Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Joumal of Safety Research, Vol. 27, No. 2, pp. 61-72.1996 Copyright &)1996 National Safety Council and Elsevier Science Ltd printed in the USA. All rights nxexved 0022-4375/96 $15.00 + .OO Pergamon PII SOO22-4375(%)00007-2 Theoretical Models of Health Behavior and Workplace Self-Protective Behavior David M. DeJoy This paper provides a critical review of the applicability of theoretical models of health behavior to workplace self-protective behavior. Valueexpectancy, environmental/contextual, and behavior change models are reviewed. On this basis, an integrative framework is proposed that conceptualizes self-protective behavior as consisting of four stages or phases: hazard appraisal, decision making, initiation, and adherence. In addition, five general constructs are identified as being of either primary or secondary importance at each stage: threat-related beliefs, response efficacy, self-efficacy, facilitating conditions, and safety climate. The proposed framework highlights the need to target interventions to each of the four stages. Particular emphasis is also assigned to environmental or situational factors in enabling and reinforcing self-protective behavior in the workplace. INTRODUCTION With very few exceptions, research on workplace self-protective behavior has been piecemeal and atheoretical. Studies have generally taken one of three approaches. First, a number of studies have examined various employee characteristics and their relationships to safety performance and injury experience. Characteristics studied have included: hazard and safety-related attitudes and beliefs (e.g., Cox & Cox, 1991; Dedobbeleer & German, 1987; Leather, 1988; Walters & Haines, 1988); personality dimensions and risktaking tendencies (e.g., Jones & Wuebker, 1985; Landeweerd, Urlings, DeJong, Nijhuis, & Bouter, 1990); subjective risk assessments (e.g., Edwards David M. DeJoy is Professor and Head of the Department of Health Promotion and Behavior at the University of Georgia. He received his PhD from Penn State University. Summer 1996Nolume 27hVumber 2 & Hahn, 1980; Goldberg, Dar-el, & Rubin, 1991; Howarth, 1987; Zimolong, 1985); and job demands and other stressors (e.g., Cooper & Sutherland, 1987; Levenson, Hirschfeld, & Hirschfeld, 1980; Murphy, 1984; Smith, Colligan, Fmckt, & Tastn, 1982). A second group of studies has attempted to modify the safety-related behaviors of workers through the use of contingent reinforcement or operant-based approaches. Reinforcers have included information feedback, goal setting, social mcognition, and praise, as well as more conventional rewards and incentives. Reinforcement protocols have been used to improve the use of protective eyewear (Smith, Anger, & Uslan, 1978) and hearing protectors (Zohar, Cohen, & Azar, 1980), and to increase compliance with safety regulations (Rhoton, 1980). Other studies within this group have been somewhat broader in scope and have tried to influence multiple safety-related behaviors 61 and overall safety performance (e.g., Komaki, Barwick, & Scott, 1978; Sulzer-Azaroff & DeSantamaria. 1980). Supervisory behavior has also been examined using this general approach (Komaki, Zlotnick, 8 Jensen, 1986; Mattila, Hyttinen, & Rantaneu 1994). The third group of studies has focused on the organizational and environmental correlates of good safety performance. Most of these studies have sought to identify the programmatic features of effective safety programs (e.g., Cohen, Smith, & Cohen, 1975; Cohen & Cleveland, 1983; Fiedler, Bell, Chemers, & Patrick, 1984; Planek & Feam, 1993; Simonds & Shafai-Sahmi, 1977; Smith, Cohen, Cohen, & Cleveland, 1978) or the major dimensions of positive or supportive organizational safety climates (e.g., Brown & Holmes, 1986; Dedobbeleer 8z Beland, 1991; Mattila, Rantanen, & Hyttinen, 1994; Niskanen, 1994; Zohar, 1980). This research has been useful but it has not provided a comprehensive understanding of worker self-protection. More is now known about the importance of certain individual and organizational characteristics, but there has been very little comparability from one study to another, and few generalizable conclusions can be drawn. For example, it is difficult to make predictions about which factors are likely to be important in a given work situation or how these factors might interact with each other. Almost nothing is known about the interplay of individual and organizational factors in influencing the safety-related actions of workers on the job. A number of conceptual models of workplace safety and health have been proposed to help organize research findings and guide the development of occupational safety and health programs (e.g., Cohen, Smith, & Anger, 1979; DeJoy, 1990; DeJoy & Southem, 1993; Smith & Beringer, 1987). These models are useful in a heuristic sense because they underscore the need to consider the characteristics of the worker, the tasks and jobs being performed, the total work environment, and the organizational structure. But for the most part, they fall far short of being comprehensive or testable models. In view of these limitations, it may be worthwhile to look to related areas for more sophisticated and applicable models. The general area of health behavior offers a number of theoretical models that have been tested and refined over the years. 62 Theoretical Models of Health Behavior A variety of models have been developed to explain why people do or do not engage in various health enhancing and medically beneficial activities. These models have been applied to a wide variety of preventive and lifestyle behaviors (e.g. vaccinations, safe sex practices, smoking, exercise, seat belt usage), screening or early detection activities (e.g., breast self-examination, cholesterol and blood pressure testing, genetic screening), sick role behaviors (e.g.. clinic utilization, physician visits), and adherence to medical and other therapeutic regimens (e.g., hypertension, diabetes). Although there would appear to be an obvious link between these models and workplace self-protective behavior, very little attention has been given to examining how they might apply to actions that workers are asked to take to protect themselves from job-related hazards. Three categories of models are considered in this paper. The first category includes the decision making or cost-benefit models derived from value-expectancy theory. These models have occupied a prominent position in the health behavior literature during the last 30 years (Glauz, Lewis, & Rimer, 1990, Weinstein, 1993). The second category is referred to here as environmental or contextual models. These models attempt to go beyond person-focused variables and take a more ecological or interactionist approach to analyzing the