استفاده از تئوری انگیزش حفاظتی در پیش بینی مصرف سیگار در نوجوانان در چین

Reducing tobacco use among adolescents in China represents a significant challenge for global tobacco control. Existing behavioral theories developed in the West — such as the Protection Motivation Theory (PMT) — may be useful tools to help tackle this challenge. We examined the relationships between PMT factors and self-reported cigarette smoking behavior and intention among a random sample of vocational high school students (N = 553) in Wuhan, China. Tobacco-related perceptions were assessed using the PMT Scale for Adolescent Smoking. Among the total sample, 45% had initiated cigarette smoking, and 25% smoked in the past month. Among those who never smoked, 15% indicated being likely or very likely to smoke in a year. Multiple regression modeling analysis indicated the significance of the seven PMT constructs, the four PMT perceptions and the two PMT pathways in predicting intention to smoke and actual smoking behavior. Overall, perceived rewards of smoking, especially intrinsic rewards, were consistently positively related to smoking intentions and behavior, and self-efficacy to avoid smoking was negatively related to smoking. The current study suggests the utility of PMT for further research examining adolescent smoking. PMT-based smoking prevention and clinical smoking cessation intervention programs should focus more on adolescents' perceived rewards from smoking and perceived efficacy of not smoking to reduce their intention to and actual use of tobacco.

Tobacco control in developing countries presents a pressing global public health challenge (Giovino et al., 2012 and World Health Organization, 2011). More than 80% of smokers are in developing countries. Smokers in China alone account for a third of the world total (Giovino et al., 2012). Despite the high rates of smoking in China and other developing countries, few evidence-based tobacco prevention and intervention studies have been conducted (Shek & Yu, 2011). Moreover, most studies that have been conducted are not guided by an established behavioral theory (see Guo et al., 2007 for a notable exception). Most behavioral theories are developed in the West. Promoting the application of Western-developed behavioral theories in China and other developing countries may advance the understanding of tobacco use for more effective tobacco control across the globe. 1.1. Theory-based behavioral research Contemporary medical and health behavior research has become increasingly grounded in and guided by theory (Glanz et al., 2008, National Research Council, 2001 and Otten et al., 2011). A large number of theories and models have been developed and used to guide behavioral research including the Information–Motivation–Behavioral Theory (Fisher and Fisher, 1992 and Fisher et al., 1996), the Theory of Reasoned Action and Planned Behavior (Ajzen, 1985, Ajzen and Fishbein, 1980 and Fishbein and Ajzen, 1975), the Transtheoretical Model of Change (Prochaska & Velicer, 1997), Social Cognitive (Learning) Theory (Bandura, 1989), and Protection Motivation Theory (Boer and Seydel, 1996 and Rogers, 1975). With the guidance of a theory or a conceptual framework, a researcher can better investigate and understand smoking behaviors, supporting more effective tobacco use prevention and smoking cessation. In addition, data derived from a theory-guided study will help theorists to refine the proposed theory. To date, little research on smoking in China has been conducted and only a small subset is theory-based. 1.1.1. Protection Motivation Theory and health behavior Among the various theories that are frequently used to guide behavior research, the Protection Motivation Theory (PMT) may be particularly well-suited for understanding and addressing smoking behavior (Floyd et al., 2000 and Milne et al., 2000). First elucidated by Rogers (1975), PMT in its current format posits two closely related pathways, Threat Appraisal and Coping Appraisal, which link environmental influences to a behavior through a series of cognitive processes (Fig. 1). Full-size image (14 K) Fig. 1. Schematic representation of the Protection Motivation Theory (PMT) components. Figure options The balance between the two appraisal pathways determines the likelihood for a risk behavior, such as smoking, to occur (Boer and Seydel, 1996 and Rogers, 1975). The threat appraisal pathway involves comparing perceived rewards (intrinsic and extrinsic) of a maladaptive health-related behavior (e.g., smoking) with perceived threats (severity and vulnerability) that the behavior poses. For example, adolescents might weigh feelings of relaxation and better concentration (potential perceived intrinsic rewards of smoking) and beliefs that happier and more popular smoke (potential perceived extrinsic rewards of smoking), against their knowledge that smoking causes cancer and other diseases (potential severity of smoking-related risk) and concerns that smoking may lead to an earlier death (potential vulnerability to smoking-related risk). The coping appraisal pathway involves comparing coping efficacy (self-efficacy and response efficacy) of an adaptive variant of the health-related behavior (e.g., avoiding smoking) with perceived response costs of such adaptive behavior. For example, adolescents might consider the health benefits that nonsmokers may enjoy (perceived response efficacy of not smoking) and how well they think they could decline a cigarette offered by a friend (self-efficacy for not smoking), as compared with their concerns about social