Adolescent daily cigarette smoking : is rural residency a risk factor ?

Introduction: Daily cigarette smoking among US adolescents remains a significant public health problem. Understanding risk is important in order to develop strategies to reduce this type of tobacco use. Purpose: The primary objective of this research was to examine whether rural residency is an independent risk factor for being a daily smoker among adolescents ages 12 to 18 years. Methods: This is a cross-sectional study where univariate, bivariate, and multivariate analyses were performed on a merged 19972003 Youth Risk Behavior Surveillance System dataset to determine whether rural residence was a significant risk factor for daily cigarette smoking, after adjusting for demographic factors. Results: Using daily smoking as the dependent variable, initial multivariate analyses revealed that adolescents who lived either in suburban (OR=.34, CI=.32, .36) or urban (OR=.33, CI=.31, .35) locales were less likely to become daily smokers than adolescents living in rural locales. Subsequent logistic regression analysis yielded that rural youths who became daily smokers were more likely to: have used smokeless tobacco products in the past 12 months (OR=1.25, CI=1.04,1.51); be female (OR=1.42, CI=1.23,


Introduction
The rates of smoking in the US are now at their lowest levels since World War II 1,2 . In addition to a drop in the percentage of adults who currently smoke, the percentage of high school students who reported smoking in the past month decreased from 35% in 1999 to 23% in 2005 3  Because rural residency is associated with higher smoking prevalence estimates among US adults 4 , it may also be a risk factor among the adolescent population.
Among the many studies examining youth cigarette smoking as tobacco use [4][5][6] , most studies that stratified adolescent smoking by rural, suburban, and urban residence were conducted outside the United States of America [7][8][9][10][11][12][13][14][15][16] . These international studies often yielded conflicting findings. For instance, a study of seventh grade Chinese students found no significant differences between rural and urban adolescent smoking rates 8 , while in Sweden researchers found significantly increased smoking in rural eighth grade students compared with students living in urban areas 14 . In contrast, the results of a cross-sectional study conducted in Scotland demonstrated higher rates of lung cancer among urban residents due to a higher probability of active tobacco smoking in the USA 10,11 .
Most US studies on youth smoking that included rurality in the analysis used either regional or otherwise limited data sets [4][5][6] . Although these studies have suggested an overall higher prevalence of smoking among rural adolescents 6,12 , the limitations of these smaller studies leave unanswered the question about whether rurality is a risk factor for youth tobacco use.
Tobacco studies also use different variables, such as experimenter, current social smoker, daily smoker or former smoker, to define tobacco use [17][18][19][20][21][22][23] . For the adolescent smoker, the evolution from experimenter to daily user is a critically important shift. A recent study 17 found that the variables that predict a shift from experimental to regular smoking differ from those that predict smoking onset.
Likewise, another study 18 comparing regular smokers with social smokers also identified variables that differed between the two groups. Neither of these studies included rurality or place of residence in their analyses, leaving a gap in the literature on the impact of residence on the risk of an adolescent becoming a regular or daily smoker.
Using a national data base, this study investigated the question of rurality and its impact on being a daily or regular smoker. Specifically, we explored the question of whether rural residency was an independent risk factor for being a regular/daily smoker among adolescents aged 12 to 18 years.
In addition, we sought to characterize rural adolescents who became regular/daily cigarette smokers. Finally, we also explored the issue of how tobacco use in adolescents changed over the study period, and whether this change differed according to place of residence. Developed in 1990, the purpose of the YRBSS is to determine the prevalence of health-risk behaviors among high-school students in order to assess whether these behaviors increase, decrease, or stay the same over time, and to examine the co-occurrence of health-risk behaviors. It is also used to provide comparable national, state, and local data as well as comparable data among sub-populations of youth. A more detailed discussion of YRBSS sampling strategies and data collection methodologies can be found elsewhere 24 .

Methods
The data made available in the weighted national YRBSS datasets minimally have a 60% response rate from the state of collection. For this analysis we merged the relevant 4 years of survey data (collected once every 2 years) into a single dataset, and re-coded a number of the factors of the original variables. Multiple years of the dataset were merged in order to ensure a sufficient number of observations of rural adolescents. The re-coding was assisted analytical ease, and also standardized the responses across the multiple years of survey data examined in this study. Originally our intent was to include more years of survey data than the 1997 to 2003 range; however, specific variables of interest for this study, such as geographic place of residence (urban, suburban and rural), were not included in the datasets outside this range. The specific survey questions and recoded categories are presented (Table 1). In addition, CDC provides a weighting variable for analysis. We used this weighting variable in the analyses conducted here to ensure the results represented the youth population of the USA.

Discussion
This study used YRBSS data sets from 1997 to 2003 to compare the tobacco use behaviors of rural youths with their non-rural counterparts. The results indicated that rural high school students were significantly more likely than metropolitan and suburban youths to both try tobacco products and to become regular smokers. This was not surprising because it confirmed the findings of other smaller, regional studies that rural residence was a risk factor for tobacco use in younger individuals 12 and that rural adults are more likely to smoke than their non-rural counterparts. An encouraging finding was that, over the study period,