Article Text
PDF
Abstract
Background The slow progress of tobacco control policy in Japan reflects the tension among the interests of the Finance Ministry (which remains the majority shareholder of Japan Tobacco, Inc), the Health Ministry, tobacco growers and consumers.
Methods We sought to examine regional disparities in compliance with national tobacco control legislation (the 2003 Health Promotion Law). Specifically, we sought to examine whether prefecture-level compliance with legislation was correlated with decreases in smoking prevalence, and decreases in lung cancer mortality rates. We also examined whether prefectural involvement in growing tobacco was associated with lower compliance with the law.
Results From 2001 to 2007, higher prefectural compliance with tobacco control laws was associated with decreased prevalence of smoking. Decreased tobacco consumption was in turn associated with declining lung cancer mortality. Prefectures involved in growing tobacco exhibited lower levels of compliance with national tobacco control laws. The same prefectures also exhibited the worst improvement in smoking prevalence.
Conclusions This study in Japan suggests that tobacco control policies are being unevenly implemented across prefectures, and that measures to counteract the influence of local tobacco culture are required to reduce the disparities in regional tobacco control outcomes in that country.
- Health policy
- tobacco
- smoking
- public health
- lung neoplasm
- litigation
- public policy
- smoking caused disease
Statistics from Altmetric.com
Introduction
Tobacco control policy in Japan has been characterised as a ‘tug-of-war’ between the powerful Ministry of Finance versus the Ministry of Health, Labour and Welfare.1 Historically, the Japanese government has maintained a direct interest in the tobacco industry, establishing a monopoly ownership of the tobacco business in 1904. Prior to the privatisation of the tobacco monopoly in 1985, it was the Ministry of Finance that directly controlled activities such as the purchase of domestically grown tobacco leaf, and the manufacture/sale of cigarettes. In April 1985, the Japan Tobacco, Inc was established to create an arm's-length distance between the Ministry of Finance and the tobacco business, although the Japanese government still held 100% of the ownership of the new company. Following the first public stock offering in October 1994, government ownership of Japan Tobacco, Inc has declined, although it still remains the majority stockholder (50.02%). In short, the Ministry of Finance's fiscal interest in the tobacco industry is directly at odds with the public health interest of the Health Ministry. The result—given the clout of the Finance Ministry—is that Japan has been much slower than most Western countries in adopting national tobacco control legislation. As Mark Levin has noted2:
“The Ministry of Finance's successful development of institutional dominance through its control over the annual fiscal budget process seems … important. When the budgetary axe swings over lesser agency heads, all players understand the dangers of seeking funding for measures likely to be unpopular with Ministry of Finance budget directors. This enables a wordless annihilation of much meaningful tobacco control development at the national governmental level, reflected in minuscule tobacco control budgets for decades.” (Levin, 2005, p 48).
More recently, however, Japanese tobacco control efforts have received a boost by the Healthy Japan 21 initiative, introduced in 2000, which included tobacco control measures including education about the harmful effects of tobacco, youth tobacco control, passive smoking control and dissemination of tobacco cessation programmes.3 This was followed in 2003 by the introduction of the Health Promotion Law, which remains one of the most influential tobacco policies introduced in recent years in Japan. The law (Article 25) states as follows2 (the hortatory language here ‘should try’ is not limited to tobacco control law in Japan and is not uncommon in many Japanese laws):
“Prevention from Passive Smoking Exposure: Persons who manage schools, gymnasiums, hospitals, theaters, public assembly halls, gallery spaces, department stores, offices, government buildings, restaurants and bars, and other facilities used by numerous people should try to take whatever steps are necessary to prevent passive smoking exposure (ie, being forced to breathe other person's tobacco smoke in indoor or functionally equivalent spaces) for the uses of such facilities.”
It mandates people who manage public places take measures to restrict exposure to secondhand smoke; however, there are no penalties for violations. Similar to other laws in Japan, the Health Promotion Law was a national enactment that was uniformly applicable throughout the country, but there was no pre-emption of more restrictive policies at the local level. Although only one prefecture, Kanagawa, introduced an ordinance with penalties to prevent passive smoking in March 2009, other prefectures in Japan are evenly encouraged to comply with the Health Promotion Law.