isolation if they do not smoke (perceived cost of not smoking). Application of PMT to research has advanced our understanding of a number of health behaviors, including alcohol consumption (Gibbons, Houlihan, & Gerrard, 2010), substance use and abuse (Feigelman et al., 1995 and Stanton and Galbraith, 1994), and sexual risk behavior for HIV infection (Chen et al., 2009, Chen et al., 2010 and Gong et al., 2009). As described in the next section, PMT has been used as a framework for tobacco research in the United States. We propose to extend the application of PMT to smoking behavior in China. The approach commonly seen in the reported studies to examine the utility of a PMT model is either to treat it as a whole or to focus on certain significant constructs (e.g., self-efficacy, response cost) (Chen et al., 2009). In this study, we will investigate PMT constructs hierarchically. Our intention is to provide data on the role of individual PMT constructs, perceptions, and appraisal pathways as an integrative system in predicting smoking behavior. 1.1.2. PMT and adolescent tobacco use research As an integrative conceptual framework, PMT has also been used in tobacco research among adolescents in the United States and other developed countries (Costa et al., 2007, Greening, 1997 and Ho, 1998). For example, guided by PMT, one early epidemiological study demonstrated the significance of cognition and appraisal in predicting adolescents' smoking behavior (Greening, 1997). In this sample of 690 American high school students, a logistic regression controlling for age and gender revealed that smokers were more likely to acknowledge the vulnerability to smoking-related diseases, perceived less severity of smoking risks, viewed male smokers as popular (an indicator of extrinsic rewards of smoking, an effect stronger for boys than for girls), and perceived lower response efficacy for not smoking than did nonsmokers. Anti-tobacco intervention programs based on PMT have also shown significant effects in discouraging tobacco use among adolescents in developed countries (Pechmann et al., 2003 and Smith and Stutts, 2003). For example, perceived social disapproval increased adolescents' intentions to abstain from smoking (Pechmann et al., 2003). The successful application of PMT in tobacco research in developed countries suggests the potential utility of the theory to guide tobacco research in developing countries such as China. However, few such studies can be found in the published literature in China. 1.2. Need for theory-based tobacco research in China As part of the global tobacco control effort, reducing tobacco use in China represents a critical challenge. China is the world's largest tobacco producer and consumer, and is home to approximately one-third of the world's smokers (Han et al., 2013 and World Health Organization, 2011). Despite much progress in tobacco control in China, data from diverse sources indicate little decline in the prevalence of tobacco use, particularly tobacco use among adolescents (Chen, Stanton, et al., 2006, Cheng, 2008, Han et al., 2013 and Zhang et al., 2011). Every day, approximately 80,000–90,000 Chinese adolescents 12–17 years of age become new smokers and 11–12 million adolescents smoked in the past month (Grenard et al., 2006, Han et al., 2013 and Weiss et al., 2008). Thus, it is a significant and timely challenge for tobacco researchers in China and across the globe to curb the high rates of tobacco use among young Chinese. There have been notable tobacco control efforts in China, but very few studies or intervention programs have been grounded in specific theory. This lack of theory-based research could be one potential explanation for the persistently high levels of tobacco use among young Chinese after several decades of tobacco control efforts. A search of the published literature in Chinese and English revealed that few studies on tobacco use in Chinese adolescents were theory-guided, and all were conducted by the same research team (e.g., Guo et al., 2012, Guo et al., 2013 and Guo et al., 2007). In addition, theories were mentioned in only three reported smoking prevention intervention trials conducted in China (Chen, Fang, Li, Stanton and Lin, 2006, Chou et al., 2006 and Wen et al., 2010). Furthermore, even in these intervention trials, no detailed information was provided about integrating the theories into the intervention programs, and no evaluation data (e.g., mediation effect analysis) were reported to support the role that the theory played in supporting the observed intervention effect. 1.3. Purposes and hypotheses of this study In the present study, we were guided by the PMT to predict tobacco use intention and behavior among adolescents in China. The purpose of this study is to demonstrate the utility of PMT, a Western-based and empirically tested behavioral theory, to advance our understanding of tobacco use behavior among Chinese adolescents. In addition, findings of this study will provide data supporting tobacco use prevention in China, with potential to extend to other developing countries, further facilitating global tobacco control efforts. As suggested by PMT principles, we hypothesized that smoking would be negatively associated with perceived threat (i.e., severity and vulnerability) of smoking effects and perceived efficacy (i.e., self-efficacy and response efficacy) of not smoking. Further, we hypothesized that smoking would be positively associated with perceived rewards (i.e., intrinsic and extrinsic rewards) of smoking and perceived costs of not smoking.

