The impact of smoking and quitting on household expenditure patterns and medical care costs in China

Transcription

1 Researh paper Appendies are published online only at tobaoontrol.bmj.om/ ontent/vol18/issue2 1 Center for Health Statistis and Information, Ministry of Health, Beijing, PR China; 2 International Health Researh Group, Liverpool Shool of Tropial Mediine, Liverpool L3 5QA, UK; 3 Institute of Health and Soiety, Newastle University, Newastle upon Tyne NE2 4HH, UK Correspondene to: Dr Julia Alison Crithley, Institute of Health and Soiety, Newastle University, Leeh Building, Newastle upon Tyne NE2 4HH, UK; j.a.rithley@nl. a.uk Reeived 28 July 2008 Aepted 9 January 2009 Published Online First 21 January 2009 This paper is freely available online under the BMJ Journals unloked sheme, see tobaoontrol.bmj.om/info/ unloked.dtl The impat of smoking and quitting on household expenditure patterns and medial are osts in China Y Xin, 1 J Qian, 1 L Xu, 1 S Tang, 2 J Gao, 1 J A Crithley 3 ABSTRACT Bakground: Smoking remains very ommon in Chinese men, and the eonomi burden aused by igarette onsumption on smokers and their families may be substantial. Using a large nationally representative household survey, the third National Health Servies Survey (NHSS, 2003), we estimated the eonomi impat of smoking on households. Methods: Smoking status of all household members (over 15 years) was olleted by interview for the NHSS, and households lassified into one of seven ategories based on their smoking status. Information on household inome and expenditure, and use of health servies was also obtained. We assessed both the diret osts (reduing funds available for spending on other ommodities suh as food, eduation, medial are, et, using a frational logit model), and indiret osts (inreasing medial expenditures, using a log-linear model). Results: Every five pakets of igarettes onsumed per apita per month redues household spending on other ommodities, most notably on eduation (by about 17 yuan per apita per annum) and medial are (11 yuan). The effets are greatest among low-inome rural households. Households with quitters spend substantially more on medial are than never-smoking households (64 yuan for households with two or more quitters). Conlusions: If a household member smokes, there is less money available for ommodities suh as eduation and medial are. Medial are expenditure is substantially higher among households with quitters, as ill-health is the main reason for quitting smoking in China. Smoking impoverishes a substantial number of poorer rural households. Smoking prevalene remains very high among Chinese men. The 2002 national smoking survey estimated an ever-smoking rate of 66% for men and 3.1% for women. 1 Although Western igarette brands sold in ities an be more expensive, loal brands in rural areas are as heap as 2 3 yuan per paket (there were approximately 7 yuan to $US1 as of November 2008). An average prie of about 3.71 yuan per paket was reported by the Chinese statistial bureau in Although the ost of smoking inreased in real terms from the late 1980s until the late 1990s, 34 China s GDP has also risen signifiantly over reent years, and igarettes may have beome more affordable. Nevertheless, approximately 9 11% of rural households live below the poverty line in China (on less than the equivalent of $1 daily), 56 and the prevalene of smoking is higher in rural than in urban parts of China. 7 The eonomi burden of smoking on households in China ould therefore be substantial. 8 The diret onsumption of tobao has an opportunity ost, restriting the household budget available to spend on other goods and servies. Smoking may also result in indiret osts to households, partiularly exess medial expenditure for treatment of smokingrelated diseases and onditions, as a result of both ative smoking and exposure to environmental tobao smoke (ETS) among other family members. Few studies have examined the impat of tobao spending on household budgets in low or middle-inome ountries, 9 and even fewer in large, nationally representative samples. Tobao expenditure may ause substantial harm to poor households with limited inomes for food and basi needs. Analysis of data from a survey of approximately 4500 households in rural China showed that smoking expenses an harm other family members by reduing expenditures on basi needs suh as foods, utilities and durable goods. 