In the present study, we provide a prefecture-level analysis of variation in compliance with the Health Promotion Law of 2003, focusing on the extent to which the 47 Japanese prefectures have measured up to the legislative directive to restrict smoking in public facilities, ban smoking on public school premises, ban smoking in local government administration buildings and ban smoking in private settings. Contemporary Japan is usually portrayed as a nation characterised by strong centralisation, with legislative directives radiating from the centre.4 In contrast to this monolithic characterisation, our analysis seeks to examine the regional differences and disparities in tobacco control outcomes within Japan. In particular, we argue that the degree of implementation of central directives is contingent upon historical forces, specifically the extent of involvement of prefectures with tobacco growing (a powerful lobby in Japan). For example, when the Healthy Japan 21 report called for the Ministry of Finance to divest their stock ownership in Japan Tobacco, Inc, one of the strongest voices of opposition to this move came from the tobacco growing sector.2 This is analogous to the situation in the USA where the influence of the tobacco farming lobby has been repeatedly noted, leading to much more tobacco-friendly policies in states such as North Carolina and Kentucky.5 One of the aims of our paper is thus to draw attention to a similar geographical variation in Japan—a point that has been missed in previous research.
A second aim of our study is to examine the potential health burden of uneven implementation of national legislation across prefectures. Given previous research, which suggested that lung cancer mortality rates can begin to show changes within a few years of the introduction of tobacco control policies,6 7 we examined prefectural variations in lung cancer mortality rates as a marker of the health impacts of the Healthy Japan 21 initiative as well as the related Health Promotion Law of 2003.
Methods
Indicators of regional compliance with tobacco control policy
We utilised four indicators of tobacco control policy compliance (described below) related to the Health Promotion Law (2003), which focuses on the restriction of exposure to secondhand smoke. Data were calculated for each of the 47 prefectures in Japan.
Smoking bans in public facilities
We calculated the prefecture-level proportions of public facilities that completely prohibited smoking throughout their buildings and grounds as of 2008, including community centres, libraries, museums, education facilities for youth, education facilities for women, gymnastic halls and cultural venues. The data were obtained from Social Education Survey of Ministry of Education, Culture, Sports, Science and Technology in Japan.8
Smoking bans in public schools
We obtained prefecture-level proportions of public schools that completely prohibited smoking throughout the entire campus (as of 2009). The data were obtained from Harada et al.9
Smoking bans in prefectural government administration buildings
As the third indicator of compliance with legislation, we used the proportions of prefectural government administration building with a complete smoking ban as of April 2009. The data were obtained from the website of NPO ‘Coalition on a Smoke-free Environment for Kids in Japan’.10 We assigned one to prefectures that completely prohibited smoking throughout the entire campus of their facilities, 0.5 to prefectures that prohibited (or planed to prohibit) smoking only in the buildings (indoor spaces) and zero to prefectures that did not prohibit smoking.
Smoking ban in taxis
Finally, as a proxy of private sector adaptation of smoking restrictions, we obtained information on the timing of the introduction of smoking bans in taxis in 43 prefectures up to October 2010.11 Each prefecture was then assigned a weight, such that the first prefecture to introduce such a ban was assigned 1, while the most recent prefecture to achieve smoke-free taxis was assigned 43, and the remaining four prefectures which have yet to introduce bans was assigned 44.
For each of the four indicators, we computed z-scores for each of the 47 prefectures.5 We then summed the z-scores and created a total compliance index (n=47).
Trends in smoking prevalence at the prefecture level
A percentage decrease in tobacco consumption at the prefecture level and per capita decrease in tobacco consumption from 2001 to 2007 (n=47)
In order to estimate the prefecture-level trends in tobacco consumption, we obtained data on the prefectural tobacco tax in 2001 and 2007 from the report ‘Statistics of Local Public Finance’ compiled by the Ministry of Internal Affairs and Communications in Japan.12 We divided the total amounts of tobacco tax collected by the prefecture by the rate of prefectural tobacco tax in 2001 (0.868¥ per a cigarette) and 2007 (0.969¥ per a cigarette), in order to obtain an estimate of the consumption levels in those years and the percentage decrease. The rates changed simultaneously across the prefectures. Finally, we divided the total consumption by the population over 15 years old in each prefecture in order to obtain the per capita decrease in tobacco consumption from 2001 to 2007 to be compatible with the previous study.13
Trends in the prevalence of current smokers and former smokers at the prefecture level, 2001 to 2007 (n=47)
The data were obtained from the National Lifestyle Survey conducted by Ministry of Health, Labour, and Welfare in 2001 and 2007.14 We defined a current smoker as a person who reported they smoked daily or occasionally, and a former smoker as a person who reported they used to smoke but did not smoke within the last month.