determinants of health behavior. The fmal category of models focuses on the behavior change process itself. Models in this category typically portray behavior change as a series of qualitatively distinct stages. Value-Expectaucy Models Value-expectancy models are based on the premise that people estimate the seriousness of risks, evaluate the costs and benefits of various actions, and then choose a course of action that will maximize the expected outcome (Cleat-y, 1987). Value-expectancy models have taken a variety of forms, and three prominent examples are discussed here: the Health Belief Model (Becker, 1974); the Theory of Reasoned Action (Ajzen & Fishbein, 1980); and Protection Motivation Theory (Rogers, 1983). The three models are each different to some extent, but they all emphasize the individual’s threat-related Journal of Safety Research beliefs or perceptions. Weinstein, (1993) argues that value-expectancy models have four characteristics in common: (a) that motivation for selfprotective behavior arises from the anticipation of negative consequences and the desire to minimize these outcomes, (b) that the impact of an anticipated negative outcome on motivation depends on beliefs about the likelihood that this outcome will occur, (c) that motivation to act arises from the expectation that the action will reduce the likelihood or severity of harm, and (d) that the expected benefits of a particular action must be weighed against the expected costs of taking the action. Health Belief Model (HBM). Of the models pre- sented in this section, the HBM has produced the largest body of health-related research. It is also the only one that was specifically developed to explain health behavior. The HBM model has four basic components: (a) perceived susceptibility to the health problem or condition in question, (b) the perceived seriousness of the problem or condition, (c) the perceived benefits associated with taking a particular action, and (d) the perceived barriers associated with taking the action. Published reviews of the HBM literature (Becker, 1974; Harrison, Mullen, & Green, 1992; Janz & Becker, 1984) show considerable support for the model and offer some general conclusions about the relative importance of its major components. Perceived barriers have been shown to be the most powerful single predictor across all studies and behaviors. Perceived susceptibility and perceived benefits are also important, with susceptibility being more important for preventive than sick role behaviors. As might be expected, benefits are more important than susceptibility for sick role behaviors. In terms of the total literature, perceived severity appears to be the weakest of the four dimensions. One of the major criticisms of the I-IBM is that is does not specify relationships among the major variables. Indeed, it has sometimes been referred to as being more a list of variables than a theoretical model (Wallston & Wallston, 1984; Weinstein, 1993). Most applications of the HBM combine the variables in a linear or additive fashion to test the model: susceptibility + severity + (benefits - barriers). However, some of the earlier writing on the model (e.g., Maiman & Becker, 1974) would appear to sugSummer 1996/Mohme 27LWmber 2 gest a multiplicative model: susceptibility severity x (benefits - barriers). x Theory of reasoned action (TRA). The TRA posits that behavioral intention is the immediate determinant of behavior and that all factors that influence a particular behavior are mediated through intention. Intention is determined by two components: (a) attitude toward the behavior, which consists of beliefs about the consequences of performing the behavior and the evaluation of those consequences; and (b) subjective norms, which consist of normrutive beliefs about what salient others think and the individual’s motivation to comply with those wishes. According to the model, intention can be predicted by the linear combination of attitude and normative beliefs multiplied by motivation to comply with the beliefs. The model is expressed as a multiple regression equation, with the weights assigned to the major components determined by multiple regression procedures. The TRA has been applied with considerable success to a number of health behaviors, inaluding exercise, weight loss, child safety seats, smoking, condom usage, and alcohol and drug use (see Cleary, 1987; Kirscht, 1983; Sutton, 1987 for reviews). Besides its focus on behavioral intention, the most obvious difference between the TRA and HBM is that the TRA includes subjective norms as a major determinant of health-related behavior. The TRA also goes further in specifying how its constructs should be measured and how they combine to form behavioral intention. On the negative side, much of the research on the TRA has been confimed to predicting behavior intention rather than aatual behavior (Baranowski, 1992-1993). More recent versions of both the HBM (Rosenstock, Strecher, & Becker, 1988) and the TRA (Ajzen, 1985) have added self-efficacy (Bandura, 1986) as an important component. The principal argument for including self-efficacy is that people must feel confident that they are capable of performing the behaviors required to produce the desired outcomes. Self-efficacy appears to be especially important for lifestyle modifications and other behaviors involving long-term change and maintenance (Streaher, DeVelLis,Becker, & Rosenstock, 1986). Protection motivation theory (PMT). This model features two cognitive processes, threat appraisal 63 and coping appraisal, which combine to form protection motivation. Protection motivation (typically measured as behavioral intention) is conceptualized as an intervening variable that activates coping behavior. Threat appraisal evaluates the intrinsic (e.g., pleasure) and extrinsic (e.g., social approval) rewards that increase the probability of making a maladaptive response against perceptions of vulnerability and outcome severity that decrease the probability. Threat appraisal is the algebraic sum of these variables. Coping appraisal consists of judgments about the efficacy of a preventive response (response efficacy) plus the assessment of one’s ability to successfully perform the necessary responses or behaviors (self-efficacy) minus the costs associated with the response. Protection motivation is assumed to be greatest when: (a) the perceived threat is severe, (b) the individual feels vulnerable, (c) the adaptive response is believed to be effective, (d) the person is confident of his or her abilities to complete the adaptive response, (e) the rewards of the maladaptive behavior are small, and (f) the costs of the adaptive behavior are low. Although this is an additive model within each of the two appraisal processes, interactive effects can