Abstract
Knowledge about social and psychological risk factors for initial cigarette smoking experience (ICSE) is sparse. The present study aimed to estimate the prevalence of ICSE and to examine the psychological and social factors related to ICSE. In a cross-sectional survey, 1,511 male college students were recruited using multistage sampling techniques from four universities located within the city of Ilam, Iran. Self-administered multiple-choice questionnaires were distributed to students from March to June 2013. Risk factors for ICSE were evaluated using logistic regression models. Participants were 22.3 ± 2.4 years of age. ICSE prevalence was 30.6%. In multivariable adjusted analysis, risk taking behavior (odds ratio [OR] = 1.61; 95% confidence interval [CI] = 1.11-2.33), perceived peer smoking prevalence (OR = 2.48; 95% CI = 1.03-5.97), positive thoughts about smoking (OR = 1.06; 95% CI = 1.02-1.10), high self-efficacy (OR = 0.95, 95% CI [0.93, 0.98]), presence in smokers’ gathering (OR = 4.45; 95% CI = 2.88-6.81), comity of smokers (OR = 2.56; 95% CI = 1.66, 3.92), very hard access to cigarettes (OR = 2.20; 95% CI = 1.16-4.16), close friends’ medium reaction toward smoking (OR = 1.38; 95% CI = 1.02-1.88), and sporting activity (OR = 0.74; 95% CI = 0.56-0.98) were significantly associated with ICSE. This study identified that a combination of psychological and social variables account for up to 78% of the probability of ICSE. The most important protective factor against ICSE was physical activity, whereas the most important risk factor for ICSE was frequent gathering in the presence of smokers.
Keywords
Introduction
Cigarette smoking is the underlying cause of various serious diseases, making it responsible for the death of 1 in 10 adults worldwide (Danaei et al., 2009; World Health Organization, 2012). It remains controversial as to which factors determine a first cigarette smoking experience. Although there is consensus regarding the adverse effects of cigarette smoking (Durazzo, Mattsson, & Weiner, 2014; Fircanis, Merriam, Khan, & Castillo, 2014), our knowledge about social and psychological risk factors of initial cigarette smoking experience (ICSE) is limited (Otten, Bricker, Liu, Comstock, & Peterson, 2011). A major challenge in smoking prevention programs is to delay ICSE in young individuals because delay of ICSE reduces the future risk of regular smoking (Hiemstra, Otten, & Engels, 2012). Supporting this, studies have reported that early ICSE predicts nicotine dependence among regular smokers (Buchmann et al., 2013; Pomerleau, Collins, Shiffman, & Pomerleau, 1993; Urban, 2010).
The causes of ICSE are manifold and complex, but there is only limited evidence to identify the most important risk factors for ICSE. Based on a thorough literature review without time limitation, only 11 related articles were identified. These studies reported that the perceived prevalence of smoking in peers (Jackson, 1997), pubertal stage (Harrell, Bangdiwala, Deng, Webb, & Bradley, 1998), liking smoking (O’Connor et al., 2005), self-efficacy (Okoli, Richardson, & Johnson, 2008), alcohol consumption, drug use, behavioral undercontrol (Myers, Doran, Trinidad, Wall, & Klonoff, 2009), and pleasant or unpleasant initial smoking experience (Urban, 2010) are strong psychological factors related to ICSE. Meanwhile, peer influence (Alexander, Allen, Crawford, & McCormick, 1999; H. Y. Chang et al., 2011), ethnic and social background (Alexander et al., 1999), family or friends smoking (Bawazeer, Hattab, & Morales, 1999; Okoli et al., 2008), urban or rural residency, and access to cigarettes (H. Y. Chang et al., 2011; Nichols, Birnbaum, Birnel, & Botvin, 2006) are known as social factors related to ICSE. Most studies have focused on risk factors for regular tobacco use, whereas risk factors for ICSE have only been sparsely investigated.