Prefecture level tobacco leaf production
The data were obtained from All Japan Tobacco Cultivation Associations.15 We used tobacco leaf production estimates from 2003, when the Health Promotion Law was introduced (n=47).
Trends in lung cancer mortality in younger adults from 2001 to 2007
Because trends in lung cancer mortality rates in younger adults are considered to be an effective monitoring method of tobacco control programmes,6 we used lung cancer mortality rates in younger adults aged from 20 to 59 years old. We calculated standardised lung cancer mortality rates for each prefecture by the direct age-standardisation method using the 2005 Japan population census as standard. The mortality data were obtained from Vital Statistics.16
We then employed two indicators to evaluate the trend in lung cancer mortality rates: decrease in lung cancer mortality rate from 2001 to 2007 (n=47), and percentage decrease in mortality rate from 2001 to 2007 (n=47).
Statistical analysis
We first calculated the correlation between per capita tobacco consumption in 2001 and lung cancer mortality in 2001. We then calculated the correlations between tobacco control policy compliance and trends in smoking prevalence (as well as trends in lung cancer mortality). We also correlated prefecture-level involvement in tobacco growing with indicators of tobacco control policy compliance, trends in smoking prevalence and trends in lung cancer mortality. The unit of correlation analysis was prefecture level (n=47). All of the correlations were estimated in separate models by using Pearson's method.
To evaluate the potential for non-linear relations between the prefecture-level involvement in tobacco leaf production and each indicator (ie, tobacco control policy compliance, trends in smoking prevalence and trends in lung cancer mortality), we first drew scatter plots and then modelled each indicator as natural splines of tobacco leaf production (4 degrees of freedom).
We classified the prefectures into three groups based on their level of involvement in tobacco production: 0 to less than 2000, 2000 to less than 6000 and 6000 to less than 10 000 million yen. We compared each indicator (tobacco control policy compliance, trends in smoking prevalence, and trends in lung cancer mortality) across the three groups using analysis of variance. We also conducted statistical tests for trend, treating the three groups as an ordinal variable (1–3).
All of the analyses were conducted by PASW version 18.0J (SPSS Japan Inc) and p values less than 0.05 (two-sided) were considered as significant. Non-linear relations were examined by the package ‘splines’ in the statistical software R version 2.10.1.
Results
As expected, per capita tobacco consumption in 2001 was significantly associated with lung cancer mortality in 2001 (r=0.30 and p=0.04; figure 1A).
(A) Per capita tobacco consumption and lung cancer mortality in 2001 (r=0.30, p=0.04). (B) Total compliance index and increased prevalence of former smokers (r=0.32, p=0.03). (C) Percentage decrease in tobacco consumption at the prefecture and decrease in lung cancer mortality rate (r=0.31, p=0.03). (D) Tobacco leaf production and total compliance index (r=−0.39, p=0.01). The lowest tobacco production group (0 to less than 2000 million yen) is shown in black, the middle production group (2000 to less than 6000 million yen) is shown in blue, and the highest production group (6000 to less than 10 000 million yen) is shown in red.
The extent of compliance with tobacco control legislation varied by prefecture, and moreover, higher compliance with the national laws was associated with a greater decrease in the prevalence of current smokers from 2001 to 2007 (r=0.39 and p=0.01), as well as an increase in the prevalence of former smokers (r=0.32 and p=0.03) (table 1 and figure 1B). Although our tobacco compliance index itself was not significantly correlated with a decrease in prefectural level lung cancer mortality rates, smoking decrease (as proxied by the decrease in per capita tobacco consumption as well as percentage decrease in tobacco consumption at the prefecture level) was positively associated with the decrease in lung cancer mortality (table 1 and figure 1C).