occur between threat and coping appraisal processes (e.g., selfefficacy x severity). Presumably, this allows the model to predict outcomes that are contrary to totally rational decision making (Prentice-Dunn & Rogers, 1986). Although less extensively researched than the two previous models, the major components of PMT have been supported (Prentice-Dunn & Rogers; Rogers, 1983). Much of this work has involved fear-arousing communications and attitude change. Applicability Behavior to Workplace Setf-Protective Aside from some consideration of HBM constructs with respect to personal protective equipment (e.g., Cleveland, 1984; DeJoy, 1986; Terrell, 1984), very little effort has been made to apply the value-expectancy models to worker safety and health. However, viewed as a group, these models do highlight several constructs that should be relevant to workplace self-protective behavior. These constructs include: (a) threatrelated beliefs, (b) self-efficacy, (c) response efficacy, (d) barriers, and (e) normative expectations (see Table 1). 64 Threat-related beliefs. Beliefs about susceptibility (probability) and severity are featured in each of the models. The interrelationship of susceptibility and severity tends to be treated differently in the various models, but a multiplicative relationship would appear to be most in line with the value-expectancy tradition. For example, a high level of perceived susceptibility may not necessarily lead to self-piotective behavior if the severity of the threat is minimal. On the other hand, for a highly lethal outcome, perceived susceptibility may be the major dimension. Self-efficacy. In one form or another, self-efftcacy has been incorporated into each of the valueexpectancy models. Because most instances of worker self-protection involve the performance of a set of prescribed actions on a long-term basis, self-efficacy should be an important factor. Workers need to feel confident about their ability to perform required behaviors on a regular and long-term basis. efficacy. Beliefs about the consequences or effectiveness of preventive action also play a role in each of the models. They are treated as benefits in the HBM, as part of attitudes in the TRA, and as response efficacy in PMT. In most workplace applications, response efficacy involves perceptions about the effectiveness of prescribed work practices or protective equipment in preventing hazardous exposures. Response Barriers are considered explicitly in the HBM and the PMT, and indirectly in the TRA. Indeed, the HBM literature (Janz & Becker, 1984) suggests that barriers or costs are the single best predictor of health behavior. Relevant to the workplace, research on the use of personal protective equipment shows that job related barriers are often a major factor in noncompliance (Acton, 1977; Cleveland, 1984; Terrell, 1984). Barriers. expectations. The TRA is the model that considers the effects of the social environment most directly. Social influences are also considered, somewhat tangentially, in the intrinsic-extrinsic rewards portion of PMT. Research on safety program effectiveness (e.g., Cohen & Cleveland, 1983; Fiedler et al., 1984; Planek & Feat-n, 1993; Simonds & Shafai-Sahrai, 1977) and safety climate (e.g., Dedobbeleer & Beland, Normative Journal of Safety Research VALUE-EXPECTANCY COll~tnrCt TABLE 1 APPLIED TO WORKPLACE CONSTRUCTS Definition SELF-PROTECTIVE Workplace Example Construction Threat-related beliefs Beliefs about hazard susceptibility Beliefs about one’s ability Self-efficacy safety measures and severity to follow indicated successfully likelihood Response efficacy Emergency medical safety expectations that interfere worker Contextual or Environmental with the use of available Physical hearing factors that influence self-protection Models The models discussed thus far have been primarily concerned with how the individual’s attitudes, beliefs, and expectations influence his or her reaction to various health threats. Only limited attention is given to social or environmental factors, or to the context in which the individual operates. The need for a person x situation or interactionist approach to workplace safety and health has been discussed in general terms by several authors (e.g., DeJoy & Southern, 1993; DeJoy, Wilson, & Huddy, 1995; Sheehy & Chapman, 1987; Smith & Beringer, 1987). The interactionist perspective holds that individual and situational (environmental) factors combine multiplicatively in influencing worker behavior. PRECEDE model. The PRECEDE model developed by Green and colleagues (Green & Kreuter, 199 1; Green, Kreuter, Deeds, & Partridge, 1980) would appear to qualify as a contextual model. “PRECEDE” is an acronym for “predisposing, reinforcing, and enabling causes in educational diagnosis and evaluation,” and this model was developed as a framework to be used in plan2 7LVumber 2 automated technician’s his/her beliefs of used needles coal miner’s carbon about confidence monoxide about safely in detection system 1991; Mattila et al., 1994; Niskanen, 1994; Zohar, 1980) highlight the important role of social-organizational factors in supporting good safety performance. The safety literature also contains a number of studies indicating that performance and other types of feedback from supervisors and coworkers can be an important factor in shaping work-related safety behavior (see review by McAfee & Winn, 1989). Summ~ I996Nolume safety measures Social/organizational Normative of available beliefs off scaffolding being able to dispose measures Factors Barriers effectiveness worker’s of falling Underground Perceived BEHAVIOR discomfort protectors Supervisor’s oractices associated in textile indifference in warehousino with wearing plant to safe lifting ooeration ning health education programs. This basic framework has been adapted to workplace selfprotective behavior (Dedobbeleer & German, 1987; DeJoy, 1986; Peters, 1991). In the PRECEDE model, three sets of diagnostic or behavioral factors drive the development of prevention strategies. Predisposing factors are the characteristics of the individual (beliefs, attitudes, values, etc.) that facilitate or hinder self-protective behavior. Predisposing factors are conceptualized as providing the motivation for behavior. The threat-related beliefs and efficacy expectancies that are prominent features of the value-expectancy models would be included here. Enabling factors refer to objective aspects of the environment or system that block or promote self-protective action. Green and colleagues (Green et al., 1980) define enabling factors as “factors antecedent to behavior that allow motivation or aspiration to be realized” (p. 68). The skill and knowledge necessary to follow prescribed actions would be included here, as would the availability and