Health behaviors in major chronic diseases patients: trends and regional variations analysis, 2008–2017, Korea

Background Improving the health behaviors of those with chronic diseases such as hypertension and diabetes is important for disease management. Few in-depth studies have been conducted in Korea on the health behaviors of chronic disease patients. This study examined the health behaviors of chronic disease patients over time and compared them with those of the general population. Methods Cross-sectional time-series data obtained from the Korea Community Health Survey from 2008 to 2017 were analyzed. Thirteen diseases were included in this analysis, namely, hypertension, diabetes, dyslipidemia, stroke, myocardial infarction, angina, osteoarthritis, osteoporosis, asthma, allergic rhinitis, atopic dermatitis, cataract, and depression. The current smoking rate, drinking rate, and the walking rate, which are leading health behaviors necessary for preventing chronic diseases, were analyzed by disease type. We compared patients’ health behaviors with those of the general population and identified regional variations. Results Although the current overall smoking rate was seemingly declining, the overall monthly drinking and high-risk drinking rates were increasing. In 2017, patients experiencing depression symptoms had a higher smoking rate than did the general population; hypertension and diabetes patients had a higher risk-drinking rate than did the latter. The general population’s walking rate was highest. There were considerable variations by region among chronic disease patients. Conclusions Chronic disease patients displayed worse health behaviors than those of the general population, in some instances. Rather than focusing only on chronic disease patients’ medication adherence, strategies must be devised to increase their smoking cessation rate, decrease their drinking rate, and increase their walking rate. Supplementary Information The online version contains supplementary material available at 10.1186/s12889-020-09940-7.


Background
Noncommunicable diseases (NCDs) are characterized as chronic, slow-progressing diseases, with a sharply increasing burden. NCDs contribute to 41 million deaths annually worldwide (71% of total deaths) [1]. Deaths from NCDs are expected to reach 52 million by 2030 [2]. Cardiovascular disease, cancer, chronic respiratory diseases, and diabetes are NCDs with sequentially the highest mortality rates [2]. Korea has a considerably high NCD burden; in 2015, the number of disability-adjusted life years (DALYs) due to NCDs was 25,683 for 100,000 people (87.1% of the total DALYs) [3].
To reduce the NCD burden, most public health policies aim to identify and manage modifiable health behaviors that can effectively reduce the number of deaths caused by NCDs [4]. The World Health Organization (WHO) compiled and introduced 16 cost-effective, easily implementable interventions as "best buys," believing that these can reduce the 9.6 million global premature deaths caused by NCDs from 2018 to 2025. The following are modifiable behavioral risk factors targeted by interventions cited as "best buys": reducing tobacco use, harmful alcohol use, an unhealthy diet, and physical inactivity [1].
This study used large-scale, nationally representative data, to examine and compare the health behaviors of patients diagnosed with various chronic diseases with those of the general population. This study also analyzes annual trends to identify health behavior changes in the general population and chronic disease patients.

Source of data
Using cross-sectional time-series data of the Korea Community Health Survey from 2008 to 2017, this study compared the health behaviors of patients with major chronic diseases across 18 cities and provinces in Korea (for Sejong since 2012). The Korea Community Health Survey, conducted by the Korea Centers for Disease Control and Prevention since 2008, collects data from 253 community health centers in Korea on more than 220,000 adults aged ≥19 years (900 × 253 community health centers) to plan public health policies [29]. The survey population is all adults aged ≥19 years living in residential houses at each sample point and survey households were selected using systematic sampling by stratification into Tong/Ban/Lee (smallest administrative district units) and housing type. All adults ≥19 years living in the selected survey households were interviewed. The survey collects information about health behaviors such as smoking and drinking, medical history, healthcare use, health-related quality of life, and demographic characteristics.

