| |
Environmental risk factor assessment: a multilevel analysis of childhood asthma in China
Fei Li, Ying-Chun Zhou, Shi-Lu Tong, Sheng-Hui Li, Fan Jiang, Xing-Ming Jin, Chong-Huai Yan, Ying Tian, Shi-Ning Deng, Xiao-Ming Shen
Shanghai, China
Author Affiliations: Department of Developmental and Behavioral Pediatrics, Shanghai Institute of Pediatric Translational Medicine, Shanghai Children's Medical Center, Shanghai Jiaotong University School of Medicine, Shanghai, China (Li F, Jiang F, Jin XM, Deng SN, Shen XM); Shanghai Key Laboratory of Children's Environmental Health, Shanghai Jiaotong University School of Medicine Shanghai, China (Li F, Li SH, Jiang F, Jin XM, Yan CH, Tian Y, Deng SN, Shen XM); The Key Laboratory of Children's Environmental Health, Ministry of Education, China (Li F, Li SH, Jiang F, Jin XM, Yan CH, Tian Y, Deng SN, Shen XM); Department of Statistics and Actuarial Sciences, East China Normal University, China (Zhou YC); School of Public Health and Institute of Health and Biomedical Innovation, Queensland University of Technology, Australia (Tong SL)
Corresponding Author: Xiao-Ming Shen, MD, 1678 Dongfang Road, Shanghai 200129, China (Tel: 86-21-38626161-6020; Fax: 86-21-38626161-6020; Email: xmshen@shsmu.edu.cn)
doi: 10.1007/s12519-013-0413-5
Background: Rapid changes in socioeconomic environ-ment and their diverse patterns in China raise a question: how socio-environmental factors affect childhood asthma in China. We performed a multilevel analysis based on a 2005 national survey to understand the association between environmental factors and asthma, and to provide insights on developing prevention strategies.
Methods: A multi-center, cross-sectional survey was conducted in 2018 school-aged children chosen from eight Chinese cities. Children of 6-13 years old were chosen randomly from schools of 39 centers in 8 cities. The multilevel analysis was made to assess both individual-level and city-level risk factors. The effect of gross domestic product (GDP) was further investigated by analysis of the factors.
Results: Analysis of city-level environmental factors showed that GDP [adjusted odds ratio (OR)=1.88], particulate matter with aerodynamic diameter ≤10 μm (PM10) (adjusted OR=1.37), and average humidity (adjusted OR=1.33) were strong risk factors. Further analysis of the factors decomposed GDP into two major factors, the first represented by urban construction, energy consumption, nitrogen dioxide concentration, and the second represented by health-system coverage. This suggested that the negative effects of GDP outweighed its positive effects on asthma.
Conclusions: The prevalence of childhood asthma varies significantly in the eight Chinese cities. Socio-environmental factors such as GDP, PM10 and average humidity are strong risk factors controlling individual attributes, suggesting that balance is needed between public health and economic development in China.
Key words: asthma; environmental health; risk factors; socioeconomic factors
World J Pediatr 2013;9(2):120-126
Introduction
Asthma is one of the most common chronic diseases which has caused substantial health care burden worldwide. According to World Health Organization (WHO) estimates, 235 million people suffer from asthma and 255 000 people died of asthma in 2005.[1] In China, several surveys during the last two decades showed a fast increase in asthma prevalence.[2] Although these studies indicated wide geographic variations in asthma prevalence and the variations were likely to be caused by geography-associated environmental risk factors, the association between environmental factors and asthma was not adequately analyzed. Hence, there is a significant relationship between the rapid development of China's economy and fast environmental changes.
To study the risk factors of asthma, we concentrated on individual-level variables such as allergen exposure, access to health care resources, socioeconomic status, and life style. Though these factors may partially explain the disparity in asthma outcomes, their changes are incapable of fully explaining the diverse pattern across different regions in China. In some US-based studies, factors in different geographic and/or social levels were put together and evaluated using multilevel analysis.[3-6] However, no studies have been conducted in China, which incorporated factors at multiple levels into one model.
Considering multilevel structures, researchers from the USA and Europe studied neighborhood and/or school level variables, including racial/ethnic make up, socioeconomic status, violence exposure, allergen exposure, in addition to individual-level risk factors and described the asthma variation in those levels.[3-6] However, assessments of environmental factors at the city level were rarely found. Since there were different systems in China's urban and rural areas, and the socio-environmental data were more easily collected in cities, as an initial study, we considered a multilevel model is needed to include both individual attributes and city-level environmental factors to understand the etiology of asthma in China.
Methods
Subjects
This study was a multi-center, cross-sectional study of 6-13 year-old school children chosen randomly from schools of 39 centers in 8 cities: Shanghai, Guangzhou, Xi'an, Wuhan, Harbin, Chengdu, Hohhot and Urumqi. The cities were capital cities of provinces located in four different regions cited in the 2006 China Statistics Yearbook according to their geographic locations, economic standards, and population densities (Table 1).[7] Why we choose capital cities? In China, capital cities generally have higher levels of health care and public awareness of allergic diseases compared to other cities or rural areas within the same province, thus the reporting of allergic conditions and risk factors would be more reliable. Three to ten districts were randomly selected from each city, and 1-2 elementary school(s) from each district of the city. Since the selection was proportional to size, the number of districts was determined by the sizes of the 8 cities, and the number of schools was determined by the sizes of the districts.
