Accident prevention. Presentation of a model placing emphasis on human, structural and cultural factors
Introduction
Injuries from accidents are a threat to health in every country. In 1990 they were responsible for 6–7% of world mortality (Murray and Lopez, 1997). In Norway, the rate of disability pensions issued due to accidents increased 4% annually in the years 1992–1997 (Lund and Bjerkedal, 2001). Various preventive measures have been used to reduce this health burden. One common set of measures has been, and remains, the use of media campaigns, information leaflets, and films to inform the population about certain hazards and how to avoid or minimise the risks associated with them. The underlying assumption is that increased knowledge will modify attitudes with regard to hazards and prevention. As a consequence, behaviour will change and accidents and injuries will be avoided (KAP-model: Knowledge→Attitude→Practice {Behaviour}). However, conclusions from reviews of research on the effects of attitude change interventions on behaviour and accidents/injuries are rather pessimistic (Wilde, 1993, OECD, 1994, Aaro and Rise, 1996).
Several explanations may be proposed for the problem of documenting behavioural effects of attitude change interventions in the field of accident prevention. Existing knowledge on how to influence attitudes may have been neglected or underutilised (Ajzen, 1988, Zimbardo and Leippe, 1991). The importance of alternative and competing factors and influences may have been underestimated (Kok et al., 1996), and the complexity of the processes influencing accident-related behaviours may have been poorly understood.
Furthermore, research has indicated that the association between attitudes and behaviour often is weak and non-significant (Wicker, 1969). When correlations are found, several competing underlying causal mechanisms may also be suggested. Attitudes may sometimes be influenced by behaviour rather than vice versa (Festinger, 1957, Bem, 1972). Bandura (1969) proposed that consistency between indicators of attitudes and behaviour might represent correlated co-effects rather than outcomes of a process in which modification of one factor produces changes in the other.
Over the last 30 years, theories and models used in health behaviour research have gradually become more sophisticated and better able to predict behaviour and inform the planning of health education and health promotion interventions (Stroebe and Stroebe, 1995). Several models are based on the assumption that behavioural intention is the best predictor of behaviour. Behavioural intentions are regarded as products not only of attitudes, but also of social influences (including social norms) and self-efficacy (or perceived behavioural control). The relationship between intentions and behaviour depends on whether or not the person actually possesses the skills needed to perform such behaviour. Attitudes are seen as influenced by cognitive processes, primarily outcome expectations (Ajzen and Fishbein, 1980, Ajzen, 1988, Kok et al., 1996). Some researchers have highlighted the importance of factors such as cues to action (Janz and Becker, 1984), self standards (Bandura, 1986), emotions (Røysamb, 1997), and heuristic processing (Eagly and Chaiken, 1993). Although there is considerable consensus regarding which factors are the most important to take into account when predicting health behaviour, there is, however, yet no consensus on the underlying causal model (Fishbein et al., 1991).
The dominant theories in risk behaviour research are borrowed from social psychology, and tend to emphasise cognitive processes as well as immediate social influences on the individual. One of the most obvious shortcomings of such theories and models is their neglect of factors beyond this domain. Although economic factors, legislation, enforcement of legislation, and aspects of our cultural, organisational, and physical surroundings are widely accepted as crucial in health promotion and health behaviour change, few models have incorporated them.
The concept of ‘safety culture’ has received considerable attention and represents an important widening of theoretical perspectives in research on injury prevention. Moghaddam (1998, p. 18) links culture to social norms, defining culture as “a normative system that prescribes how one should behave in given contexts”. Through socialisation by families, peers, and the wider society, individuals become skilled in recognising, influencing, and following social norms and rules. Groups can establish normative systems with distinct characteristics, and in this way create greater cultural and behavioural diversity in society (Moghaddam, 1998). According to Reason (2000, p. 11), “culture transcends the psychology of any single person”. Safety behaviour (which here includes risk behaviour) is claimed to depend on cultural factors, and to vary systematically across cultures.
The concept of safety culture is an outgrowth of ‘safety climate’, which is in turn a derivative of ‘organisational climate’ first introduced into the industrial psychology literature in the early 1970s. The concept of a ‘safety climate’ was introduced by Zohar (1980a) as a derivative of the broader climate concept. It constitutes a nexus of safety hazards, several aspects of management control, and employees’ perceptions of its effectiveness (Williamson et al., 1997, Flin et al., 2000).
After the Chernobyl catastrophe, increased attention was paid to safety climate or safety culture in workplaces with advanced technology, complex systems, and potential for catastrophes in order to understand the influence of cultural factors on safety behaviour (Pidgeon, 1991, Reason, 1997, Ostvik et al., 1997). A literature review (Guldenmund, 2000) has shown that safety culture and safety climate remain ill-defined concepts. He proposes the following definition (p. 251): “Safety culture is defined as those aspects of the organisational culture which will impact on attitudes and behaviour related to increasing or decreasing risk.” According to Reason (1997, p. 192), Uttal's (1983) definition of organisational culture most closely captures its essence: “Shared values (what is important) and beliefs (how things work) that interact with an organisation's structures and control systems to produce behavioural norms (the way we do things around here)”. A question to ask is if the concept ‘safety culture’, which now is used in organisations with tightly coupled systems, such as nuclear plants (Perrow, 1984), may also be used in loosely coupled systems such as local communities.