accessibility of protective equipment and other resources. Most barriers or costs would be classified as enabling factors. Reinforcing factors involve any reward or punishment that follows or is anticipated as a consequence of the behavior. Performance feedback and the social approval/disapproval received from coworkers, supervisors, and managers would qualify as reinforcing factors in workplace settings. The PRECEDE model was developed as a program-planning framework rather than as a causal model of health behavior (Green et al., 1980). The three diagnostic categories were 65 intended primarily as reference points for analyzing the determinants of behavior that would be most responsive to health education efforts. As such, little attention was given to delimiting the categories or to specifying combinational rules. The PRECEDE model has been used extensively to plan and evaluate health education and related programs in a variety of settings, including the workplace (Green & Kreuter, 1991). However, there have been relatively few formal tests of this model. Applicability Behavior to Workplace Self-Protective The PRECEDE model increases the saliency of environmental or contextual variables in two important ways: (a) by directing attention to the skills and resources that are prerequisite to the achievement of behavioral goals, and (b) by viewing the environment as an important source of support and reinforcement for behavior change and maintenance. The recently revised version of the PRECEDE model, the PRECEDE/PROCEED model (Green & Kreuter, 1991) is even more distinctly environmental. The revised version proposes that an environmental diagnosis should occur along with the behavioral diagnosis, and that special attention should be given to the interaction of behavioral and environmental factors. This encourages the selection of both behavioral and environmental targets for change within the intervention program. However, it is the interactive nature of the three diagnostic factors that is probably of greatest potential importance to understanding worker behavior. In particular, efforts to influence the beliefs and attitudes of workers and, thus, motivate them to follow safe practices may fail if the environment is nonsupportive. This calls attention to the importance of jobrelated barriers, the ready availability of safety equipment and devices, and the importance of skill-based training in facilitating self-protective behavior. A second point is that even wellmotivated and well-trained workers may not respond appropriately if doing so is not acknowledged or reinforced by peers, supervisors, and management. Performance feedback and the safety-related attitudes and actions of management would appear to be especially important in this regard. 66 Behavior Change Models The final category of models includes those that focus on the process of behavior change. The Transtheoretical Model (Prochaska & DiClemente, 1982) and the Precaution-Adoption Process (Weinstein, 1988) are perhaps the best exemplars of this general category. Both of these models portray the change process as a series of stages. A fundamental tenet of the stage change perspective is that people at different points in the change process require different types of information and assistance to move to the next stage. The variables or factors important at one stage may be quite unimportant at another stage. For example, beliefs about susceptibility or severity may be important early in the change process, such as during initial awareness and decision making, but become less important once actual behavior change is underway. Transtheoretical Model. Prochaska and colleagues (Prochaska, DiClemente, & Norcross, 1992) describe the process of behavior change as consisting of five principal stages: (a) precontemplation, (b) contemplation, (c) preparation, (d) action, and (e) maintenance. Precontemplation is the stage in which the person is not seriously thinking about changing his or her behavior. Contemplation begins when the person starts to think seriously about changing in the near future (i.e., about 6 months). Preparation denotes that the person intends to make a change in the very near future, and that he or she has a plan for action and has already made small or preliminary behavior changes (thus, preparation has both intentional and behavioral dimensions). Action is the period following initial behavior change (usually about 6 months). Maintenance extends from this point until termination. A growing body of evidence exists for the basic stages of change construct, mostly involving addictive behaviors and psychotherapy (Prochaska & DiClemente, 1992). Work with this model has also sought to uncover the processes of change, or the activities that individuals engage in when they attempt to modify problem behaviors. Of considerable importance is the observation that the processes of change appear to be quite similar across a variety of different behaviors (Prochaska et al., 1992). That is, for many behaviors, information-seeking and consciousness-raising activities are relevant for people entering contemplation, and stimulus control Journal of Safety Research and social support are important processes for those in the action and maintenance stages. The ultimate goal is to integrate the stages and pro cesses of change, and to identify the most effective strategies for moving people from one stage to the next (DiClemente et al., 1991). Precaution-adoption process. The precautionadoption process (Weinstein, 1988) also has five stages: (a) has heard of hazard, (b) believes in susceptibility for others, (c) acknowledges personal susceptibility, (d) decides to take precaution, and (e) takes precaution. A central feature of this model is that personal susceptibility is treated as a series of three stages rather than as a single dimension or continuum. This conceptualization derives from the idea that people are likely to have little interest in taking precautions unless they feel personally vulnerable. A further complication is that people tend to be optimistically biased in judging their personal levels of risk. That is, for many different hazards, people consider their own risk to be considerably less than that of most other people (Weinstein, 1980, 1982, 1987). The decision making stage in this model involves many of the same hazard-related beliefs and cost-benefit considerations that are featured in the value-expectancy literature. However, since the stages are cumulative, decision making does not begin until the requirements of stage three have been met. Barriers are particularly important in translating a decision into action (stage 5). Weinstein and colleagues have tested this model in several studies of home radon testing (Weinstein & Sandman, 1992). Applicability