9 The main aims of this study were therefore: 1. To examine the effets of tobao onsumption on other household expenditures 2. To estimate the exess medial spending attributable to smoking and exposure to ETS. METHOD Data soure This study used data derived from the third National Health Servies Survey (NHSS) onduted in 2003, a nationally representative survey whih overed both urban and rural populations, and olleted information on healthare utilisation and expenditure. The survey adopted multistage stratified luster random sampling, and the methodology has been published elsewhere The overall sample inluded 95 ounties (distrits), 475 townships (streets) and a total of 950 villages (resident ommittees). The basi sampling unit was the household, and 60 households were randomly sampled in eah village or resident ommittee. Enumeration was undertaken by seleted loal health workers who undertook faeto-fae interviews in the respondent s home. The survey olleted data on demographi and soioeonomi fators, self-reported health status, healthare utilisation, household inomes and expenditures on a variety of ommodities, medial expenditure, and behavioural fators suh as smoking and drinking at an individual level. A total of households, inluding people, took part. Over 99% of households initially seleted agreed to take part, and 99.6% of respondents answered questions on smoking status. In total, households answered questions on both smoking status and household expenditures, and were therefore inluded in our study. 150 Tobao Control 2009;18: doi: /t

2 Table 1 The number and perentage of households by smoking status, National Health Servies Survey (NHSS) survey National Urban Rural Household groupings No (%) No (%) No (%) Never-smoking households* (NS) (37.9) 7284 (43.4) (35.6) Low tobao onsumption (Low) (18.8) 3499 (20.9) 7205 (18.0) Medium tobao onsumption (Medium) (26.6) 3638 (21.7) (28.7) High tobao onsumption (High) 7702 (12.1) 1440 (8.6) 5465 (13.6) One former smoker{ (1 quit) 1049 (1.8) 341 (2.0) 708 (1.8) Two or more former smokers{ (2+ quit) 934 (1.6) 389 (2.3) 545 (1.4) Mixed (smokers and quitters){ (Mixed) 594 (1.0) 192 (1.1) 402 (1.0) Total (100) (100) (100) *No urrent or former smokers in the household, NS (household of never-smokers). {No urrent smokers but there are one (1 quit) or at least two former smokers (2+ quit) who have quit smoking in the household. {At least one urrent smoker, and also at least one former smoker who has quit in the household. Low, low tobao onsumption (0 25% of distribution of household tobao onsumption); Medium (25 75%); and High tobao onsumption (75 100%). The ategories are mutually exlusive ie, mixed households with smokers and quitters are not also ounted as low, medium or high onsumption households. Smoking status of households Those who had smoked at least five paks of igarettes during their lifetime (100 igarettes in total), and reported smoking at the time of the survey are defined as smokers in this study. Former smokers are defined as those who had smoked at least 100 igarettes over their lifetime, but had stopped by the time of the survey. All urrent smokers were asked how many igarettes they onsumed daily (oded as the exat number). All members of the household over 15 were interviewed; we an therefore estimate total household tobao onsumption. All households were divided into seven mutually exlusive ategories aording to the smoking status of family members (see table 1). Never-smoking households are those with no urrent or former smokers. Households with at least one urrent smoker (and no former smokers) were lassified based on the total household onsumption of igarettes (in perentiles). Those in the 0 25 perentile of tobao onsumption were defined as lowtobao onsumption households. Likewise, those in the perentile or perentiles were ategorised as moderate or high-tobao onsumption households. Households with former smokers only were ategorised aording to the number of former smokers (one, two or more). Households with both urrent smokers and quitters were lassified as mixed onsumption (table 1), and not inluded in the low, medium or high ategories. Exposure to ETS is a serious problem in China. 1 ETS ours in the workplae and other publi plaes, and in the home. We ould only assess exposure to ETS at home, as the NHSS requested only limited information on exposure elsewhere. The survey asked whether there were any restritions on smoking in the household so, onservatively, we defined ETS exposure as ourring among households with at least one urrent smoker, and no smoke-free poliies in the home. ANALYSIS Diret osts of smoking household onsumption and expenditure patterns Initially, we alulated the perentage of total household expenditures spent on seven distint expenditure ategories (food, lothing, transportation, housing, eduation, medial are, and others). These seven ategories are exhaustive and mutually exlusive. Table 2 gives the rude mean expenditure for these seven ategories, by household smoking status. Expenditure on tobao was not diretly olleted