Introduction

Tobacco smoking is known risk factor for early morbidity and mortality worldwide [1–3]. Smoking is the second largest contributor to global disability-adjusted life-years (DALYs), contributing to 148 million DALYs annually [2]. The World Health Organization (WHO) has estimated that smoking is responsible for 12% and 6% of worldwide deaths among males and females, respectively [4]. It is predicted that one billion people will die in the 21st century as a result of smoking if there is no change in smoking habits, with most of these deaths striking low- and middle-income countries [5].

Prior studies have investigated the association of tobacco and increased mortality rates [6–8], and cohort studies in the UK and US have found that the relative risk of mortality and the effects on the populations vary between countries [9, 10]. Different environmental aspects such as socioeconomic status, stress and genetic factors are among the possible explanations of that variation [11]. The evidence on how smoking increases all-cause mortality rates in low-to middle-income countries is limited. For example, Gu et al. found that 673,000 deaths were attributable to smoking in China in 2005 [12]. In other report, Correa et al. showed that cigarette smoking was responsible for 24,222 (13.64%) deaths of persons aged 35 years and older in 2003 in 16 Brazilian regional capitals [13]. In the Casablanca region of Morocco 9.7% of deaths among people older than 35 were caused by smoking in 2012 [14]. However, less is known regarding the link between smoking prevalence and mortality in Indonesia.

Smoking is responsible for a high proportion of morbidity and mortality in Indonesia. A 2015 study showed that 925,611 males (93.27%) and 66,719 females (6.93%) in Indonesia were hospitalized as a result of diseases attributable to smoking such as hypertension (42.6%), chronic obstructive pulmonary disease (COPD) (40.2%), and stroke other diseases (12%), accounting for 21,05% of all chronic diseases in the country [15]. A 2017 study showed that cerebrovascular attack and ischemic heart disease are leading causes of death in Indonesia. During the period from 2007 to 2017, the mortality rate of diseases attributable to smoking such as a cerebrovascular attack, ischemic heart disease, and COPD increased by 29.2%, 29.0%, and 10.5%, respectively [16].

The incidence of smoking-attributable diseases in Indonesia is predicted to remain high. As of 2016, 39.5% of Indonesian aged 15 years and older are active smokers; this proportion is 7.4% higher than the global average [17]. Furthermore, the percentage of Indonesian men who smoke tobacco has increased dramatically, rising from 56.2% in 2000 to 76.2% in 2015 [17]. As one of the countries that has not signed the World Health Organization Framework Convention on Tobacco Control (WHO FCTC), Indonesia also lacks of policies related to the tobacco market [18, 19]. Taken together, these circumstances indicate that Indonesia has a significant effect on the incidence of tobacco-related illness worldwide.

There has as yet been no research on the hazard risk of smoking for all-cause death in Indonesia. As a developing country with the highest smoking consumption in the world, Indonesia’s patterns are a different from those of other countries, even within Asia [20]. It is essential to determine the relationship between smoking behaviour and the developing risk caused by tobacco smoking. With this study, we aimed to investigate whether smoking status predicts all-cause mortality risk in a developing country.

Materials and methods

Results

Table 1 presents the statistics on demographic and socio-economic characteristics along with health behaviours by mortality status. The sample consisted of 53.8% females. The respondents’ average age was 56.9 years (standard deviation = 10.4). Only approximately 6% of respondents had completed college. Most of the respondents (78.7%) were married and lived on the islands of Sumatra or Java (90.8%). The respondents’ average waist circumference was 80 cm for males and 82 cm for females. The proportions of respondents with cardiovascular diseases, diabetes and history of stroke were 1.9%, 2.4%, and 0.8%, respectively. More than half of the respondents had low HDL cholesterol or hypertension, and almost 40% had high cholesterol levels.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of study participants.

https://doi.org/10.1371/journal.pone.0242558.t001

Approximately 40.3% of the respondents were current smokers. Approximately 66% respondents started smoking when they were 17 years or older. The summary of descriptive statistics of the sample by mortality, the bivariate analysis and the mortality rates appear on the right pane of Table 1. The data show that the mortality rates were higher among older people, current smokers, those with less educational attainment, and those with no comorbidities. S1 Table provides the baseline characteristic by smoking status.