Given the strong relationship between ICSE and future regular smoking, the identification of risk factors for ICSE is an important clinical concern. In an attempt to increase our understanding of the phenomenon of cigarette smoking addiction, an analytical study aiming to estimate the prevalence of ICSE and to examine the effect of psychological and social factors on the odds of ICSE in Iranian male college students was conducted.
Method
Survey Design
The methods used in this survey were similar to those described in previous studies (Nazarzadeh et al., 2013; Nazarzadeh, Bidel, & Carson, 2014). Subjects were male students recruited using multistage sampling techniques from four universities located within the city of Ilam from March to June 2013. Sampling was performed considering university type as the stratum combined with the approximate number of students at each university and the number of classrooms as clusters. All students within the 76 randomly selected classrooms, 2,000 students, were invited to complete the questionnaire.
Trained research assistants visited all 76 classes and explained the study topics to the students, invited them to participate, and provided each student with a questionnaire. Students were informed that participation was optional and that complete as well as partial completion of the questionnaire was acceptable. Furthermore, students were informed that all data would be protected and handled confidentially.
Teachers were present in the classrooms to maintain discipline, but they stayed at their desks to ensure the confidentiality of the responses. Students not willing to complete the questionnaire in the classroom were encouraged to complete it elsewhere and return it to the research assistants within 1 week.
Data Collection
All information related to the smoking stages questionnaire, including design, type of variables, variable scales, reliability, and validity has previously been reported in detail (Alireza Ayatollahi, Mohammadpoorasl, & Rajaeifard, 2005). This questionnaire was designed for measurement of the demographical characteristics, socioeconomic status, smoking stages, drug abuse, as well as psychological and social characteristics related to smoking.
Measures
Cigarette smoking stages were measured using a validated algorithm (Mohammadpoorasl et al., 2013). According to Kaplan, Napoles-Springer, Stewart, and Perez-Stable (2001), students were categorized into never smokers (respondents who have never smokers), experimenter smokers (respondents who have tried cigarette smoking, even a puff, but have smoked less than 100 cigarettes in total), and regular smokers (respondents who have smoked 100 cigarettes or more irrespective of current smoking status). In the present study, regular smokers were excluded and experimenter smokers considered as ICSE.
Psychological variables were measured using the following multiple-choice questions (translated herein from Persian to English). Risk taking behavior was measured according to Kaplan et al. (2001) using the statement “I enjoy doing things that are a little dangerous or risky”; perceived peer smoking prevalence (PPSP) was measured using the question “How many of your same-age peers do you believe to be smokers?”; and self-harm was measured using the question “Have you ever hurt yourself?” The propensity to get involved in fights or quarrels (physical fights or verbal fights) was assessed using the question “Have you been involved in a physical or verbal fight more than five times in the last year?” Self-esteem was measured using the Rosenberg 10-item questionnaire (Nazarzadeh et al., 2014). Positive thoughts about cigarette smoking were measured using the following questions: smoking increases my concentration/reduces my anger/reduces my sorrow/is a sign of maturity/is a sign of independence? Self-efficacy was measured using a 10-item questionnaire concerning self-control and problem solving (Nazarzadeh et al., 2014).
Social characteristics were measured as follows: presence in smokers’ gathering and comity of smokers were measured using direct questions and access to cigarettes was measured using the question “How easy is it for you to obtain cigarettes?” The respondents’ close friends’ reaction to smoking was assessed using the question “If you lit up a cigarette in front of your closest friend, how do you think he/she would react?” Physical activity was defined as being a member of a sports team, going to the gym, or doing other regular sporting activities.
The study protocol, including the questionnaire, was approved by the Ethics Committee of Ilam University of Medical Sciences.
Statistical Methods
Comparison of continuous variables between never smokers and experimenter smokers was performed with an unpaired t test. Distributions of discrete variables were compared across smoking status using the chi-square test. For discrete variables, the number of items reported and the relevant Cronbach’s alpha of internal consistency was calculated.