by the survey, so is inluded in the Others ategory. Researh paper Our aim was to estimate how household spending on tobao affets spending on other goods and servies, after ontrolling for other household harateristis. Beause a ategory for expenditure on smoking was not inluded in the NHSS, we annot estimate smoking osts diretly. The amount of tobao onsumed by the households was therefore the preditor variable (per five pakets, or 100 igarettes per month). Our outome variable is the perentage of total household expenditure alloated to eah ategory; we therefore used the frational logit model to predit the effet of the number of igarettes smoked on household expenditure patterns after ontrolling for other fators Using the frational logit model, the predited values of these perentages will be bounded by the unit interval [0, 1]. Speifially onfounders onsidered a priori and inluded were household head s age, gender, marital status, level of eduation, oupation, insurane and self-reported health status; household loation (urban or rural area), household inome, number of household members, household members under 15 years or over 65 years and the number of former smokers (see appendix 1). We ran the following model on eah of the seven household expenditure ategories to predit the household tobao onsumption effets: where L represents the logit funtion, Y represents the perentage of expenditures spent on eah of our seven ategories, X represents tobao onsumption, as well as our ontrol variables and m i is an error term. After onstrution of the predition model for eah expenditure ategory, a reyling predition method was used to predit the perentage for eah ategory based on the estimated oeffiients. 17 Inrease in medial expenditures assoiated with smoking and ETS exposure There are two main approahes to assess the additional medial spending attributable to smoking inlusive and disease-speifi. The disease-speifi approah attempts a priori to Tobao Control 2009;18: doi: /t

3 Researh paper Table 2 Desriptive statistis on household expenditure (yuan and %) by household tobao onsumption NS Low Medium High 1 Quit 2+ Quit Mixed Average household onsumption by smoking status Number of households Number of individuals Expenditure by ategory (yuan)* Food Clothing Transportation Housing Eduation Medial are Others Perentage of total household expenditure for eah ategory Food Clothing Transportation Housing Eduation Medial are Others *There are approximately $68, or J8.9 to 1 yuan (data as of November 2008). NS, household of never-smokers; Low, low tobao onsumption (0 25% of distribution of household tobao onsumption); Medium (25 75%); High tobao onsumption (75 100%); 1 Quit, household with one quitter only and no urrent smokers; 2+ Quit, household with two or more quitters and no urrent smokers; Mixed, household with both urrent smokers and quitters. identify smoking-related diseases and their osts, 20 or uses the absolute death rates from lung aner as an indiator of tobaorelated deaths. 21 The inlusive approah developed here reognises that the health effets of smoking are omplex and multiple, and assesses the impat of smoking on all medial spending. This latter does not require aurate disease diagnosis (whih may not be available in rural parts of China). Crude and age-standardised analyses demonstrated higher health servie use and expenditure among those who had quit smoking, but not among urrent smokers, from the NHSS data. To examine the effet of smoking, the following regression equation of medial are utilisation is estimated using a loglinear approah 19 : where Y is the annual self-reported medial expenditure of a household, whih might be zero (hene the need to add 1 before logging); CS and FS are dummy variables equal to 1 if the individual is a urrent smoker (CS) or former smoker (FS); and X is the vetor of demographi and individual ovariates with oeffiient vetor B. For ETS, a similar regression equation is estimated: where PS is a dummy variable, equal to 1 if the individual lives in a household in whih there is at least one urrent smoker, and no smoke-free household poliy; and X is the vetor of the same demographi and individual ovariates as before with oeffiient vetor B. We have therefore onservatively assumed that there are no medial osts relating to ETS among households with former smokers. We used the log-linear model to ontrol for respondents age, gender, marital status, level of eduation, oupation, health insurane status (yes, no) and drinking status (see appendies 2 4). Household level variables we ontrolled for inluded household loation (urban or rural area), household inome and the distane from respondent s home to the nearest medial institution. In both of the models, our