During the eight years of follow-up, there were 509 mortalities. Current smokers had a higher risk of all-cause mortality. Of the non-current smokers (n = 2,059), 250 died (12.1%); of the current smokers (n = 1,294), 259 died (20.0%). Table 2 shows the estimated hazard ratio and 95% confidence intervals for mortality in two models. We included demographic, marital status, socio-economic status, area of living, and the presence of comorbidities in the Model 1. We than added smoking status in Model 2. The results in Model 1 shows that being younger and college-educated are related to lower hazard risks of all-cause mortality. Respondents who were married or separated/widowed have lower odds of all-cause mortality than those who were single. Living on Kalimantan Islands, and having hypertension or cardiovascular disease were associated with a higher risk of mortality among all participants. All those associations remained significant when we included smoking behaviour in Model 2. It shows that when demographics and other potential covariates were equal, current smokers had a 48% higher risk of death (95% confidence interval = 1.11 to 1.98) than non-current smokers. We further performed those analyses separately by smoking behaviour in S2 Table; S3 Table, and S1 Fig.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Hazard ratios of all-cause mortality stratified by smoking status.

https://doi.org/10.1371/journal.pone.0242558.t002

Discussion

Indonesia has the fourth largest population in the world. 2015 study found that more than 25% of the entire Indonesian population were current smokers [1]. In this study, we found that 40.3% of our respondents were current smokers, 89% of them are men. This finding is higher than the global average: 25% for men and 5.4% for women in the world [1]. After China and India, Indonesia has the largest number of smokers in the world. The number of smokers in these three countries accounts for more than 50% of smokers worldwide. The number of smokers in Indonesia did not decrease over the 15 years from 1990 to 2015. One contributing factor may be of the Indonesian government’s lack of tobacco control policies, including policies on smoking in open spaces, marketing cigarette products, and selling cigarette products. Indonesia is not signatory of or party to the WHO FCTC. The FCTC comprehensively prohibits cigarette promotion, advertising and sponsorship, while the Indonesian government has not banned cigarette advertising. Cigarette companies have a large market in Indonesia, and they can aggressively promote their products. In addition, the price of cigarettes in Indonesia is relatively low compared with other countries; the cigarette tax in Indonesia is far below the WHO recommendation of 70% of retail price. Indonesia's average 2017 excise tax rate was 49.1% of retail price. The Indonesian government has been unable to meet the WHO recommendation as its excise law sets the maximum tobacco excise tax at 57% of retail price. This renders Indonesia one of the world’s largest markets for global cigarette marketing. Smoking-related health problems are increasing gradually along with the number of smokers.

The hazard risk of current smokers in this research is shown to be higher than that of non-smokers. The impact of smoking on increased all-cause mortality rates has been widely acknowledged in previous research [7, 8, 23–25]. Similar to this finding, current smokers have a higher risk of all-cause mortality than non-smokers. A meta-analysis on smokers in other Asian countries shows that Asian male smokers in China, Japan, Korea, Singapore, Taiwan and India have a higher risk of all-cause mortality. That risk factor is increased if the person starts smoking at a young age [26]. The death rate of current smokers in Australia is three times higher than among non-smokers. This finding matches research results in Western countries such as the US and UK. The relative risk (RR) caused by smoking consistently is the 2.8 to 3 over 50 years [27, 28].

Age show significant association with the risk of all-cause mortality. Older people have a higher risk of all-cause mortality than younger people. It is well known that ageing affects mortality because of the accumulation of damage and deterioration at the cell, tissue, organ, and organism levels, ultimately leading to death [29].

Our result indicates that education inversely associated with all-cause mortality. Various previous study showed similar result [30, 31]. The people with high education tend to have a better job and higher income, which make people with high education to set aside money to address health concerns [32]. Moreover, education also influences health awareness and health believes [32].

Marital status also influences the hazard risk of all-cause mortality. Physiological distress plays an important role in explaining this finding. Single people are less likely to have a high survival rate than separated, widowed and married people. Married people tend to have more social advantages. Furthermore, married people are more likely to have support from their partner and family, which leads to healthier life [33].

Our result indicated that people who live in urban area more likely to have higher hazard all-cause mortality rate than people who live in rural area. Urban area has a higher population density than in rural area [34]. High population density related to urban stress such as vandalism, noise from neighborhood, air pollution and low quality [34]. The previous report in Denmark and Japan showed that population density influence mortality rate [35, 36].