3.1. Preliminary analyses Just under half (45%, n = 251) of students indicated that they had ever smoked. 2 In the past month, 74% (n = 393) of students indicated not smoking, 12% (n = 63) reported smoking occasionally, 7% (n = 39) indicated weekly smoking, and 6% (n = 33) reported daily/near daily smoking. Of the approximately one-quarter of students who smoked in the past month, 35% (n = 46) smoked one cigarette, 52% (n = 69) smoked 2–5 cigarettes, and 14% (n = 18) smoked more than 5 cigarettes per day that they smoked. Further, 17% of students (n = 94) believed it likely and 5% (n = 25) believed it very likely that they would be smoking in one year. 3.2. Correlation analyses 3.2.1. Demographic relationships Table 1 includes descriptive statistics and correlations among study variables. Boys were significantly more likely than girls to intend to smoke in one year and to have smoked in the past month; among past-month smokers, boys also smoked more cigarettes per day than girls. Girls indicated higher vulnerability and severity of smoking-related health issues, lower rewards for smoking, higher efficacy for not smoking, and lower costs of not smoking than boys. Neither age nor school performance was significantly associated with any of the smoking measures nor with the PMT constructs, with the exception of the significant association between younger age and higher response efficacy. Table 1. Descriptive statistics of and correlations among study variables. 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Gender – 2. Age − .07 – 3. School performance .02 .21⁎⁎⁎ – 4. Smoking intention − .31⁎⁎⁎ − .06 − .08 – 5. Frequency smoked in past month − .43⁎⁎⁎ .02 − .11⁎ .57⁎⁎⁎ – 6. Number cigarettes per day − .23⁎⁎ .02 − .08 .25⁎⁎ .61⁎⁎⁎ – 7. Vulnerability .08⁎ − .03 .01 − .16⁎⁎⁎ − .10⁎ − .02 – 8. Severity .14⁎⁎⁎ .04 –.01 − .23⁎⁎⁎ − .20⁎⁎⁎ − .06 .52⁎⁎⁎ – 9. Extrinsic rewards − .26⁎⁎⁎ .01 − .03 .24⁎⁎⁎ .36⁎⁎⁎ .13 − .05 − .11⁎⁎ – 10. Intrinsic rewards − .22⁎⁎⁎ .01 − .07 .31⁎⁎⁎ .37⁎⁎⁎ .29⁎⁎⁎ − .08⁎ − .22⁎⁎⁎ .61⁎⁎⁎ – 11. Response efficacy .12⁎⁎ − .09⁎ .01 − .07 − .14⁎⁎ .02 .24⁎⁎⁎ .28⁎⁎⁎ − .15⁎⁎⁎ − .11⁎ – 12. Self-efficacy .17⁎⁎⁎ − .01 .08 − .19⁎⁎⁎ − .29⁎⁎⁎ − .10 .18⁎⁎⁎ .29⁎⁎⁎ − .17⁎⁎⁎ − .23⁎⁎⁎ .41⁎⁎⁎ – 13. Response cost − .28⁎⁎⁎ .06 − .01 .24⁎⁎⁎ .25⁎⁎⁎ .15 − .12⁎⁎ − .23⁎⁎⁎ .42⁎⁎⁎ .44⁎⁎⁎ − .12⁎⁎ − .27⁎⁎⁎ – N 553 553 550 538 528 133 551 551 551 550 550 550 550 M 1.50 16.28 3.20 1.61 1.45 1.79 4.76 5.88 2.71 2.36 5.14 5.94 2.58 SD .50 .98 .92 .93 .88 .66 1.37 1.36 1.42 1.44 1.70 1.40 1.41 Notes. Gender was coded as 1 = boys, 2 = girls. Number of cigarettes smoked per day was restricted to students who had smoked in the past month (n = 133). ⁎ p < .05. ⁎⁎ p < .01. ⁎⁎⁎ p < .001. Table options 3.2.2. Relationships of smoking measures with PMT variables Stronger smoking intentions were significantly associated with more frequent past-month smoking, smoking more cigarettes per day, lower vulnerability and severity, higher extrinsic and intrinsic rewards, lower self-efficacy, and higher response cost. Past month smoking frequency was associated with PMT constructs in the same manner, and frequency was also significantly associated with lower response efficacy of not smoking. Number of cigarettes smoked only included those who indicated smoking in the past month (n = 133); more cigarettes smoked was significantly associated with stronger intrinsic rewards for smoking. PMT constructs were moderately intercorrelated. Most of the strongest relationships were among pairs of variables forming perception scores (i.e., r = 0.52 for vulnerability and severity, r = 0.61 for extrinsic and intrinsic rewards, and r = 0.41 for self-efficacy and response efficacy). 