analysis interview meetings were organised. Interviewees were competent national safety authorities and technical level decision-makers, i.e., those who are responsible for drawing up advice or making presentations of matters to decision-makers responsible for the practical implementation of the actions. The theme of the meetings was to study how uncertainties could be included in the decision-making process and whether pre-structured generic attributes and value trees would help this process and save time. The approach was to present a generic value tree, a decision table and a selected information package at the beginning of the interviews. The interviewees then examined the suggested value tree in order to ensure that no important factors have been omitted and they made changes when necessary. Also, the decision table was examined and altered by some participants and some of them asked for further information on some issues. But all in all the selected approach allowed for more time and effort to be allocated to value trade-offs and elicitation of risk attitudes. All information was calculated with the support of the rodos system.

Predefined value trees were found to ensure that all relevant factors are considered. The participants also felt that rodos could provide the required information but, as in previous rodos exercises, they found it more problematic to use decision analysis methods when planning countermeasures in the early phase of a nuclear accident. Furthermore, it was again noted that understanding the actual meaning of ‘soft’ attributes, such as socio-psychological impacts, was not a straightforward issue. Consequently, the definition of attributes and training in advance would be beneficial. The incorporation of uncertainties also proved to be difficult and participants felt uneasy about probabilities.

The application of decision analysis in exercises has proven useful. Structuring the problem provided insight and many new issues could be analysed and discussed. Using a decision interview technique forces participants to think about the issue more carefully. Opinions seem to be more coherent and harmonised compared with earlier decision conferencing. Further meetings, however, need to be organised in order to deepen insight into the features of the decision-making process and to familiarise decision-makers with decision analysis techniques. More research is needed on how to implement decision conferencing or interviews in nuclear emergency management.