People living in Kalimantan Island have a higher risk of all-cause mortality rate compared to residents of other islands. Premature mortality such as stroke, ischemic heart disease, diabetes and COPD, was more frequent in all provinces in Kalimantan than in Indonesia overall [16]. One plausible explanation for this is that Kalimantan has a number of industrial areas which may affect air quality [37, 38].

The presence of comorbidities such as CVD, diabetes, stroke, and hypertension, is more likely to increase all-cause mortality risk. According to the data from the Institute for Health Metrics and Evaluation, CVD, diabetes, stroke and hypertension are included in the top ten diseases causing the death in Indonesia [39]. Individuals with non-communicable diseases are usually asymptomatic and progressive as they are chronic conditions. Poor screening system for non-communicable diseases in Indonesia causes the person to not recognize that they have this disease. Those diseases are often diagnosed if the person already has a symptom when the disease already disturbs the physiology of body and causing a problem in the organs [40]. Non communicable diseases are strongly related with poor health behaviors, such as smoking consumption, diet and drinking alcohol. Changing the behaviors of a patient with non-communicable diseases requires a large effort. This circumstance causes the patient with non-communicable disease to struggle to control their disease, leading to a higher mortality risk [41].

Among smokers, people with lower educational levels, and those who are unmarried are more likely to have a higher risk of mortality; among non-smokers, we observed no such association. Education has an inverse effect on smoking. People with lower educational levels are more likely to be smokers than people with higher educational levels; this result is supported by much previous research [33, 42, 43]. The effect on smoking may be explained by awareness of health facts, awareness of preventive treatment and concern about health, and seeking treatment, all of which are more prevalent in people with more formal education; this group also has greater ease accessing health facilities as those with more education are generally also wealthier. In contrast, people with lower levels of education are less likely to seek assistance at a smoking cessation facility. In families with lower educational levels, several smokers usually live together [44]. Furthermore, people who graduate from college are 17% more likely to try to stop smoking than those with less education [42].

Marital status also influences the hazard risk of all-cause mortality in the smoking group. The spouse is twice as important as a friend or even a sibling in an individual’s decision to quit smoking [45]. Among Hispanic and non-Hispanic-white singles in the US, previous research has shown that those who were single/never married had the highest prevalence of smoking. The same study found that single people experience more psychological distress. Unmarried people suffer more from loneliness and dissatisfaction than married people, and they may try smoking to relieve their stress [46].

Our result shows that individuals with comorbidities, such as CVD, stroke, and diabetes, have higher all-cause mortality risks than those without the comorbidities in the non-smoking group, while in the smoking group there is no significant difference in the mortality risk. The lack of association between chronic diseases and mortality risk among current smokers may due to several plausible explanations. Firstly, the respondents may follow the advice from the doctors and stop smoking when they were diagnosed by chronic diseases. A study in Indonesia showed that current smokers had 26% lower odds of having hypertension, while those who quit smoking had higher odds of having hypertension [47]. The second plausible explanation is the presence of other diseases causing mortality among current smokers, such as lung cancer and chronic obstructive pulmonary disease [48]. Further research to establish this finding is required.

This study has a number of limitations. Firstly, it investigates all-cause mortality risk and its association with smoking behaviour as the data source (IFLS Wave 5) provides no acceptable information on specific causes of mortality, such as lung cancer or cardiovascular events. Another limitation is that the IFLS sample is representative of about 83% of the Indonesian population mostly in the west and central parts of the country. IFLS is not representative of all Indonesian provinces. The survey mostly in the west and central parts of the country, in the wave 5 still and excluded most eastern Indonesian provinces, which are considered underdeveloped compared to their western counterparts. Additional data sets on population health that encompass all Indonesian provinces are available (e.g. Indonesian Socioeconomic Survey [SUSENAS] and Riset Kesehatan Dasar [RISKESDAS]), but these data are not longitudinal and do not provide information on the deaths of their respondents. Despite its limitations, this study has several strengths, including its prospective cohort, represent more than 80% of Indonesian population and the high rate of follow-up.

Conclusions

In conclusion, our results show that the number of current smokers in Indonesia remains high and that smoking increases the risk of all-cause mortality. This situation is expected to directly impact many health problems in the future. It is therefore essential to review the Indonesia government’s tobacco control policies with regard to many aspects such as tobacco regulation, higher taxes on cigarettes, and regulation of smoking areas. The government must also consider potential interventions in order to encourage smoking cessation among current smokers.

Supporting information