Definition of major chronic diseases
The Korea Community Health Survey checks for major chronic disease among participants. This analysis focused on hypertension, diabetes, dyslipidemia, stroke, myocardial infarction, angina, osteoarthritis, osteoporosis, asthma, allergic rhinitis, atopic dermatitis, cataract, and depression. A physician's diagnosis confirmed a patient's disease status. For instance, those answering "Yes" to the question, "Have you ever been diagnosed with diabetes by a physician?," were defined as diabetes patients. Depression included not only whether participants had been physician-diagnosed, but also whether they had experienced depression symptoms. Those answering "Yes" to the question, "Have you ever felt so sad or devastated in the past 1 year that it affected your daily life for 2 weeks or longer consecutively?," were defined as experiencing depression symptoms. The survey did not investigate annual occurrence of all the above conditions; Table 1 presents the survey's annual tracking of diseases per year.

Health behaviors
This study analyzed the current smoking (total vs. male), drinking (monthly and high-risk drinking rates), and walking rates as leading health behaviors. The current smoking rate was defined by dividing the number of current smokers (every day or sometimes) among those who had smoked ≥5 packs (100 cigarettes) in their lifetime, by the total number of respondents. The monthly drinking rate was defined by dividing the number of those who had drunk once per month in the past year by the total number of respondents. The high-risk drinking rate was defined by dividing the number of those who drank 7 (men) or 5 glasses (women) in one occasion ≥2 times a week in the past year, by the number of those who drank in the past year. The walking rate was defined by dividing the number of those who walked 30 min daily for ≥5 days per week in the past week, by the total number of respondents.

Analysis
With descriptive analysis, the health behaviors of patients per disease were organized by year and region and compared with those of the general population. The general population was operationalized as all the participants in the Korea Community Health Survey. In addition, we obtained each standard deviation (SD) for health behaviors in 18 cities and provinces, to identify regional variations.
To examine any significant changes in health behaviors, joinpoint regression analysis was performed [30]. We permitted no more than 1 joinpoint, calculated the annual percentage change (APC) and average APC (AAPC) (when the number of joinpoints was 1), and presented the p-value for each. Stata/SE13.1 (StataCorp, Texas, TX) was used for descriptive analysis. For joinpoint regression, we used the Joinpoint Regression Program version 4.6.0 (US National Cancer Institute, Bethesda, MD, USA).

Results
Overall, chronic disease patients and the general population did not show major health behavior differences. In 2017, while the general population's current smoking rate was 20.2%, those of hypertension and diabetes patients were 16.6 and 19.2%, respectively (Fig. 1). The current smoking rate of patients experiencing depression symptoms was 21.9%-higher than that of the general population. The general population's monthly drinking rate was highest (59.3%), with that of patients experiencing depression symptoms at > 50%. Overall, high-risk drinking rates have been on the rise since 2010, in particular, the high-risk drinking rates of hypertension and diabetes patients were 22.3 and 20.6%, respectively-higher than that of the general population (18.7%). The walking rate was highest in the general population (45.1%) and lowest (42.0%) in patients experiencing depression symptoms.

Current smoking rate trends
The current overall smoking rate showed a declining trend (Fig. 2, Supplement 1 The current male smoking trend was similar to the overall trend (Fig. 3, Supplement 1). In the general population,

Drinking rate trends
The overall monthly drinking rate showed an increasing trend (Fig. 4 It also increased for patients with cerebrovascular and cardiovascular diseases such as stroke, myocardial infarction, and angina. The overall high-risk drinking rate showed an increasing trend, except from 2008 (Fig. 5, Supplement 1). The general population's high-risk drinking rate was 17.6% in Fig. 2 Trends in the current smoking rate (total) by disease

Trends in the walking rate
The overall walking rate showed a declining trend (Fig. 6  Jeju (40.6%) had the highest current male smoking rate for the general population, and Sejong (34.8%), the lowest rate (SD: 1.6%). In the same year, Busan (36.0%) had the highest current smoking rate for diabetes patients, and Gwangju (25.5%), the lowest rate (SD: 2.8%).
Unlike other health behaviors, the walking rate showed larger regional variations in the general population. As of 2017, Seoul (61.1%) had the highest walking rate for the general population, and Gangwon (32.4%), the lowest rate (SD: 7.5%). In the same year, Seoul (57.6%) had the highest walking rate in hypertension patients, and Jeju (29.5%), the lowest rate (SD: 7.3%).