Behavior to Workplace Self-Protective Inherent in both stage models is the view that different kinds of information and interventions will have different saliencies depending on where the individual is in the change process. For example, minimizing the costs or barriers associated with following a particular self-protective action will have very little impact on people who are not yet aware of the threat or who do not think that they are personally susceptible to it. Value-expectancy models, in contrast, assume that the probability of action is essentially an algebraic function of the individual’s beliefs and that this same predictive equation applies across the entire behavior change proSummer 1996Nohne 27/Number 2 cess. The stage change perspective argues that different equations are needed for each stage. Providing workers with information about particular job-related hazards is likely to be most useful for increasing awareness and a sense of personal susceptibility. But beyond this point, such information is likely to be of limited effectiveness in changing and maintaining relevant work practices. Once awareness and personal susceptibility are established, attention might better be focused on related skill development and self-efficacy enhancement, and with actions that reduce barriers and create more favorable cost-benefit ratios for safe behavior. Long-term adherence ultimately requires constructing task and work environments that support safe behavior, even under the most adverse workload conditions. Borrowing from relapse prevention research (e.g., Marlatt & Gordon, 1985), conscious efforts are needed to assess the situations and circumstances under which adherence is likely to fail. Alternative strategies or coping responses should be available to workers for use in such situations. The identification and analysis of special requirements and high risk situations should be an important feature of a comprehensive safety program. The importance of environmental supports during the action and maintenance stages blends nicely with the preceding discussion of enabling and reinforcing factors within the PRECEDE framework. In the final analysis, high levels of both individual and collective control may be critical to successful long-term adherence. Presumably, efforts to enhance self-efficacy have the direct effect of increasing the individual’s sense of personal control. However, groups of people may also possess a sense of control. Peterson and Stunkard (1989) describe collective control as: “a norm - or shared belief - about the way that the group works, what it is that the group can and cannot accomplish by what actions” (p. 822). As described by these authors, collective control bears considerable resemblance to Zohar’s (1980) concept of safety climate. For Zohar, safety climate is: “a summary of molar perceptions that employees share about their work environments” (p. 96). It is generally thought that safety climate serves as a frame of reference for guiding relevant behavior in the workplace and that employees develop reasonably coherent expectations regarding behavioroutcome contingencies in their environment (cf. 67 Schneider, 1975). It follows that collective control or safety climate should be an important consideration in fostering long-term and broad-based adherence to safe work practices. Au Integrative Framework for Workplace Self-Protective Behavior Each of the categories of models reviewed thus far has something to contribute to understanding workplace self-protective behavior. The value-expectancy models provide a fairly circumscribed set of person-focused variables that are likely to be important to any type of self-ptotective behavior. The PRECEDE model, as an environmental or contextual model, directs attention to the interaction of person and situational factors, and to how the work environment, broadly defined, can enable and reinforce selfprotective actions. The stage models emphasize that precautionary behavior is inherently dynamic and comprised of qualitatively different phases or stages. Figure 1 portrays self-protective behavior as consisting of four stages: hazard appraisal, decision making, initiation, and adherence. These stages may be viewed as representing different levels of motivational readiness to engage in self-protective behavior. However, the factors that influence this readiness are not limited to person-focused variables. Table 2 features five general constructs that are likely to be important to self-protective actions in the workplace. The constructs build on those portrayed earlier (see Table l), with two modifications. The termfacili&.&g conditions expands the concept of barriers relative to its usage in the value-expectancy models and emphasizes the importance of environmental supports in self-protective behavior. Also, safety climate is used to represent the manifold of social and organizational factors that may impinge on workplace behavior. Each of the five constructs can be further described in terms of specific variables. Because current knowledge is not sufficient to exclude any of the constructs at a particular stage, the five constructs in Table 2 are labeled as being of either primary or secondary importance for each stage. As with other stage models, the stages should be treated as cumulative; for example, the individual must believe that a particular hazard represents a serious threat before he or she will enter the decision making stage. Efficacy considerations and facilitating conditions will be of little importance for those who are unaware of the hazard or for those who have appraised the hazard as insignificant. A discussion of the four stages follows. Hazard appraisal. The individual’s threat-related beliefs are of primary importance during this stage. As discussed earlier, perceived susceptibility (likelihood) and severity can be expected to interact to form perceived threat. Response efficacy, or the availability of effective preventive strategies, may also be important at this stage. Defensive or self-protective motives (e.g., Miller & Ross, 1975) may cause the individual to dismiss or discount a hazard for which preventive strategies are unavailable or of limited effectiveness. Unwarranted optimism or overconfidence may also play a role in hazard appraisal (Weinstein, 1987). At this stage, workers should benefit from information about the hazard itself, including risk estimates, exposure modes, available control measures, and so forth. The importance of having workers personalize the risks should not be ignored. FIGLJRE 1 STAGE OR SEQUENTIAL MODEL OF WORKPLACE SELF-PROTECTIVE BEHAVIOR T Yes Appraisal I_C DecisionMaking- No + No Yes I.tiation ererice No + No + Exposure to injury/illness 68 Journal of Safety Research TABLE 2 SELF-PROTECTIVE BEHAVIOR CONSTRUCTS AND THE STAGES OF SELF-PROTECTIVE BEHAVIOR Construct Threat-related Response Safety Note: ................ .................. efficacy Self-efficacy Facilitating Hazard beliefs ...................... conditions ................ climate P= primary unponance: S =secondary Appraisal Adherenoe s S S P P S S S P S S S P P P S P P P importance Initiation. Facilitating conditions and safety climate are the principal constructs during the initiation stage. At this stage, attention shifts to the environmental and organizational factors that support and reinforce self-protective action. These factors become very important as behavioral intention is translated into action, and as workers try out new behaviors. Although safety climate remains a rather elu27/iVumber 2 Initiation P Decision making. Efficacy considerations and the costs and benefits associated with alternative courses of action become important during the decision making stage. Efficacy considerations include those related to the effectiveness of available precautionary actions (response efficacy), as well as the individual’s perceptions of his or her ability to successfully accomplish these actions (self-efficacy). Self-efficacy can be developed through education/ training and skill-building exercises, as well as actual experience in performing the relevant behaviors. The modeling of these behaviors by coworkers may also enhance self-efficacy expectancies. The cost-benefit portion focuses on weighing the benefits of the self-protective action against the costs incurred. Costs may include time constraints, actual or imagined reductions in productivity or skilled performance, physical discomfort, and any other encumbrances that reduce “quality of life” on the job. Facilitating conditions serve to counteract costs, and might include the ready availability of needed safety equipment, training in the correct usage of this equipment, and redesigned jobs and equipment that make self-protection easier and more effective. The cost-benefit analysis might also include safety climate considerations. Summer I99tWohme Decision-making sive concept, it appears reasonable to conclude that the attitudes and actions of management play a prominent role in the safety climate of the organization (Brown & Holmes, 1986; Dedobbeleer & Beland, 1991; Zohar, 1980). Safety performance information and other types of feedback received from coworkers and supervisors is also likely to be an important aspect of safety climate. Adherence. Environmental and organizational factors can be expected to play major roles in sustaining long-term adherence. As workers successfully follow prescribed actions, their self-efficacy should increase, and with that, their resistance to nonadherence. Many self-protective actions become relatively automatic over time, and this should also help to sustain longterm adherence. Response efficacy should also increase with time as workers become more confident of the effectiveness of the actions in question. Still, efficacy considerations remain secondary to barriers and safety climate at this stage. Long-term maintenance ultimately depends on creating task and work environments that support safe behavior, even under the most stressful or demanding conditions. Although this proposed framework certainly needs to be tested and validated, there are several lines of inquiry that support the general organization of the framework and the assignment of various self-protective behavior constructs to particular stages. First, it is generally acknowledged that awareness and knowledge are necessary but often insufficient to produce and sustain behavior change, including selfprotective actions in the workplace. This supports the focus on threat-related beliefs and response efficacy in the first stage, and the 69 view that hazard appraisal is only one stage in the process. Second, there is considerable agreement that behavior change and maintenance are different processes. As such, a construct or factor that is important in bringing about initial behavior change may not play the same role in sustaining the.change over time. Third, there is broad agreement that the individual’s subjective beliefs related to costs and benefits play a central role in most health- and safety-related decision making. And fourth, organizational and environmental factors appear to be particularly important in facilitating and sustaining desired behaviors. Indeed, the occupational safety and health literature suggests that these factors may be critically important in workplace settings. CONCLUSIONS This paper reviewed three categories of theoretical models that have been used to analyze and predict health-related behavior. On this basis, an integrative framework was proposed that conceptualizes workplace self-protective behavior as consisting of four sequential stages: hazard appraisal, decision making, initiation, and adherence. The proposed framework is integrative in that its basic structure comes from the behavior change models, while the constructs or factors considered relevant at each stage are derived from the value-expectancy and environmental/contextual models. The principal implication of this framework is that different constructs or factors can be expected to be important at different stages. Threat-related beliefs and response efficacy are likely to be most important during hazard appraisal. During this stage, workers determine both what is dangerous and whether they can protect themselves by taking specific actions and/or by using specific protective equipment. Efficacy considerations (response and selfefficacy) and cost-benefit factors assume increased importance during the decision making stage. Workers determine that available safety protocols are both effective and practical. The personal weighing of costs and benefits is at the very heart of this stage. This calculus takes into consideration beliefs about the threat itself, expectancies surrounding the required actions, barriers and environmental supports, and infer70 ences about the attitudes and actions of peers, supervisors, and management. During the initiation and adherence stages, facilitating conditions and safety climate are the primary factors. Attention is focused on the environmental and organizational factors that support and reinforce self-protective actions. Performance feedback may be a particularly valuable tool for encouraging workers to accept that the self-protective actions in question are effective and practical. Adherence, as the final stage, requires the creation of task and work environments that support safe behavior under all conditions. High levels of both individual and collective control may be critical to longterm adherence, and creating a positive safety climate may be the best way to foster perceived