3.3. Multiple regression models Table 2 contains the standardized estimates for multiple regression models predicting smoking intentions and behaviors from PMT variables. Gender and age were included as control variables in all models. Table 2. Regression models of smoking variables regressed on PMT variables. PMT constructs Smoking intentions Frequency smoked in past month Number of cigarettes per day Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Model 1 Model 2 Model 3 Threat appraisal .27⁎⁎⁎ .25⁎⁎⁎ .15 Perceived threat − .15⁎⁎⁎ − .04 .01 Vulnerability − .08 .01 − .01 Severity − .10⁎ − .05 .08 Perceived rewards .20⁎⁎⁎ .30⁎⁎⁎ .21⁎ Extrinsic rewards .04 .14⁎⁎ − .07 Intrinsic rewards .18⁎⁎⁎ .18⁎⁎⁎ .32⁎⁎ Coping appraisal − .05 − .09⁎ .03 Perceived efficacy − .01 − .12⁎⁎ − .02 Response efficacy .05 .02 − .01 Self-efficacy − .07 − .17⁎⁎⁎ − .06 Perceived cost .05 − .01 .03 Response cost .04 − .03 .02 R2 0.19 0.18 0.18 0.31 0.29 0.27 0.13 0.10 0.08 Notes. Multiple linear regression was used. Models for number of cigarettes per day were restricted to students who had smoked in the past 30 days (n = 133). Model 1 includes the 7 PMT construct scores. Model 2 includes the 4 PMT perception scores. Model 3 includes the 2 PMT pathway scores. All models controlled for age and gender. ⁎ p < .05. ⁎⁎ p < .01. ⁎⁎⁎ p < .001. Table options 3.3.1. Smoking intentions The first set of results in Table 2 shows the standardized estimates of the multiple linear regression models predicting intention to smoke in one year. Model 1, in which all seven PMT constructs were entered, was significant, F(9, 525) = 14.09, p < .001. Stronger intentions were significantly associated with higher intrinsic rewards for smoking and lower perceived severity of health issues from smoking. Model 2, in which the four PMT perceptions were entered, was also significant, F(6, 528) = 19.87, p < .001. The greater the perceived rewards from smoking and the lower the perceived threat that smoking poses, the greater the intention to smoke. Model 3, with smoking intentions regressed onto the two appraisal pathways, was significant, F(4, 530) = 29.64, p < .001. Greater intentions to smoke were significantly associated with higher threat appraisals (more rewards than threats). The estimated R2 indicated that these models accounted for 18–19% of the variance in smoking intentions. 3.3.2. Frequency of smoking in the past month The second set of models shown in Table 2 provides estimates of the multiple linear regression models predicting past-month smoking. Model 1, in which all seven PMT constructs were entered, was significant, F(9, 515) = 25.46, p < .001. More frequent smoking in the past month was significantly associated with higher perceived intrinsic and extrinsic rewards of smoking and lower self-efficacy for not smoking. Model 2, in which the four PMT perceptions were entered, was also significant, F(6, 518) = 35.85, p < .001. Similar to the construct model (Model 1), more frequent past-month smoking was significantly associated with higher perceived rewards of smoking and lower perceived efficacy for not smoking. Model 3, with smoking intentions regressed onto the two appraisal pathways, was significant, F(4, 520) = 47.68, p < .001. Smoking more often was associated with significantly higher threat appraisals and lower coping appraisals. Model 3 suggests that smoking more frequently is related to believing more strongly that the rewards from smoking outweigh the potential risks. Further, more frequent recent smoking was associated with higher costs of not smoking outweighing perceived efficacy of not smoking. These models accounted for 27–31% of the variance in past-month smoking frequency. 3.3.3. Number of cigarettes smoked per day The last set of models shown in Table 2 provides estimates of the multiple linear regression models predicting the number of cigarettes smoked per day among past-month smokers. Model 1, in which all seven PMT constructs were entered, was significant, F(9, 122) = 2.11, p < .05. The stronger the intrinsic rewards from smoking, the more cigarettes smoked per day. Model 2, in which the four PMT perceptions were entered, was also significant, F(6, 125) = 2.31, p < .05. The greater the perceived rewards from smoking, the more cigarettes smoked. Model 3, with smoking intentions regressed onto the two appraisal pathways, was significant, F(4, 127) = 2.70, p < .05. Although the model as a whole was significant, neither pathway reached significance. These models accounted for 8–13% of the variance in daily cigarette consumption.