referene ategory is households that ontained only never-smokers. Additionally, as the NHSS survey only olleted data on smoking status from those aged 15 and above, we only inluded individuals over 15 in the referene ategory when estimating the medial are expenditures assoiated with smoking. We then predited the value of the dependent variable for eah urrent and former smoker using the oeffiient estimates from the regression equations. To estimate the inrease in medial expenditure assoiated with smoking, we used the same oeffiient estimates but set the CS and FS variables to zero. Then, the predited expenditure of the smokers (assuming they had never smoked) is subtrated from the predited expenditure. The differene of these two expressions is the predited effet of smoking on medial are expenditure. RESULTS Assoiation between tobao onsumption and other household expenditure Table 2 shows rude expenditure on different ategories (yuan and %) by smoking status, showing that households with quitters spent more on medial are, and households with smokers less on eduation. We estimated the marginal propensity to spend on other expenditure ategories in relation to tobao onsumption in all surveyed households using the frational logit model. Tobao onsumption was negatively and signifiantly related to spending on eduation, and medial are (table 3). Consuming every five paks of igarettes (100 igarettes in total) per apita per month was assoiated with 16.6 (5% less) yuan, and 10.9 yuan delines (4% less) in the above onsumption ategories, respetively. 152 Tobao Control 2009;18: doi: /t

4 Table 3 The marginal effets of tobao onsumption on other household expenditures in the total survey population, and the poorest 20% of rural households in the survey Marginal effets b % Total inome In smoking households, the atual redution may be muh higher as the average household with one male smoker onsumed around 450 igarettes per month in our survey. 22 The results also indiated that tobao onsumption was positively and signifiantly assoiated with inreased spending on transportation, food and spending in the other ategory, but the perentage hanges were very small (,1%, table 3). Costs of tobao are inluded in the other ategory, so spending on this ategory inreased with higher tobao onsumption. In fat, the only variable not assoiated with tobao onsumption is housing, presumably beause these generally are fixed osts for households. Assoiation between tobao onsumption and other household expenditure among the poorest 20% of rural households We repeated our analyses limiting the dataset to the poorest 20% of rural households. Tobao onsumption was assoiated with 5.1 yuan (4%) and 6.4 yuan (4%) delines in spending on eduation and medial are respetively (table 3). The average inome of the poorest 20% of rural households was only 709 Money value (Per apita per year (yuan)) Perentage of total expenditure on ategory A Total population Food , Clothing , Transport , Housing Eduation , Medial are , Others , B Poorest 20% of rural households Food , Clothing Transport , Housing , Eduation , Medial are , Others , The marginal effets of tobao spending (per 1000 yuan) on other household expenditures in (a) total dataset (n = households and individuals; 77 missing households), and (b) poorest 20% households in rural China in 2003 (n = households; individuals; 36 missing households). The unit of amount of igarette onsumption is five paks per apita per month (100 igarettes, approximately three per day) in a household. yuan per apita per year, so though absolute monetary values are small, the negative effet on households was quite severe. Medial spending assoiated with igarette smoking Ative igarette smoking Appendies 1 3 (see website) display rude data and results of regression models quantifying assoiations between household smoking and medial expenditure. Medial expenditure is 2.7% higher (p = 0.04) among urrent smokers ompared with neversmokers, and 27.0% higher among former smokers (p,0.0001) (table 4 and appendix 2 (see website)). Households with former smokers experiene the greatest inrease in medial osts, espeially those with two or more former smokers (table 4), an additional 64.0 yuan per apita per year (84.9 in urban areas; 51.6 in rural areas). Environmental tobao smoke (ETS) We also estimated medial spending as a result of ETS using the same method, exept that we inluded the whole population in Table 4 Exessive medial spending attributable to ative smoking and exposure to environmental tobao smoke (ETS) by smoking status (yuan, per apita per year) Smoking status National Urban Rural Ative smoking NS (referene ategory) Low Medium High quit quit Mixed ETS Low Medium High NS, household of never-smokers; Low, low tobao onsumption (0 25% of distribution of tobao onsumption); Medium (25 75%) and high tobao onsumption (75 100%). SE Researh paper p Value Tobao Control 2009;18: doi: /t