control at both levels. In practical terms, there is a need to match interventions and stages; the utility of any intervention depends on where the individual worker is in the change process. For example, interventions to improve worker self-protection typically revolve around hazard communication and information-based training in safe work practices. Such efforts are likely to be most beneticial during the hazard appraisal and decision making stages. These particular interventions are much less relevant to the factors hypothesized to be most important during initiation and adherence. It follows that programs to maximize self-protective behavior in the workplace need to be multifaceted and should feature interventions targeted to workers at each of the four stages. As a fmal point, the proposed framework also highlights the importance of situational or environmental factors in workplace self-protective behavior. As Table 2 shows, the individual’s personal beliefs about the hazard and the effectiveness of precautionary action are most important during the first stage (hazard appraisal). But, beyond this point, environmental factors become increasingly more important, and become critical as intention is converted into action (initiation), and as self-protective behavior is sustained (adherence). Given the interactive aspect of individual and situational factors, even well-designed efforts to influence the hazard-related beliefs and attitudes of workers may fail if the environment is nonsupportive. The importance of the environment in enabling and reinforcing safe workplace behavior cannot be overemphasized. Journal of Safety Research REFERENCES Acton, W. I. (1977). Problems associated with the use of bearing protection. Annals of Occu@mral Hygiene, 29, 387-395. Ajzen, I. (1985). From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckmann (Eds.), Action control: From cognition to behavior (pp. 1I-40). Berlin: Springer-Verlag. Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting behavior. Englewood Cliffs, NJ: Prentice Hall. Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ: Prentice-Hall. Baranowski, T. (1992-93). Beliefs as motivational influences at stages in behavior change. bttemational Quarterly of Community Health Education, Z3( 1). 3-29. Becker, M. H. (1974). The health belief model and personal health behavior. Thorofam, NJ: Slack Press. Brown, R. L., & Holmes, H. (1986). The use of a factoranalytic procedure for assessing the validity of an employee safety climate model. Accident Analysis and Prevention, 18(6), 455-470. Cleaty, P. D. (1987). Why people take precautions against health risks. In N. D. Weinstein (Ed.), Taking care: Understanding and encouraging self-protective behavior (pp. 119-149). Cambridge: Cambridge University Press. Cleveland, R. J. (1984). Factors that influence safety shoe usage. Professional Safety, 29.26-29. Cohen, A., Smith, M. J., & Anger, W. K. (1979). Selfprotective measures against workplace hazards. Journal ofsafetyResearch, 11(3). 121-131. Cohen, A., Smith, M. J., & Cohen, H. H. (1975). Safety program practices in high versus low accident rate companies - An interim report (questionnaire phase; DHEW Publication No. 75-185). Ciiinnati, OH National Institute for ocCupational Safety and Health. Cohen, H. H., & Cleveland, R. J. (1983). Safety program practices in recording-holding plants. Professional Safety, 28.26-33. Cooper, C., & Sutherland, V. (1987). Job stress, mental health, and accidents among offshore workers in the oil and gas extraction industries. Journal of Occupational Medicine, 29(2), 119125. Cox, S.. 81 Cox. T. (1991). The structure of employee attitudes to safety: A European example. Work and Stress, S(2), 93-106. Dedobbeleer, N., & Beland, F. (1991). A safety climate measure for construction sites. loumai of Safety Research, 22(2), 97-103. Dedobbeleer, N., & German, P. (1987). Safety practices in construction industry. Journal of Occupational Medicine, 29(1 I), 863-868. DeJoy, D. M. (1986). A behavioral-diagnostic model for self-protective behavior in the workplace. Professional Safety, 3 I, 26-30. DeJoy, D. M. (1990). Toward a comprehensive human factors model of workplace accident causation. Professional Safety, 35, 1l-!6. DeJoy, D. M., & Southern, D. S. (1993). An integrative perspective on worksite health promotion. Journal of Occupational Medicine, 35,122 1-l 230. DeJoy, D. M., Wilson, M. G., & Huddy, D. C. (1995). Health behavior change in the workplace. In D. M. DeJoy Summer 19%Nolume 27LNumber 2 & M. G. Wilson @is.), Critical issues in worksite health promotion (pp. 97-122). Boston, MA: Allyn and Bacon. DiClemente. C. C., Prochaska, J. O., Fairhurst, S. K., Velicer, W. E, Velasquez, M. M.. & Rossi. J. S. (1991). The process of smoking cessation: An analysis of precontemplation, contemplation, and preparation stages of change. Journal of Consulting and Clinical Psychology, 59,295-304. Edwards, D. S., & Hahn, C. P (1980). A chance to happen. Journal of Safety Research, 12,59-67. Fiedler, F. E., Bell, C. H., Chemers, M. M., & Patrick, D. (1984). Increasing mine productivity and safety through management training and organization development. Basic and Applied Social Psychology, 5. I-1 8. Glanz, K., Lewis, F. M., & Rimer. B. K. (1990). Health behavior and health education. San Francisco: Jossey-Bass. Goldberg, A. I., Dar-cl, E. M., & R&ii, A. E. (1991). Threat perception and the readiness to participate in safety programs. Joumalof OtganizationalRehavior, f2,IO!&122. Green, L. W., & Kreuter. M. W. (I 991). Health promotion planning: An educational and environmental approach (2nd ed.). Mountain View, CA: Mayfield. Green, L. W., Kreuter. M. W.. Deeds, S. G.. & Partridge, K. B. (1980). Health education planning: A diagnostic upproach. Palo Alto, CA: Mayfield. Harrison, J. A., Mullen. P. D., & Green, L. W. (1992). A meta-analysis of studies of the health belief model with adults. Health Education Research, 7( I ), 107-l 16. Howarth, C. I. (1987). Perceived risk and behavioural feedback: Strategies for reducing accidents and increasing efficiency. WonGand Stress, I, 6 l-65. Janz, N. K., & Becker, M. H. (1984). The health belief model: A decade later. Health Education Quarterly, /I( I), 147. Jones, J., & Wuebker, L. (1985). Development and validation of the safety locus of control scale. Perceptual and Motor S&ilk. 61,151-161. Kirscht, J. P. (1983). Preventive health behavior: A review of research and issues. Health Psychology, 2.277-301. Komaki, J., Barwick, K. D., & Scott, L. R. (1978). A behavioral approach to occupational safety: Pinpointing and reinforcing safe performance in a food manufacturing plant. Journal of Applied Psychology, 63.434445. Komaki. J. L., Zlotnick, S., &Jensen. M. (1986). Developpment of an operant-based taxonomy and observational index of supervisoty behavior. Joumal ofApplied Psychology, 71(2), 260-269. Landeweerd, J. A., Urhngs, I. J. M., Dc.Jong, A. H. J.. Mijhuis, F. J. N.. & Bouter, L. M. (1990). Risk taking tendency among construction workers. Journal of Occupational Accidents, II, 183-196. Leather, P. J. (1988). Attitudes toward safety performance on construction work: An investigation of public and private sector differences. Work nnd Stress, 2. 