First, unadjusted (crude) odds ratios (ORs) with 95% confidence intervals (CIs) were calculated using binary logistic regression for each variable category (psychological and social). Afterwards, multivariable logistic regression models were made for psychological and social variables, separately. Finally, all variables were included in a combined model. The goodness of fit of the logistic models was assessed using the Hosmer–Lemeshow Test, as well as Cox and Snell R2 and Negelkerke R2 were calculated. Meanwhile, receiver operating characteristic curves and area under the curves were calculated to assess the goodness of fit of the models using the predicted probability of the models. P ≤ .05 was considered statistically significant. All statistical analyses were performed using the SPSS software package version 19 for Windows.
Results
Student Characteristics
Of 2,000 students invited to participate in the study, 1,824 (91.2%) responded to the questionnaire. After exclusion of all regular smokers (n = 302, 16.6%) and 11 respondents with missing data on smoking status, a total of 1,511 respondents were considered for further analysis.
The final study cohort had a mean age of 22.3 ± 2.4 years. The prevalence of ICSE was 30.6%, and their psychological and social characteristics are presented in Tables 1 and 2. In terms of psychological characteristics, all measured variables were significantly different between never smokers and students with ICSE (all p < .001). Students with ICSE were more likely to display risk taking behavior, report previous deliberate self-harm, have more PPSP, quarrel more than five times, have a low self-esteem level, and have more positive thoughts about smoking. In terms of social characteristics, all measured variables differed significantly between never smokers and people with ICSE except that access to cigarettes was equal (p < .65). Furthermore, students with ICSE tended to be less physically active than never smokers (p < .07). In terms of social characteristics, students with ICSE had further presence in smokers’ gathering (p < .001) and had further exposure to comity of smokers (p < .001), had a more positive reaction about smoking from close friends (p < .001), and had no sporting activity (p < .07).
Prevalence of Initial Cigarette Smoking Experience Across Psychological Characteristics of Male College Students.
Note. PPSP = perceived peer smoking prevalence; α = Cronbach’s alpha (reliability index).
p calculated using chi-square test.
p calculated using unpaired t test.
Prevalence of Initial Cigarette Smoking Experience Across Social Characteristics of Male College Students.
Note. PSG = presence in smoker’s gathering.
p was calculated using chi-square test.
Logistic Regression Modeling and Goodness of Fit
Reported in Table 3 are the two separate univariable and multivariable models for each category of variables to calculate the unadjusted and adjusted OR and 95% CI for people with ICSE compared with never smokers. Variables (categories) expected to be associated with nonsmoking were considered as the reference group. In the multivariable analysis for prediction of ICSE, the base model (multivariable Model 1 in Table 3) included the psychological characteristics that are presented in Table 1. According to the adjusted Model 1, risk taking behavior (OR = 1.55), self-harm (OR = 1.60), PPSP (OR of very high vs. very low = 3.01), quarrelling (OR = 1.93), and positive thoughts about smoking (OR = 1.07) were significantly associated with an increased risk of ICSE; and self-efficacy (OR = 0.95) was significantly associated with a reduced risk of ICSE. The models were well fitted as indicated by nonsignificant Hosmer–Lemeshow goodness-of-fit tests (p = .19). According to the Cox and Snell R2 as well as the Negelkerke R2, 23% and 30% of the observed variance was explained by the psychological model.
Odds Ratios (ORs) and 95% Confidence Intervals (CIs) of Initial Cigarette Smoking Experience of Male College Based on Two Separate Binary Logistic Regression Models.
Note. PPSP = perceived peer smoking prevalence; Ref = reference category; PSG = presence in smoker’s gathering.
Hosmer–Lemeshow Test: chi-square = 11.13; p = .19; −2 log likelihood = 1590.1; Cox and Snell R2 = 0.23; Negelkerke R2 = 0.30.
Hosmer–Lemeshow Test: chi-square = 45.77; p < .001; −2 log likelihood = 1638.4; Cox and Snell R2 = 0.24; Negelkerke R2 = 0.32.
