JECH Online First, published on February 27, 2013 as 10.1136/jech-2012-201859 Research report
Impacts of alcohol duty reductions on cardiovascular mortality among elderly Chinese: a 10-year time series analysis Vivian C Pun,1 Hualiang Lin,2 Jean H Kim,1 Benjamin H K Yip,1 Vincent C H Chung,1 Martin C S Wong,1 Ignatius T S Yu,1 Sian M Griffiths,1 Linwei Tian1 1
The Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong, Prince of Wales Hospital, Sha Tin-NT, Hong Kong 2 Center for Disease Control and Prevention of Guangdong Province, Guangzhou, Guangdong, China Correspondence to Professor Linwei Tian, 4/F, The Jockey Club School of Public Health and Primary Care, The Chinese University of Hong Kong, Prince of Wales Hospital, Sha Tin-NT, Hong Kong;
[email protected] Received 20 August 2012 Revised 4 January 2013 Accepted 10 February 2013
ABSTRACT Background In March 2007, the Hong Kong Government halved its heavy excise taxes on beer and wine, and 1 year later, it eliminated all duties on these beverages. This study examines the impact of such duty reductions on cardiovascular disease (CVD)-related mortality among the elderly in Hong Kong. Methods Box-Jenkins autoregressive integrated moving average intervention time series analyses were applied to monthly morality data from 2001 to 2010 to quantify the impacts of duty reduction and exemption on CVD death rates among those aged 65 years or older. Results The alcohol duty reduction in March 2007 was associated with an estimated 13% increase (95% CI 2% to 24%) in CVD death rates among elderly men, after controlling for the other intervention, outlier, trends and seasonal variations. This was equivalent to an extra 11 CVD deaths per 100 000 elderly men each month. Much of the observed impacts on CVD death rates were found to have contributed only by that on ischaemic heart disease mortality (18% increase in rate for men (95% CI 4% to 34%); 15% increase for women (95% CI 0.4% to 31%)), not by mortality due to stroke or hypertension. The alcohol duty exemption on March 2008 was not found to have impacted the CVD death rates. Conclusions The increase in CVD death rates among the Chinese elderly after alcohol duty reduction suggest that the purported beneficial effect of moderate alcohol use may not apply to certain Chinese populations, adding fuel to the ongoing debate on the risks and benefits of moderate alcohol consumption on mortality.
BACKGROUND
To cite: Pun VC, Lin H, Kim JH, et al. J Epidemiol Community Health Published Online First: [ please include Day Month Year] doi:10.1136/jech-2012201859
The debate of the health risks and benefits of alcohol consumption, which began over a century ago in Europe, continues today in the global arena.1 With greater globalisation of alcohol trade, the public health effects of alcohol consumption are increasingly being directed towards regions with traditionally low alcohol consumption. Considerable international research has also shown that high alcohol consumption is not only associated with increase in risky behaviours and injuries, but also it is shown to be harmful to both the liver and circulatory system.2–4 A large body of evidence has demonstrated that low-to-moderate alcohol consumption is associated with lower risk of mortality due to cardiovascular diseases (CVD), especially coronary heart disease.2 5 Nonetheless, a majority of the evidence comes from observational studies of Western populations (ie, the USA, Canada and Europe). In Asia,
PunCopyright VC, et al. J Epidemiol Community Health doi:10.1136/jech-2012-201859 Article author (or2013;0:1–5. their employer) 2013. Produced
cohort studies of the elderly population have found selective cardioprotective effects of moderate alcohol consumption.6–12 Most of the studies conducted in Hong Kong, however, showed no beneficial effect of moderate alcohol consumption on CVD morbidity and mortality.13–15 A meta-analysis of 112 studies documented that a 10% decrease in alcohol prices results in an approximately 5% increase in alcohol drinking.16 A number of intervention studies focused on determining the impact of alcohol policies (ie, duties and prices) on health outcomes. A majority of the studies have examined these impacts on alcoholrelated mortality and morbidity, and one also examined CVD mortality.17–24 Upon finding beneficial impacts on CVD mortality, Herttua et al17 suggested that cheaper alcohol may have fostered a moderate consumption and its beneficial effects among the older population in Finland. All intervention studies on the impacts of alcohol duties/ prices on mortality were conducted in Western countries. No time series intervention studies addressing the relationship between alcohol duties/ prices and CVD mortality have been conducted in Asia. It is therfore uncertain whether the magnitude and direction of the associations observed among Western populations can be directly applied to the general Asian population, given their lower serum cholesterol levels, lower body mass index and different overall CVD risk profile.9 Hong Kong, a duty-free port with a population of seven million, has traditionally levied heavy excise taxes (referred to as ‘duties’) on imports of alcoholic beverages.25 Since 2001, the ad valorem duties on alcoholic beverages were levied at 40% for beer, 80% for wine and 100% for spirits.26 In March 2007, the Hong Kong Government halved excise taxes on beer, wine and all other alcoholic beverages, except spirits, to help promote the development of the hospitality industry.27 In 2008, the Government went on to eliminate all duties on these beverages in order to promote Hong Kong.28 The elimination of alcohol duties have led to a rapidly increasing number of social drinkers in Hong Kong, as evidenced by year-on-year wine import growths of 80%, 45% and 73% for 2008–2009, 2009–2010 and 2010–2011, respectively.29 The current study takes advantage of this natural experiment in which the Hong Kong population experienced two successive large-scale alcohol duty reductions, leading to higher rates of social drinking, to assess their impacts on CVD death rates among the elderly.
by BMJ Publishing Group Ltd under licence.