Characteristics and correlations of each indicator (tobacco control policy compliance, trends in smoking prevalence, trends in lung cancer mortality and tobacco leaf production) in 47 prefectures, Japan
Notably, prefectural level tobacco leaf production was inversely associated with our tobacco compliance index (r=−0.39 and p=0.01)—that is, prefectures that produced a larger amount of tobacco leaf had worse compliance (table 1 and figure 1D). Moreover, tobacco leaf production was inversely correlated with smoking prevalence decrease—that is, prefectures that produced a larger amount of tobacco leaf had a smaller decrease in smoking (table 1).
Scatter plots and (non-linear) relations between the prefecture-level involvement in tobacco leaf production and each indicator are shown in figure 2A–C. As shown in table 1, although the relation was not linear, prefectures that produced a greater volume of tobacco leaf exhibited worse compliance (figure 2A). An inverse association between tobacco leaf production and decrease in tobacco consumption was also observed (figure 2B). However, the relation between tobacco leaf production and decrease in lung cancer mortality was U-shaped (figure 2C).
(A) A scatter plot and non-linear relation between the prefecture-level involvement in tobacco leaf production and tobacco control policy compliance. A curve of natural spline (4 degree of freedom) (solid line) and its 95% CI (broken lines) are shown. (B) A scatter plot and non-linear relation between the prefecture-level involvement in tobacco leaf production and per capita tobacco decrease. A curve of natural spline (4 degree of freedom) (solid line) and its 95% CI (broken lines) are shown. (C) A scatter plot and non-linear relation between the prefecture-level involvement in tobacco leaf production and percentage decrease of lung cancer mortality. A curve of natural spline (4 degree of freedom) (solid line) and its 95% CI (broken lines) are shown.
Table 2 shows the prefectures categorised according to three levels of tobacco leaf production. The means as well as the p values for the analysis of variance and trend are shown. The profiles of indicators for individual prefectures are shown in supplementary tables 1–3. Kanagawa prefecture demonstrated the highest compliance index (5.21). As demonstrated in figures 2A,B, prefectures with higher levels of tobacco leaf production exhibited worse indicators of compliance and smoking reduction. On the other hand, decrease in lung cancer mortality was smallest in the middle production group.
Characteristics of each indicator (tobacco control policy compliance, trends in smoking prevalence and trends in lung cancer mortality) separated by tobacco leaf production categories in 47 prefectures, Japan
Discussion
To our knowledge, this is the first paper from Japan to examine prefectural disparities in compliance with national tobacco control legislation. We draw three broad conclusions from our analyses: first, tobacco control in Japan is not a monolithic entity, notwithstanding the fact that the Health Promotion Law was uniformly applicable throughout the nation. Instead, we found prefectural disparities in the extent of enforcement and compliance with legislation. Second, we found that the extent of prefectural compliance with national legislation is highly correlated with prefectural involvement in tobacco growing. Here again, it has not been widely reported or appreciated that Japan exhibits a similar regional pattern of tobacco control ‘culture’ to the USA, based upon historical patterns of agriculture. According to Gilpin et al,5 US tobacco-growing states belong to the lowest tertile of tobacco control, presumably because of their significant economic dependence on tobacco. Indeed, it is well known that one Diet (Japanese national parliament) member of Kumamoto (a medium-scale prefecture in southern-western part of Japan), where tobacco production was the second largest in Japan, lobbied in favour of the tobacco agriculture industry.17 However, no quantitative analyses are available in Japan that link donations from tobacco interests and political behaviour. Third, we found that residents of tobacco-growing prefectures exhibit slower declines over time in smoking prevalence. While reverse causality cannot be ruled out (ie, tobacco control policies are weak in places where population rates of smoking are high), our findings are also consistent with the notion that residents of the places where tobacco policy compliance is low pay a price in terms of slower decline in smoking prevalence. This aspect of health disparity has not hitherto received attention in tobacco policy debate in Japan.