155-l 67. Levenson. H., Hirschfeld, M., & Hirschfeld, A. (1980). Industrial accidents and recent life events. Journal of Occupational Medicine, 22( I ), 53-57. Maiman, L. A., & Becker, M. H. (1974). The health belief model: Origins and correlates in psychological theory. Health Education Monographs, 2.336-353. Marlatt, G. A., & Gordon, J. R. (1985).Relapse prevention: A self-control strategy for the maintenance of behavior change. New York: Guilford Press. Mat& M., Hyttinen, M., & Rantanen, E. (1994). Effective supervisory behaviour and safety at the building site. International Journal of Industrial Ergonomics, 13, 85-93. 71 Mattila, M., Rantanen. E.. & Hyttinen, M. (1994). The quality of work environment, supervision, and safety in building construction. Safery Science, 17.251-268. McAfee. R. B., & Wmn, A. R. (1989). The use of incentivesl feedback to enhance workplace safety: A critique of the Iiterature. Journal of Snferv Research, 20(l), 7-19. Miller, D. T.. & Ross, M. (1975). Self-serving biases in the attribution of causality: Fact or fiction? Psychological Bulletin, 82.213-225. Murphy, L. R. (1984). Stress management in highway maintenance workers. Journal of Occupational Medicine, 26,436-442. Niskanen, T. (1994). Safety climate in the road administration. Safety Science, 17.237-255. Peters, R. H. (1991). Strategies for encouraging self-protective employee behavior. Journal of Safety Research, 22.53-70. Peterson, C., & Stunkard, A. J. (1989). Personal control and health promotion. Social Science and Medicine, 28(8), 819-828. Planek, T. W., & Fearn, K. T. (1993). Reevaluating occupational safety priorities: 1%7 to 1992. Professional Safery. 38, 16-2 1. Theory, Research, and Practice, 20, Prochaska, J. 0.. & DiClemente. C. C. (1992). Stages of change in the modification of problem behaviors. In M. Hersen, R. M. Eisler, & P M. Miller (Eds.), Progress in behavior modification (pp. 184-214). Sycamore, IL: Sycamore Press. Prochaska, J. 0.. DiClemente, C. C.. & Norcross, J. C. (1992). In search of how people change. American Psychologisr, 47(9), 1102-l 114. Rhoton. W. W. (1980). A procedure to improve compliance with coal mine safety regulations. Jownnl qf0~@&011& Behavior Management, 2.243249. Rogers, R. W. (1983). Cognitive and psychological processes in fear appeals and attitude change: A revised theory of protection motivation. In J. T. Cacioppo & R. E. Petty (Eds.). Social psychophysiology (pp. 153-176). New York: Guilford Press. Rosenstock, I. M., Strecher, V. J.. & Becker, M. H. (1988). Social learning theory and the health belief model. Health Education Quarterly, 1.5, 175-183. Schneider, B. (1975). Organizational climates: An essay. Personnel Psychology, 28.447-479. Sheehy, N. I!, Sr Chapman, A. J. (1987). Industrial accidents. In C. L. Cooper & I. T. Robertson (Eds.), International review of industn’al and organizational psychology. New York: John Wiley & Sons. Simonds. R. H., & Shafai-Sahrai, Y. (1977). Factors apparently affecting the injury frequency in eleven matched pairs of companies. Journal of Safety Research, 9.120-127. Smith, M. J., Anger, W. K., & Uslan, S. S. (1978). Behavior modification applied to occupational safety. Jouml of 72 Journal of Sr$ety Research, IO, 5-15, Smith, M. J., Colligan, M. J.. Fro&t, I. J., & Tasto, D. L. (1982). Occupational injury rates among nurses as a function of shift. Journal of Safety Research, 4,181-l 87. Strecher, V. J.. DeVellis, B. M., Becker, M. H., & Rosenstock, I. M. (1986). The role of self-efficacy in achieving health behavior change. Health Education Quarterly, 13.7392. Sulzer-Azaroff. B., & DeSantamaria, M. (1980). Industrial safety hazard reduction through performance feedback. Journal of Applied Behavioral Analysis, 13,281-295. Sutton, S. (1987). Social-psychological approaches to understanding addictive behaviors: Attitude-behaviour and decision making models. British Journal of Addicton, 82.355-370. Prentice-Dunn, S., & Rogers, R. W. (1986). Protection motivation theory and preventive health: Beyond the health belief model. Health Education Research. I(3). .,. 153-161. Pro&&a, J. O., & DiClemente, C. C. (1982). Tramtheoretical therapy: Toward a more integrative model of change. Psychotherapy: 161-173. Safer>,Research, IO, 87-88. Smith, M. J.. & Beringer. D. B. (1987). Human factors in occupational injury evaluation and control. In G. Salvendy (Ed.), Handbook of human factors (pp. 767-789). New York: Wiley-Interscience. Smith, M. J., Cohen, H. H., Cohen, A., & Cleveland, R J. (1978). Characteristics of successful safety programs. Terrell, P G. (1984). How to increase worker acceptance of respirators. Professional Safety, 29. 15-20. Wallston. B. S.. & Wallston, K. A. (1984). Social psychological models of health behavior: An examination and integration. In A. Baum, S. E. Taylor, & J. E. Singer (Eds.). Handbook of psychology andhealth (Vol. 4, pp. 23-54). H&dale. NJ: Erlbaum. Walters,’ V., & Haines, T. (1988). Workers’ perceptions, knowledge, and responses regarding occupational health and safety: A report on a Canadian study. Social Science ondMedicine, II, 1189911%. Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Persona&y and Social Psycho&y. 39.806-820. Weinstein, N. D. (1982). Unrealistic optimism about susceptibility to he&b problems. Journal of Behavioral k4edicine, 5.441-460. Weinstein, N. D. (1987). Unrealistic optimism about illness susceptibility: Conclusions from a community-wide sample. Journal of Behavioral Medicine, 10.481-500. Weinstein, N. D. (1988). The precaution adoption process. Health Psychology, 7,355-386. Weinstein, N. D. (1993). Testing four competing~theories of health-protective behavior. Heabh Psychology, 12(4), 324-333. Weinstein, N. D., & Sandman, P. M. (1992). A model of the precaution adoption process: Evidence from home radon testing. Healrh Psychology, 1113). 170-l 80. Zimolong, B. (1985). Hazard perception and risk estimation in accident causation. In R. E. Eberts & C. G. Eberts @ids.), Trends in ergonomics/human factors II $q~9~3-470). Amsterdam, The Netherlands: NorthZohar, D. (1980). Safety climate in industrial organizations: ‘Iheotetical and applied implications. Journal of Applied Psychology, 65(l), 96-102. Zohar, D., Cohen, A., & Azar, N. (1980). Promoting increased use of ear protectors in noise through information feedback. Human Factors, 22(l), 6979. Journal of Safety Research