Results of Model 2, which was based on social variables, are reported in Table 3. This model indicated that presence in smokers’ gathering (OR = 4.36), comity of smokers (OR = 2.84), limited access to cigarettes (OR = 3.48), and light reaction of close friend to smoking (OR = 2.01) were significant risk factors for ICSE. The models were not well fitted as indicated by significant Hosmer–Lemeshow goodness-of-fit tests (p < 0.001). Cox and Snell R2 as well as Negelkerke R2 showed that 24% and 32% of the observed variance was explained by the social model.
Model 3 included all variables and reported that risk taking behavior (OR = 1.61), PPSP (OR = 2.48), positive thoughts about smoking (OR = 1.06), self-efficacy (OR = 0.95), presence in smokers’ gathering (OR = 4.45), comity of smokers (OR= 2.56), very limited access to cigarettes (OR = 2.20), close friends’ medium reaction (OR = 1.38), and sporting activity (OR = 0.74) were significantly associated with the risk of ICSE. The models were well fitted as indicated by nonsignificant Hosmer–Lemeshow goodness-of-fit tests. Cox and Snell R2as well as Negelkerke R2 showed that 21% and 29% of the observed variance was explained by psychological or social variables (see Table 4).
Adjusted Odds Ratios (ORs) and 95% Confidence Intervals (CIs) of Initial Cigarette Smoking Experience of Male College Students for Related Psychological and Social Factors, Based on Binary Logistic Regression Model.
Note. PPSP = perceived peer smoking prevalence; PSG = presence in smoker’s gathering.
Hosmer–Lemeshow Test: chi-square = 7.62; p = .47; −2 log likelihood = 1368.0; Cox and Snell R2 = 0.21; Negelkerke R2 = 0.29.
The receiver operating characteristic curves and area under the curve for assessing the percentage of predicted probability that is explained by each model is exhibited in Figure 1. According to this figure, 69%, 72%, and 78% of the predicted probability was explained by Models 1, 2, and 3, respectively.
Receiver operating characteristic (ROC) curve analysis to assess goodness of fit for three binary logistic regression models.
Discussion
The effects of psychological and social factors on the path that leads to ICSE were examined. Overall, the current study identified that psychological and social variables are strongly associated with the risk of a first time cigarette smoking experience. In particular, this study has identified a number of specific risk factors, of which some are changeable, whereas others are not.
These results are in agreement with previous studies suggesting that risk taking behavior is strongly associated with smoking stages and regular smoking (Mohammadpoorasl, Nedjat, Fakhari, et al., 2012; Mohammadpoorasl, Nedjat, Yazdani, et al., 2012; Nazarzadeh et al., 2013). In these studies, the same questionnaire had been used to demonstrate that risk taking behavior emerges as an independent predictor of ICSE. Importantly, these studies had a cross-sectional design, which implies that exposure and outcome data were collected at the same time. Therefore, the present study cannot firmly conclude whether risk taking behavior is the actual cause of ICSE. Longitudinal cohort studies are warranted to clarify this matter.
An unanticipated finding was that “very limited access to cigarettes” increased twofold the odds of ICSE. This finding may reflect the very equal access to cigarettes in the Iranian community. In Iran, cigarettes are generally easy to obtain, no matter the age, and according to Table 2, most students had easy access to cigarettes. Moreover, given the low number of respondents with hard or very hard access to cigarettes, the unanticipated findings may be due to chance.
In terms of social variables, close friend’s light reaction to smoking, comity of smokers, and presence in smokers’ gathering were strong risk factors for ICSE. Presence in smokers’ gathering more than 10 times was the strongest risk factor for ICSE, and these individuals were four times more likely to report ICSE. In previous longitudinal studies, various measures of peer bonding have been employed, including attachment to peers, agreement with peers (Krohn, Massey, Skinner, & Lauer, 1983), association with friends who were regular smokers (Skinner, Massey, Krohn, & Lauer, 1985), and number of friends (Brunswick & Messeri, 1983). Also, a systematic review reported that 11 of 15 studies conducted before 1992 reported a strong influence of peers on ICSE in adolescents (Conrad, Flay, & Hill, 1992). Our study is the first to report that presence in smokers’ gathering is significantly associated with ICSE in young adults. Moreover, this association seems to follow a dose-relationship pattern implying that the risk of ICSE increases with the number of exposures to this environment. A plausible explanation for this is that college students are not yet independent enough to resist a social pressure toward smoking (Katz, Robisch, & Telch, 1989). Accordingly, the aim of many smoking prevention programs is to improve the target group’s ability to resist such pressure (Botvin & Eng, 1982; Evans, 1976; Katz et al., 1989; Kreuter et al., 2014).