5 Researh paper the referene ategory (households with never-smokers), regardless of age. Appendix 3 (see website) shows that among households with exposure to ETS there is a statistially signifiant (p,0.0001) 7% higher medial expenditure among non-smoking household members, ompared with households with no ETS exposure. For low-tobao onsumption households, the average exessive medial spending attributable to ETS is 13.5 yuan per year (with similar results for moderate and high onsumption households, table 4). DISCUSSION Tobao onsumption drives out spending on basi household ommodities, espeially eduation, and on medial are. Our results reinfore those published from other Chinese surveys. A 2002 rural health insurane survey in two poor rural provines in China showed that rural households spent 6.5% of their total expenditure on tobao, and showed that for every 100 yuan spent on tobao, there was a 30 yuan derease in spending on eduation and a 15 yuan derease on medial are A further study found that poor urban households spent 6.6% of their expenditure on igarettes, and 11.3% in poor rural areas. Again, tobao spending was assoiated with redutions in spending on other goods and servies. 24 This finding is a partiular onern as expenditures on eduation and medial are both tend to inrease produtivity that is, are human apital investments. As eduation expenditures mainly benefit hildren (from shool fees), it is oneivable that hildren in poorer smoking households may be fored to give up some eduation. This has partiular onerns for China s ontinued eonomi and soial development. Tobao onsumption was assoiated with higher food osts, but the food ategory inluded the osts of alohol, and smokers are far more likely to drink alohol regularly than non-smokers (rude OR = 8.99, 95% CI 8.63 to 9.37), whih may aount for this. Tobao onsumption was also assoiated with slightly inreased spending on transport; it is possible that this is beause of the assoiation of smoking with soial ativities in China. 25 There are two main onlusions from our analyses of medial expenditure by smoking status. One is that households with former smokers have very high medial spending, and the seond is that ETS exposure is also assoiated with raised medial are expenditures. These osts are proportionally greater for the poorest 20% of rural households, where exessive medial spending due to smoking as a perentage of inome is 3.5 times that for the highest inome quintile. Frational logit modelling estimating the diret osts of smoking indiated that inreases in tobao onsumption were assoiated with redutions in spending on medial are as a perentage of total household expenditure, while log-linear models estimating the indiret medial are osts assoiated with tobao found that households with smokers had slightly inreased absolute medial are osts. Although we ontrolled for household inome within quintiles in the log-linear models, there is likely to be some residual onfounding as even within the poorest quintile (and deile) of rural households, smoking households still had slightly higher inomes than non-smoking households, allowing for differenes in perentage and absolute tobao expenditure to arise. Partiularly in rural areas, it is likely that those in the very poorest households do not smoke beause of the eonomi burden. The appropriate interpretation is therefore that tobao onsumption redues medial expenditure, but those in slightly wealthier households (even among poorer rural areas) are more likely to smoke, and also spend more on medial are, probably beause of their slightly higher inomes. It may seem surprising that the medial are osts of urrent smoking households are not higher, although the osts among households with former smokers were very high. Only diret household osts were olleted by the survey. Smokers with health insurane, and in some ases their families, may be partly sheltered from the diret osts of exessive medial expenditure due to smoking-related diseases. This is unlikely to be a major fator though, as health insurane overage was low in 2003 (nearly half of the urban population had no overage at all), 13 and out-of-poket payments represent a high proportion of total healthare expenditure in China (59% in 2000); in partiular outpatient servies (inluding primary are) are not overed at all in most parts of China. 