Discussion
This study used the Korea Community Health Survey to examine the health behaviors of patients with major chronic diseases with regard to smoking, drinking, and physical activity, and compared these with those of the general population. We analyzed these trends from 2008 to 2017. Patients with major chronic diseases did not show notably better health behaviors; some showed  worse health behaviors than did the general population. Although the smoking rate decreased, the decrease in patients with major chronic diseases was smaller than that in the general population. Regional variations were larger in the health behaviors of patients with major chronic diseases, compared to those of the general population. Improved health behaviors can prevent NCDs and reduce their recurrence risk and severity, improve health-related quality of life, and extend life expectancy [31,32]. Hence, the importance of health behaviors should be conveyed to the general population, to improve NCD patients' health behaviors. Given the general difficulty of changing health behaviors [12], effective strategies should be developed, and resources assigned accordingly. Rather than focusing only on chronic disease patients' medication adherence, strategies must be devised to encourage smoking cessation, reduce alcohol consumption, and increase patients' walking rate. This study holds significance, as it presented empirical evidence to enable goal setting, so as to improve chronic disease patients' health behaviors, and provided reasons for such improvement.
Smoking cessation has been highlighted for its potential to prevent or stabilize NCDs [33]. In particular, an NCD diagnosis is expected to serve as an important opportunity for smokers to change their smoking behaviors; in fact, the smoking rate has shown decline after a stroke [34], cancer [22], and diabetes [23] diagnosis. However, findings that > 3 out of 4 patients still smoked 2 years after a stroke diagnosis, and that only 20% of patients with lung diseases quit smoking within 2 years [14], suggest that a diagnosis does not necessarily translate into smoking cessation. In this study, the current smoking rate of hypertension (16.6% as of 2017) and diabetes patients (19.2% as of 2017) did not differ substantially from that of the general population (20.2% as of 2017), suggesting a need for more active treatment for smoking cessation for chronic disease patients.
We assumed that smoking was particularly concerning in patients with mental disorders. Smoking increased only among depression patients (17.2% in 2009 to 18.9% in 2013) in this study, and the current smoking rate of patients experiencing depression symptoms (21.9% as of 2017) was higher than that of the general population. These results were consistent with those of previous studies, showing a higher smoking rate among patients with mental disorders, that these patients accounted for a higher proportion of smokers, and that they were less likely to quit smoking successfully [35,36]. A study reported that smoking cessation led to depression [37], suggesting that it might be better to let depressed patients smoke, thereby discouraging the smoking cessation recommendation to these patients. However, given mounting evidence on how to treat smoking patients with mental disorders [38], and the latest empirical evidence suggesting that smoking cessation actually decreases depression [39,40], it is worth focusing more on smoking cessation among patients with conditions such as depression.
This study also found a higher drinking rate in chronic disease patients. While chronic disease patients' monthly drinking rate was lower than that of the general population, those with cerebrovascular and cardiovascular diseases displayed a higher rate. Of more concern was that the high-risk drinking rate of patients with depression symptoms, hypertension, and diabetes was higher than that of the general population. Drinking is generally associated various NCDs including cancer, hypertension, hemorrhagic stroke, and liver diseases [41]. While small amounts of alcohol reportedly show preventative effects on diabetes [42], they also increase the risk of chronic diseases such as atrial fibrillation [43] and cataracts [44]. We cannot recommend small alcohol amounts to prevent some diseases, at the risk of others developing, nor can we recommend such to current chronic disease patients. Moreover, given drinking's high correlation with smoking [45], we cannot rule out the possible effect of drinking on smoking behaviors.
