Identifying accident patterns using the FAC and HAC: their application to accidents at the engine workshops of an automobile and truck factory

Abstract

This paper raises the issue of the optimum utilization of information from accident registers for the achievement of one main goal: the facilitation of safety planning. It proposes a non-traditional way of analyzing accident data that provides a condensed and comprehensive overview of the most common accident problems encountered by a target group. A description of typical accident patterns is obtained by the simultaneous treatment of the entire body of information compiled. Data analysis is performed on the basis of the following question: What are the most typical injury events and in terms of which factors may their characteristics be explained? Two statistical methods complement one another: the Factorial Analysis of Correspondence (FAC) and the Hierarchical Ascendant Classification (HAC). The target group comprised blue-collar workers at the engine workshops of a large automobile and truck factory in Sweden. As well as the identification and characterization of the main accident patterns, the study was also designed to estimate the risk of accidents for six occupational groups and to establish whether levels of risk were similar between workshops for each occupational group. Statistics on accident characteristics were compiled on the basis of the 178 accident declaration forms filled in during 1986 and 1987. These forms were filled in when an injured worker had been away from work at least one day after an accident. Data on a total of 35 variables were collected. These variables covered seven themes: (1) the time of the accident; (2) its location; (3) the injured worker; (4) the activity performed by the injured worker; (5) the machinery and/or tool involved, if any; (6) the accident event; (7) the injury. Seven typical accident classes (patterns) were extensively portrayed. It was observed that accidents of any class might occur in either workshop, but that their distribution varies according to the occupational group, and the tasks performed by the injured workers. It was also shown that accident frequency varies considerably from one occupational group to another; and for some groups, from workshop to workshop. It is emphasized that the multivariate analysis of the data available in accident registers both broadens and enriches understanding of the accident problems encountered by a target group. It also represents an important supplement to traditional accident statistics. This non-traditional way of defining accident problems may, through the treatment of one accident pattern (class) at a time, provide assistance in understanding and evaluating the need for preventive measures. Further accident prevention and, in a broader sense, safety improvements, may be made considerably easier on the basis of the type of reference material obtained.