We did not find that the degree of prefectural compliance was correlated with decreases in lung cancer mortality. There could be several reasons for this. First, although the Health Promotion Law also mandates restriction of secondhand smoke in the workplace, we could not obtain estimates of workplace smoking restrictions at the prefecture level. Second, we lacked data on other tobacco control attempts targeting active smoking (eg, education, youth tobacco control, and dissemination of tobacco cessation programmes) emphasised in the Healthy Japan 21 initiative. Future studies should not only incorporate indicators of these attempts but also attempt to differentiate the effects of various forms of tobacco control policies (youth access, smoking bans, cigarette taxes, quit-programme promotion, etc). Third, there may have been insufficient time elapsed for the legislation to affect lung cancer mortality rates. We selected lung cancer mortality among younger adults, because as Doll and other researchers suggested,6 18 cancer trends in younger adults tend to reflect relatively recent changes in the exposure to carcinogenic agents and are not confounded by the effects of exposures in the distant past. For example, Barnoya and Glantz observed a beneficial effect of tobacco control programmes on lung cancer incidence after a 1-year lag,19 and several cohort studies have shown that lung cancer risk began to decrease within a few years after tobacco cessation—for example, 40% reduction in less than 2 years20 and 33% within 2–4 years.21 Hence, the present finding that reduced tobacco consumption was positively associated with the decrease in lung cancer mortality is plausible. Monitoring the trends in lung cancer mortality rates in young adults could thus be an effective strategy for evaluating progress in tobacco control programmes in Japan.
Our study has several limitations. Mortality data can be statistically unstable especially for smaller prefectures (data not shown). However, higher tobacco consumption at the prefecture level in 2001 was cross-sectionally correlated with greater lung cancer mortality in 2001, consistent with a previous Japanese study,13 indicating the validity of our mortality measures. Proportions of current smokers and former smoker were calculated by self-report, thereby introducing potential misclassification which may have biased the associations that we examined towards the null. Finally, some of the indicators were measured in different time periods based on data availability—that is, tobacco production was measured in 2003, the compliance indicators were measured in the period from 2008 to 2010 and trends in smoking prevalence and lung cancer mortality were measured during the period from 2001 to 2007. Caution is therefore warranted in interpreting the relations between compliance indicators and smoking/lung cancer trends. In particular, the associations cannot be interpreted to mean that poor policy compliance causes more lung cancer or reduced declines in smoking. Our findings could be equally consistent with higher population smoking rates leading to lagging policy compliance.
Although the Japanese government—particularly aided by consumer movements and public health advocacy—has made important strides in recent years towards tobacco control, our findings suggest that important prefectural disparities remain. Moreover, these disparities reflect historical patterns of involvement in the tobacco growing business. It has been previously suggested that the Ministry of Finance, Japan Tobacco, Inc and tobacco retailers together form a triad of political interests arrayed against tobacco control policy formation in Japan.2 This study adds yet a third group to that equation—namely, the interests of local tobacco growers. That agricultural interests can affect local policy dynamics in Japan is illustrated by the anecdote of the Diet member of Kumamoto. The findings of our statistical analyses further suggest that in order to succeed, tobacco policy needs to incorporate countermeasures to support not only labourers in the tobacco sector of the prefectural economy (distribution, marketing, cigarette manufacture, etc) but also tobacco farmers so that they can produce alternative crops in an economically viable manner, as recommended by the Article 17 of WHO Framework Convention on Tobacco Control.22
What this paper adds
We examined prefectural disparities in compliance with national tobacco control legislation (the 2003 Health Promotion Law) in Japan. We also examined whether prefectural involvement in growing tobacco was associated with lower compliance with the law.
We draw three broad conclusions from our analyses: first, tobacco control in Japan is not a monolithic entity, notwithstanding the fact that the Health Promotion Law was uniformly applicable throughout the nation. Second, we found that the extent of prefectural compliance with national legislation is highly correlated with prefectural involvement in tobacco growing. Third, we found that residents of tobacco growing prefectures exhibit slower declines over time in smoking prevalence and residents of the places where tobacco policy compliance is low pay a price in terms of slower decline in smoking prevalence.
Acknowledgments
We thank Toshihide Tsuda, Masaaki Yamaoka and Jun Hamada for their valuable advice and support of this study.
Supplementary materials
Web Only Data
Files in this Data Supplement:
Footnotes
Funding The study was funded by a grant for young researchers from Okayama University.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.