Various factors and sub-components have been proposed for identifying and measuring safety culture (Reason, 1997, Lee, 1998). The criticism has been made that safety culture is reduced to a combination of administrative procedures and individual attitudes to safety (Pidgeon, 1998), and that considerations of politics and power are absent. Also, there is some doubt of whether “culture can be ‘measured’ at all using quantitative psychometric methodologies such as questionnaires or surveys” (Pidgeon, 1998, p. 204), or as the only measurement tool (Cooper, 2000). Hale (2000) questions whether there is such a thing as a ‘safety culture’, or whether it might not be better to talk about ‘cultural influences on safety’.
The purpose of the present study is to propose a model that takes into account this wide range of human, structural and cultural factors. It includes a consideration of social norms and culture, concepts that until now have been insufficiently dealt with by theoretical models in this field. By reviewing the literature on accident prevention, the correspondence of the model with existing evidence will be examined.
Section snippets
A model for accident prevention placing emphasis on human, structural and cultural factors
The measures used for prevention of accidents might be divided into three main categories:
- 1.
Attitude modification: attitudes are changed by means of persuasive messages in mass media campaigns, leaflets, booklets, films, posters, or direct mail. Also included in this category are one-way counselling schemes, such as counselling on car safety to mothers of new-born children.
- 2.
Behaviour modification: behaviour is changed through more direct approaches, without assuming that attitudes have an
Method
Relevant literature was identified through psychological and medical databases (Psychlit from 1966 and Medline from 1967), searching journals, scanning reference lists, and consultations with colleagues in Norway, England, and the USA. When searching on databases, more general terms such as ‘accidents’, ‘injuries’, and ‘safety’ were combined with ‘prevention’, ‘control’, and ‘campaign’. It turned out, however, that more specific searches were also needed, and in these searches various
Information measures such as mass media campaigns, leaflets, booklets, films, posters, or direct mail
Studies on attitude modification interventions using information measures alone are summarised in Table 1. Most of them do not prove any effect on behaviour or on the incidence of accidents and injuries (Fleischer, 1973, Robertson et al., 1974, Anderson, 1978, MacKay and Rothman, 1982, Saarela, 1989, Damoiseaux et al., 1991, Robertson, 1994, Delhomme et al., 1999).
A small proportion of the information-based intervention projects that use communication devices such as leaflets and films turned
Variations in safety dimensions across cultures
Studies of attitudes among more than 13,000 pilots of 25 airlines in 16 countries (Helmreich et al., 1996) have revealed considerable national differences in norms and attitudes. Some of them might be related to safety behaviour, including, for example, whether or not written procedures are required, and whether or not organisation rules are complied with. These differences might refer to cultural differences with regard to, for example, collectivism versus individualism, or attitudes of
New roles for attitude-forming initiatives
The main impression from the overview given above is that the more we rely on attitude change through one-way information and education campaigns only, the less likely we are to succeed in changing injury-related behaviour. The more we use behaviour modification techniques, the more we allow for interpersonal interaction to take place, and the more informational and educational measures are combined with other measures (such as restrictions, organisational change, changes in the physical
Conclusion
In the present article we have focussed on connections between the elements of Fig. 1, and we have examined various pathways from preventive action to reduction in accidents and injuries. Further reviews and findings, especially from databases outside of the psychological and medical field, and from publications in other languages than English, might extend or modify parts of the proposed model. However, based on findings from this literature review, some of the paths in the model appear to be
Acknowledgements
The authors acknowledge with gratitude the contribution from Tor Bjerkedal and Rune Elvik during the preparation of this manuscript. Also, the help from members in the reference group of a previous project on attitude change: Terje Assum, Gunnar Breivik, Jan Hovden, Reidar Ommundsen, Kjersti Simonsen and Terje Sten, where a first draft of the model was developed. An especially helpful contribution was given by Jostein Rise, who participated in the literature review and assessment of attitude
References (189)
The effectiveness of traffic safety material in influencing the driving performance of the general driving population
Accident Analysis and Prevention
(1978)The crash helmet legislation in Nigeriaa before-and-after study
Accident Analysis and Prevention
(1980)Self perception theory. An alternative interpretation of cognitive dissonance phenomena
- et al.
The community safety approach in Falun, Sweden—is it possible to characterise the most effective prevention endeavours and how long-lasting are the results?
Accident Analysis and Prevention
(2000) - et al.
The effectiveness of antilock braking systems in reducing accidents in Great Britain
Accident Analysis and Prevention
(2002) - et al.
The effect of anticipatory guidance on mothers’ self-efficacy and behavioral intentions to prevent burns caused by hot tap water
Patient Education and Counselling
(1993) - et al.
Helmets, injuries and cultural definitionsmotorcycle injury in urban Indonesia
Accident Analysis and Prevention
(1996) - et al.
The effect of variable-rate cash incentives on safety belt use
Journal of Safety Research
(1986) - et al.
An evaluation of three corporate strategies for safety belt use promotion
Accident Analysis and Prevention
(1986) - et al.
Increasing child safety-seat use and proper use among toddlersevaluation of an enforcement and education program
Accident Analysis and Prevention
(1994)