The Effect of Depression Status Change on Daily Cigarette Smoking Amount: a eleven-year follow up study of the Korea Welfare Panel Study


 Background: In the past decade, the Korean smoking rate has only decreased by 3%, despite several smoking control policies. There is a need for such policies to take smokers’ psychological characteristics into account. Depression is a well-known contributor to failed smoking cessation. This study aimed to examine the effect of smokers’ depression status changes on their daily cigarette smoking amount (DCA).Methods: This study used a sample drawn from the Korea Welfare Panel Study (KoWePS) waves 3 (2008) to 13 (2018). The DCA refers to the number of the cigarettes smoked per day at the time of the survey. Depression was measured using an 11-item version of the Center for Epidemiologic Studies Depression Scale (CESD-11). A generalized estimating equation (GEE) model was employed to analyse the effect of change of depression status on DCA.Results: The Yes→No male depression status group had lower DCA (β=-0.631, p-value=0.0248) than the No→No group. The Yes→No male depression status group that began smoking before age 19 had lower DCA (β=-0.881, p-value: 0.0089) than the No→No group that started smoking before 19.Conclusions: We found that a change from depressed to non-depressed and non-depressed to depressed status is associated with decreasing and increasing DCA, respectively. Also, for smokers who began smoking before 19 years of age, the subgroup that went from depressed to non-depressed had much a lower DCA than general smokers. Thus, when treating people participating in smoking cessation programs, counsellors should check for depression symptoms and encourage individuals to pursue depression treatment simultaneously.


Introduction
In 2010, 8,284 people died of either lung cancer or ischemic heart disease due to ongoing smoking [1]. The Korea National Health and Nutrition Examination Survey (KNHANES) calculated the national smoking rate as 22.3% in 2017, with 38.1% of men falling under the 'current smoker' category. Moreover, the past decade has only seen a 3% smoking decrease, down from 25.3%.
Since 1982, the Korean government has implemented various smoking control policies, such as regulating cigarette advertisements, adding warning labels to cigarette packs, restricting the sale of cigarettes to adolescents, banning smoking in public places and restaurants, and increasing cigarette prices [2]. Clearly, smoking control policies have not translated to a signi cant change in the national smoking prevalence. This indicates a need for research and policy development informed by both the sociodemographic and psychological characteristics of smokers.
To predict the success or failure of smoking cessation attempts, researchers have focused on the mitigating role of 'major depressive disorder' or 'depressed mood'. A history of regular smoking was observed more frequently among those who experienced major depressive disorder at some point in their lives than it was among those who had not (non-depressed individuals) [3]. Some studies report that the history of major depressive disorder might imply the failure of smoking cessation [4,5]. In addition, the relationship between depressed mood and past major depressive disorder has been supported in many studies on smokers [6,7]. On the contrary, attempts to cease smoking, or successful smoking cessation may increase one's risk of experiencing a major depression episode, irrespective of past major depressive disorder history [8,9].
Self-reported daily cigarette amount (DCA) is an important factor used by many studies to assess the level of cigarette dependence. DCA has been used to determine what quali es as 'heavy' dependence, or categorize those deemed 'hardcore smokers' [10][11][12]. Some studies have used DCA as a predictor for successful smoking cessation or of attempting cessation. Indeed, Hymowitz reported that individuals with lower daily cigarette consumption levels were statistically signi cant predictors of smoking cessation [13]. Similarly, Hyland found that individuals who reduced their smoking by more than 50% over 5 years were 1.7 times more likely to quit smoking in the 13 years after than those who did not [14].
Previous studies also demonstrate the relationship between depression and failure to quit smoking. However, the impact of smokers' depression status change (e.g. from depressed status to non-depressed status) on their DCA has not been adequately assesses. Thus, this study sought to examine the effect of smokers' depression status changes on their DCA.

Data and population
The study used data from the Korea Welfare Panel Study (KoWePS), conducted by the Korea Institute for Health and Social Affairs and Seoul National University. The KoWePS is an annual longitudinal panel survey that was undertaken from 2006 to 2018. This data is suitable for low-income policies or poverty studies because about 50% of samples were low income earners with a median income of 60% or less. This study used a sample drawn from waves 3 (2008) to 13

Change in depression status
The main independent variable is change in depression status, which was measured by the 11-item version of the Center for Epidemiologic Studies Depression Scale (CESD-11). The CESD-11 is a shorter version of the original 20-item version, and is a self-reported screening tool which is well validated [15]. The total score of the CESD-11 was calculated by adding the scores for all 11 questions and multiplying these values by 20/11. Scores of 16 and higher were considered evidence of likely depressed status.

Covariates
Demographic, socioeconomic and health-related factors were included. Demographic variables included gender, age, and region. Socioeconomic variables included education level, marital status, and income level (quartile). Health-related factors included age at smoking commencement, alcohol consumption, and the absence or presence of chronic disease.