1
Research report Table 1 Demographic description of all CVD deaths among the elderly in Hong Kong between 2001 and 2010
Age of death Average monthly CVD deaths (deaths per 100000) IHD Stroke Hypertension
Men
Women
79 341 (88.0) 145 (37.4) 116 (29.9) 25 (6.5)
83 380 142 131 35
(84.2) (31.3) (29.0) (7.8)
CVD, cardiovascular diseases; IHD, ischaemic heart diseases.
MATERIALS AND METHODS Study population According to the Government’s 2008 census statistics, 13% of the general population in Hong Kong were aged 65 years or older.30 More than one-third of the elderly population had attained primary education and another third had attained secondary or higher education. Their median personal income was HKD3300 (∼USD420) per month. Seventy per cent of the elderly population reported having chronic diseases, and a majority of them (63%) suffered from hypertension. Smoking prevalence in this population was low, as 11% and 13% were smokers and ex-smokers, respectively. Twelve per cent were alcohol drinkers and a third of them drank daily. The habit of regular alcohol drinking was much more prevalent among elderly men than women.
of two major alcohol duty interventions on CVD death rates that could not be accounted for by the trend due to normal cyclical fluctuations within the time series. Seasonal ARIMA models without constant term were specified to account for the inherent dynamics in the series, and are expressed in ( p, d, q) (P, D, Q). The p, d and q in the first parentheses specify the order of the autoregressive, differencing and moving average processes of the regular noise modelling, whereas the P, D and Q in the second parentheses specify the corresponding parameters for the seasonality component.31 The Ljung-Box Portmanteau (or Q-) test was used to test the randomness of residuals of the estimated model. Log transformation was applied to the death rate series to assure the normality and homogeneity of variance of the residuals. The two duty interventions were treated as dummy variables (0=preintervention period, 1=postintervention period) in the predictor list of the ARIMA model. The first dummy was assumed to take place on 1 March 2007 (50% duty reduction), and the second dummy on 1 March 2008 (duty exemption). The effect of each intervention was assessed by interpreting the coefficient β for the indicator variable; the per cent change in the postintervention period was estimated as exp(β)−1. If significant association(s) was found in either intervention, subsequent analysis was conducted to identify which disease(s) may explain that association. ARIMA model with maximum likelihood estimator from the TSA package in the R statistical software was used for the analysis.32 33
RESULTS
Mortality data Monthly mortality data, which covered all deaths in Hong Kong between 2001 and 2010, were obtained from the Hong Kong Census and Statistics Department. They were coded according to the 10th revision of the International Classification Diseases (ICD). Data on mortality due to CVD (ICD-10: I00–I99), excluding alcoholic cardiomyopathy (I426), among people aged 65 years or older at the time of death were extracted. CVD was further subdivided into ischaemic heart diseases (IHD) (ICD-10: I20–I25), cerebrovascular diseases/strokes (ICD-10:I60–I69) and hypertension (ICD-10:I10–I15). These disease outcomes were selected based on findings from previous alcohol-related studies. Gender-specific time series death rates (deaths per 100 000) were examined.
Statistical analysis Box-Jenkins autoregressive integrated moving average (ARIMA) intervention time series analysis was used to quantify the impact
Between 2001 and 2010, there were 86 592 CVD deaths among the elderly in Hong Kong, 53% of who were elderly women. Twenty-one deaths due to alcoholic cardiomyopathy were excluded. The average age of death was slightly higher among women than men (table 1). The average monthly death rates due to specific CVD outcomes, which accounted for 80% of all CVD deaths, are presented in table 1. Death rates by gender did not differ significantly in Hong Kong. Findings from the ARIMA intervention analysis that assessed the effects of two alcohol duty interventions on CVD death rates are shown in table 2. Model diagnostics suggested that the models provided a good fit to the data ( p>0.05). A statistical outlier on 1 February 2008 was detected after examining the time series and residual plots, and thus was adjusted to prevent distortion of the model specification and intervention effect estimates.34 Overall, the only significant effect resulting from the models was the effect of the March 2007 intervention, which was associated with an estimated 13% increase (95% CI 2% to
Table 2 Impacts of two alcohol duty interventions on monthly cardiovascular disease death rates among the elderly in Hong Kong, stratified by gender, 2001–2010, autoregressive integrated moving average (ARIMA) intervention models† Change Alcohol duty intervention March 2007 (50% duty reduction) Men Women March 2008 (duty exemption) Men Women
ARIMA model and noise
Percentage of change
95% CI
p Value
Specification
Q24a
p Value
12.6* 5.5
2.2 to 24.0 −3.5 to 15.3
0.02 0.24
(0,1,1) (0,1,1)12 (0,1,1) (0,1,1)12
24.3 32.1
0.45 0.12
10.3 2.4
−2.1 to 24.2 −7.9 to 13.8
0.11 0.66
(0,1,1) (0,1,1)12 (0,1,1) (0,1,1)12
24.3 32.1
0.45 0.12
*p Value