26 Probably more importantly, health servie utilisation data do not neessarily reflet health need, and there is lear evidene from the NHSS that smoking households are less likely to obtain seondary medial are, even when this has been reommended by a dotor (data not shown). Finally, the survey was ross-setional in design; smokers who beome ill may deide to quit, resulting in a healthy smoker bias. 22 Our data soure was a very large, nationally representative household survey, the NHSS. We believe this is one of the first attempts to estimate the household-assoiated osts of smoking from a nationally representative survey in a developing world ountry. However, our analyses have several limitations. The main limitation is that the NHSS survey did not ollet data on spending on igarettes diretly; we therefore used household tobao onsumption as the dependent variable. Cigarette pries vary onsiderably in China, and it has been reported elsewhere that while poor rural households overwhelmingly hoose heap loal brands, riher urban households may purhase more expensive brands. We ould have estimated igarette prie using data from other Chinese surveys, but we preferred to use the tobao onsumption data, olleted by the NHSS. Critially, we found the expeted dose-response relation in both urban and rural areas inreasing ategories of tobao onsumption redued spending further on other ategories, suh as eduation and medial are. However, it is possible that our approah may underestimate the effets of smoking osts on wealthier urban households, if these are purhasing more expensive igarette brands. We may also underestimate the diret osts of smoking through mislassifiation of smoking osts as some households, ategorised as low onsumption may be buying more expensive igarettes, while some ategorised as high onsumption may be buying heap igarettes. Most of our survey data (on igarette onsumption, household inomes and medial are osts) were self-reported, and thus may be subjet to bias. However, we used trained loal interviewers, and were able to validate some inome data against household benefits and expenditures. Finally, our data are ross-setional; even if smokers quit we do not know how households would use potential savings. Reduing smoking is not only ritial in improving health, it also has an important role in poverty redution in China, partiularly in poorer rural areas. As well as the diret opportunity osts of smoking, reduing expenditure on eduation, the ost of medial are has been inreasing rapidly in China over reent years ompared with inome, and the average ost of a single hospital admission is over double the average annual inome of the poorest 20% of the population. 26 It is therefore likely that the household medial osts assoiated with adult smoking have inreased sine the NHSS was arried 154 Tobao Control 2009;18: doi: /t

6 What this paper adds Households with urrent smokers spend signifiantly less on other goods and servies, partiularly eduation and medial are. Household medial are expenditure osts are substantially higher among households with one or more ex-smoker. Exposure to environmental tobao smoke is also assoiated with inreased medial are osts. The effets are most aute among poorer rural households, ontribute to impoverishment and may have a substantial effet on China s ontinued eonomi and soial development. out in 2003; and even if smoking falls, in absolute terms medial osts as a result of smoking are likely to rise further. It has been estimated that 35% of urban households and 43% of rural households have diffiulty affording healthare, go without healthare, or are impoverished by health osts. 28 Insurane overage for most of the Chinese population is also inadequate. 26 We found partiularly high medial are osts among households with quitters; the true osts are probably far higher as many poorer smokers may be foregoing the healthare they need. 22 Poliy initiatives should fous partiularly on rural areas in China, where smoking prevalene and poverty are higher, and where sustained anti-smoking ampaigns have been rare. Information on the opportunity osts of smoking should be widely disseminated in these areas. Smoking essation advie and niotine replaement therapies are not generally available in rural areas; their use should be promoted by dotors and overed by health insurane shemes. Extending and enforing smokefree poliies in publi plaes and, more ontroversially, inreasing taxes, have both been shown to redue smoking prevalene in similar areas, and may have both eonomi and health benefits for China In China, assuming a prie elastiity of 20.54, it has been estimated that a 40% inrease in tax from 1.60 yuan per pak to 2.00 yuan tax per pak would redue onsumption by 4.57 billion paks, generate additional entral government revenues of nearly 25 billion yuan, and save million lives. 30 The inrease in entral government tax revenue would be twie as large as the total losses to industry and agriulture. 