Statistical analysis
We calculated the distribution of the general characteristics at baseline. An analysis of variance (ANOVA) test was used to analyze the mean DSA at the rst time point, for change of depression status categories. To analyze the effect of change of depression status on DSA, we used the generalized estimating equation (GEE) model. The GEE model is a more e cient and unbiased regression estimate for analysing longitudinal or repeated measures in research designs with non-normal response variables [16]. Statistical analyses were performed using the GENMOD procedure in SAS Page 4/13 version 9.4 (SAS Institute, Inc., Cary, NC, USA). The results were considered statically signi cant if the p-value was less than 0.05. Table 1 presents the general characteristics at the rst change time point (2008→2009). Of the 1,821 participants, 86.9% of males and 64.2% of females were in the No→No group. The Yes→Yes group was the smallest, accounting for 2.7% of males and 10.2% of females. Among males, the No→Yes group had the highest mean DCA (18.6, SD: 9.88), and the Yes→No group had the smallest (16.40, SD: 6.29). Among females, the Yes→No group had the highest mean (13.95, SD: 8.48) DCA, and the No→No had the smallest (10.54, SD: 6.15).

Discussion
This study examined the effect of smokers' depression status change on their daily cigarette smoking amount (DCA).
Findings suggest that emerging from a depressed status is related to a decreased DCA among males. For males who smoke more than 15 cigarettes per day, depressed status appears to increase DCA. Among males who began smoking before age 19, and males living with spouses, Yes→No depression status groups had a lower DCA than the No→No groups.
Smokers experiencing signi cant depression are more likely to begin smoking, due to the antidepressant effect of nicotine. In addition, nicotine boosts bioavailability of serotonin, whose mechanism is similar to some antidepressants [16]. Thus, depression contributes to increased DCA. Moreover, depression encourages smokers, particularly those who smoke more than 15 cigarettes per a day, to persist in this behaviour despite the health risk. It is, therefore, necessary to help smokers seek alternative, healthier antidepressant resources, or to encourage smokers with depressive symptoms to consult mental problem professionals.
Much like smoking from an early age was associated with increased DCA, several studies reported the association between the age of smoking initiation and smoking cessation as well as strong nicotine dependence [17][18][19]. Indeed, a study focused on smokers who visited the Center for Cancer Prevention and Early Detection (in Korea's National Cancer Center) reported that nicotine dependence was much higher among those who started smoking aged 19 or earlier (compared to those who started aged 25 or later) [20]. Of signi cance in our study is the fact that smokers who began before 19 years of age and changed from depressed to non-depressed status had a much lower DCA than general smokers with the same change pattern. This may provide evidentiary support for the implementation of smoking cessation programs by detecting depression symptoms and encourage to participating in antidepressant programs in parallel with smoking cessation programs for smokers with early initiation age.
This study demonstrated no statistically signi cant results among females. This is likely related to the low selfreporting rate among female smokers in Korea. This small number of female smokers in this study data might be underestimated by self-reported responses. The notion of an underestimation issue in this demographic is supported by the Korea National Health and Nutrition Examination Survey (KNHANES) study. By comparing the number of females who self-reported as smokers with the number con rmed to be through urine cotinine examination, a signi cant discrepancy between the prevalence of self-reported and biochemically-veri ed female smokers [21]. To identify the effect depression status change on female smokers, further study is needed, using data obtained from other smoking detection methods, e.g. urine cotinine examination.
This study has several limitations. First, since the KoWePS largely represents low income households, study ndings are somewhat limited when it comes to generalizing to the national population, and high income households in particular. Second, because of data limitations, the analysis could not include clinical laboratory results (such as those from urinary cotinine level exams) that show the degree of smoking intensity. Third, the use of self-reported responses means that measurement issues such as reporting bias and recall bias may have confounded the ndings.
Despite the limitations, the present study has several strengths. First, we used data from a nationwide survey with randomly sampled longitudinal data and a 10-year follow up. As the KoWePS largely represents low income households, it can be generalized to South Korea's low-income population. Second, previous studies have examined the association between depression and smoking cessation success. However, this study used a lagged GEE model to determine the ability of depression status change (which may occur through treatment) to lessen daily smoking cigarettes among smokers.
In conclusion, this study examined the effect depression status change on smokers' DCA. We found that change from depressed to non-depressed and non-depressed to depressed status is associated with decreasing and increasing DCA, respectively. Among those who began smoking before age 19, the subgroup that changed from depressed to nondepressed status had a much lower DCA than general smokers with the same change pattern. These ndings may support the incorporation of depressive symptom treatment into smoking cessation programs.

Declarations
Ethics approval and consent to participate The KoWePS data are openly published; thus, ethical approval was not required for this study. This study did not require informed consent from the participants, as their information was fully anonymized and unidenti ed prior to analysis.

Consent for publication
Not applicable.
Availability of data and materials The data supporting this article is openly available in https://www.koweps.re.kr:442/main.do by submitting oath and data utilization plan.

Competing interests
The authors declare that they have no competing interests.

Funding
This research received no speci c grant from any funding agency in the public, commercial or not-for-pro t sectors.