More focus should be directed towards drinking in Korea, considering increases in the monthly and highrisk drinking rates in both the general population and patients with major chronic diseases. A permissive drinking culture has developed in Korea, with drinking considered to represent non-verbal communication, demonstrate a sense of community, and serve as a bridge between work and leisure. This permissive culture may have contributed to 6 out of 10 Korean adults drinking once or more per month. The higher the average drinking rate, the likelier problematic drinking is. Thus, drinking reduction strategies for the general population and chronic disease patients are necessary. To reduce drinking among the latter, physicians should assess their drinking patterns at each consultation, inform them of the harm associated with drinking, and actively recommend drinking cessation.
Physical activity is important for chronic disease management. The benefits of physical activity for hypertension and diabetes patients are well known; regular physical activity is recommended in the treatment guidelines for these diseases [46,47]. Furthermore, exercise's confirmed depression alleviation effect [48] has prompted its recommendation to depression patients. Moreover, despite exercise seeming somewhat risky for diseases like myocardial infarction, exercise-centered cardiac rehabilitation improves quality of life and reduces the mortality rate for myocardial infarction [49]. Therefore, exercise must be recommended to chronic disease patients, and the level of physical activity (e.g., the walking rate) must be evaluated as an indicator. The current study showed that the walking rate of patients with major chronic diseases was lower than that of the general population; there were annual declining trends in the walking rates of the general population and patients with major chronic diseases. Although the obesity rate in Korea is lower than that in other countries, a declining walking rate may be related to a consistently increasing obesity rate. Programs for increasing the walking rate in chronic disease patients and the general population are necessary.
Notably, although variations are affected by the sample sizes, regional variations in health behaviors were larger in chronic disease patients than in the general population. For instance, the standard deviation for the general population's current smoking rate was 1.1% as of 2017, and 2.9% for patients experiencing depression symptoms. While for the former, the standard deviation for the monthly drinking rate was 3.2%, it was 4.5% for hypertension patients. Although the life expectancy gap between different regions in Korea has narrowed, a quality-adjusted life expectancy gap of 4.6 years has been shown between regions [50]. While many factors such as unequal health resources and income gaps ostensibly contribute towards this gap, a regional gap in health behaviors may also explain this phenomenon. More studies are needed to explain the regional health behavior gap.
This study had limitations. First, changes in health behaviors, based on a chronic disease diagnosis, were not examined due to the nature of this study's data source. Future studies could use panel or patient cohort data to examine changes in chronic disease patients' health behaviors over time before and after diagnosis. Second, chronic diseases were self-reported. The accuracy of these self-reports of physician diagnosis and health behaviors used in the Korea Community Health Survey, cannot be determined. A follow-up study would need to confirm the diagnoses and the health behaviors in relation to physical assessments or healthcare use data. Third, although this study focuses on patients with 13 chronic diseases, cancer-one of the leading NCDswas excluded. Studies could analyze annual trends and regional variations in cancer patients' health behaviors, using this study's methodology. In addition, the gender differences in the smoking rate or health behavior of chronic disease patients could be a good research topic which we will consider in the future.

Conclusions
This study using large-scale, nationally representative data confirmed that patients with major chronic diseases did not show particularly better health behaviors than did the general population, with some even displaying worse health behaviors. This study will enable goal setting aimed at improving chronic disease patients' health behaviors, in particular in regional level. Goal setting and management, and strategizing in this regard, are important not only regarding chronic disease patients' medication adherence, but also health behaviors such as smoking cessation, alcohol abstinence, and physical activity. Physicians who diagnose chronic disease patients could also implement strategies that identify and improve patients' health behaviors. Furthermore, considering the difficulty of improving the general population's health behaviors, an effective public health approach may be necessary in this regard, instead of reliance on individuals to improve their health behaviors.