16 Our study suggests it would also redue poverty and inrease resoures available for household spending on eduation and medial are, important for China s future development. Aknowledgements: We thank Mrs Pat Barker for seretarial assistane. Funding: International Development Researh Centre, Canada. Competing interests: None. REFERENCES 1. Yang GH, Ma JM, Liu N, et al. Smoking and passive smoking in Chinese,2002. Chinese Aad Med Si 2005;26: Researh paper 2. Wang H. Tobao ontrol in China: the dilemma between eonomi development and health improvement. Salud Publia Mex 2006;48(suppl I):S Hu T-W, Mao Z. Eonomis analysis of tobao and options for tobao ontrol. China ase study. In: Preker AS, ed. Health, nutrition and population (HNP) disussion paper. Washington, DC: The International Bank for Reonstrution and Development/ The World Bank. %20China%20Case%20Study-whole.pdf, aessed Deember Guindon GE, Tobin S, Yah D. Trends and affordability of igarette pries: ample room for tax inreases and related health gains. Tob Control 2002;11: Rural Survey Assoiation of National Bureau of Statistis. Poverty Statistis in China. 4_Poverty%20Statistis%20in%20China.pdf, aessed September Woo WT, Shi L, Ximing Y, et al. The poverty hallenge for China in the new millennium. Report to the Poverty Redution Taskfore of the Millennium Development Goals Projet of the United Nations: faulty/woo/1%20ot%2004.woo.china%20poverty.pdf, aessed 1 Otober Yang G, Fan L, Tan J, et al. Smoking in China: findings of the 1996 National Prevalene Survey. JAMA 1999;282: Tobao Free Initiative, University of California. Tobao and poverty: a viious irle. Tobao Control, WHO Tobao Control Papers, org/ontext/t/artile/1149/type/pdf/viewontent/ 9. Wang H, Sindelar JL, Bush SH. The impat of tobao expenditure on household onsumption patterns in rural China. So Si Med 2006;62: Gao J, Qian J, Tang S, et al. Health equity in transition from planned to market eonomy in China. Health Poliy Planning 2002;17(suppl): Gao J, Tang S, Tolhurst R, et al. Changing aess to health servies in urban China: impliations for equity. Health Poliy Planning 2001;16: Tang S, Li X, Wu Z. Rising esarean delivery rate in primiparous women in urban China: evidene from three nationwide household health surveys. Am J Obstet Gyneol 2006;195: Xu L, Wang Y, Collins C, et al. Urban health insurane reform and overage in China using data from national health servie surveys in 1998 and BMC Health Serv Res 2007;7(37). 14. Papke L, Wooldridge JM. Eonometri methods for frational response variables with an appliation to 401(k) plan partiipation rates. Tehnial Working Paper No 147. Cambridge, MA: National Bureau of Eonomi Researh, Regression models for quality of life outome. Third International Health Eonomi Assoiation Conferene, York, UK, Gould BW, Villarreal HJ. An assessment of the urrent struture of food demand in urban China. Agr Eons 2006;34: Yip WC, Wang H, Liu Y. Determinants of patient hoie of medial provider: a ase study in rural China. Health Poliy Planning 1998;13: Cutler DM. How good a deal was the tobao settlement? Assessing payments to Massahusetts. Cambridge, MA: National Bureau of Eonomi Researh, Liu Y, Rao K, Hu T-W, et al. Cigarette smoking and poverty in China. So Si Med 2006;63: MGhee SM, Ho LM, Lapsley HM, et al. Cost of tobao-related diseases, inluding passive smoking, in Hong Kong. Tob Control 2006;15: Peto R, Lopez AD, Boreham J, et al. Mortality from tobao in developed ountries: indiret estimation from national vital statistis [see omment]. Lanet 1992;339: Qian J, Cai M, Gao J, et al. Trends in smoking and quitting in China from 1993 to 2003: National Health Servie Survey. Am J Pub Health 2008 (in press). 23. Wang H, Zhang L, Hsiao W. Ill health and its potential influene of household onsumptions in rural China. Health Poliy 2006;78: Hu T, Mao Z, Liu Y, et al. Smoking standard, of living, and poverty in China. Tob Control 2005;14: Yang T, Li F, Yang X, et al. Smoking patterns and soiodemographi fators assoiated with tobao use among Chinese rural male residents: a desriptive analysis. BMC Publi Health 2008;8: Hu S, Tang S, Liu Y, et al. Reform of how health are is paid for in China: hallenges and opportunities. Lanet 2008;372: Hesketh T, Wang H, Ye X, et al. Smoking, essation and expenditure in low inome Chinese: ross-setional survey. BMC Publi Health 2007;7(29). 28. Liu Y, Rao K, Hu SL. People s Republi of China: toward establishing a rural health protetion system. Asian Development Bank, Fihtenberg CM, Glantz SA. Effet of smoke-free workplaes on smoking behaviour: systemati review. BMJ 2002;325: Hu TW, Mao Z. Effets of igarette tax on igarette onsumption and the Chinese eonomy. Tob Control 2002;11: Tobao Control 2009;18: doi: /t