Physical activity often plays an important role in smoking prevention programs (Escobedo, Marcus, Holtzman, & Giovino, 1993). In agreement with previous studies, the current study identified that physical activity is a strong protective factor against ICSE (Escobedo et al., 1993; Rodriguez Garcia, Lopez Villalba, Lopez Minarro, & Garcia Canto, 2014). A study on American college students reported that students doing no regular exercising had an OR of 1.33 of having smoked within the last 3 months (Halperin, Smith, Heiligenstein, Brown, & Fleming, 2010). The present study identified that physical activity was negatively associated with the risk of ICSE (OR = 0.74). Thus, there is strong evidence that physical activity affects smoking patterns, but mechanisms driving this association remain elusive.
A large-scale study in Iran previously reported that having a positive attitude toward cigarette smoking in adolescents was associated with having the intention of smoking (OR = 1.30; Mohammadpoorasl, Nedjat, Yazdani, et al., 2012). A similar trend in the current study was identified, although the association was weaker (OR = 1.06), which may be due to the differences in study populations. College students are more mature and have a greater knowledge about health than teenagers, and consequently they may be less susceptible to positive thoughts about smoking. The current study’s findings are supported by other studies (Andrews & Duncan, 1998; Sargent & DiFranza, 2003; Tyas & Pederson, 1998).
Ordinal logistic regression model indicate that a lower risk of ICSE could be expected with increasing self-efficacy. This is in line with previous reports (Bidstrup et al., 2009; F. C. Chang et al., 2006; Lotrean, Dijk, Mesters, Ionut, & De Vries, 2010). Moreover, it has been reported that a decrease in self-efficacy is associated with early smoking onset, even after adjustment for possible confounders (Hiemstra et al., 2012). Hiemstra et al. (2012) reported that low refusal self-efficacy was associated with an increased 1-year risk of smoking debut in adolescents. In conclusion, self-esteem and self-efficacy should be considered as important elements in smoking prevention programs.
Knowledge about psychological factors such as PPSP in the community is limited, especially in young people and adults. The current study identified that high levels of PPSP increase the risk of ICSE threefold. More research on this topic is warranted.
The current cross-sectional study should be considered hypothesis generating. Based on the study results, the authors suggest the performance of longitudinal cohort studies to further explore determinants of ICSE and to more precisely define the temporal relationship between the measured variables. Meanwhile, meta-analyses and systematic review studies may be useful.
A number of limitations should be considered when interpreting the results of the current study: (a) temporal relationships cannot be inferred because of the cross-sectional design; (b) self-administrated questionnaires were used for data collection, which implies a risk of response bias and false negative result that may lead to misclassification and a reduction of OR estimates; (c) the study cohort was restricted to male students, and the results might not reflect smoking habits of female students.
Conclusion
The current study reported that ICSE prevalence in college students was 30.6%. Determinants of ICSE were risk taking behavior, PPSP, positive thoughts about smoking, self-efficacy, presence in smokers’ gathering, comity of smokers, friends’ reaction to smoking, and physical activity. Overall, the study identified that a combination of psychological and social variables accounts for up to 78% of the probability of ICSE. The most important protective factor against ICSE was physical activity, while the most important risk factor for ICSE was frequent gathering in the presence of smokers.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was financially supported by the Ilam University of Medical Sciences, and under a research grant (Grant No. 908668) from the Prevention of Psychosocial Injuries Research Center.