Data collection
The survey was conducted by The Key Laboratory of Children's Environmental Health in Shanghai, China. A national steering committee which comprised highly experienced pediatricians and epidemiologists from each survey center was established; a uniform research protocol was used by each survey center; and formal training for the survey interviewers was provided. Each questionnaire was completed by a parent or guardian of a child after an informed consent form was signed. To ensure credibility and accuracy of the survey, we randomly selected 239 children for re-evaluation of responses one month after the initial evaluation. A total 123 questionnaires for testing parallel consistency (the consistency between the results when father and mother completed the same survey at the same time) and 116 questionnaires for assessing test-retest reliability were completed. The internal consistency of overall questionnaire was good (Cronbach's alpha's coefficients were 0.73). The parallel consistency between mother and father was represented by intro-class correlation coefficients (ICCs)=0.89, and the test-retest reliability was also high (ICCs=0.85). There were 23 791 children from six grades in the chosen schools who participated in the study, of whom 22 018 (92.55%) returned completed questionnaires. We excluded 1250 participants who had missing data of either dependent variables or independent variables considered.
Outcome variables
We used key questions from the International Study of Asthma and Allergies in Children questionnaire to examine the prevalence of current asthma. Subjects with asthma were identified as reported by one of the child's parents or guardians. The question regarding asthma was: "Has the child had asthma in the past 12 months?" (Yes or no).[8]
Individual-level variables
All individual-level variables were those have been found to be associated with asthma in prior studies.[9] They belonged to demographic information, health conditions, socioeconomic status, life style, etc., including specific variables such as age, gender, city, diagnosed gastroesophageal reflux (GER), diagnosed obesity, diagnosed recurrent otitis media, snoring, mode of delivery, exclusive breast feeding, intake of carbonated drinks, diagnosed childhood attention deficit hyperactivity disorder (ADHD), diagnosed allergic rhinitis, diagnosed eczema, diagnosed prepartum and postpartum depression, family structure, household income per capita, parental education level, common cold, sleep-disordered breathing, frequent computer use for amusement, and schoolwork burden. The variables related to health conditions were dichotomous based on answers "yes" or "no" to the specific questions. Other variables were as follows: mode of delivery was categorized as vaginal or caesarean, exclusive breast feeding was categorized as "≥4 months" and "<4 months", intake of carbonated drinks was categorized as "≥5 times a week" and "<5 times a week", family structure was categorized as "nuclear family" or "single-parent or extended family", household income per capita was categorized as "≥1500 RMB/year" and "<1500 RMB/year", parental education level was "≥high school graduate" and "<high school graduate", common cold was ">5 times/year" and "<5 times/year", frequent computer use for amusement was "≥5 times/week" and "<5 times/week", and schoolwork burden was "heavy" and "not heavy".
Additional variables
In addition to field-collected data, environment-related variables, which were obtained at the city level, were retrieved later from official database.[7,10] These variables were categorized as natural environmental factors and socioeconomic factors. The natural environmental factors included altitude, humidity variation, temperature variation, average temperature, and average humidity. The socioeconomic factors included gross domestic product (GDP), nitrogen dioxide (NO2), sulphur dioxide (SO2), particulate matter <10µm (PM10), urban construction, standard coal consumption, and health-system coverage. Table 2 shows definitions of these factors. In particular, according to a recent study, health-system coverage was identified as a composite measure of coverage constructed with simple averages of 11 interventions to summarize the overall service delivery at regional level, which included access to safe drinking water, access to sanitary toilets, smoking cessation, antenatal care, hospital delivery, postnatal care, immunization, examination of suspected tuberculosis cases, treatment of confirmed tuberculosis cases, treatment of hypertension, and effective treatment of hypertension.[10]
Statistical analysis
The outcome of asthma was coded as a binary variable (present/absent). Binary logistic regressions with single individual-level independent variable were used and variables were selected (P<0.05). With these selected individual-level variables Xi, our first model was a fixed effect logistic model where "city" was treated as a fixed effect:
log it (πi) = log (πi/1-πi) = β0 + β1Xi+β2 cit yi+ ei
where πi is the probability of having asthma for subject i and ei is the random differential of subject i. Stepwise procedure was used to select risk factors at the individual level and to assess their influence on asthma by calculating odds ratios (ORs). The probabilities for variable entry and removal were 0.05 and 0.10, respectively.
In order to investigate city-level risk factors, our second model was a two-level mixed effect model, where "city" was treated as a random effect, to disclose both the effects of city-level variables and the variation among cities:
log it (πij) = log (πij/1-πij) = β0 + β1Xij+β2 Xj+ uj+ eij
where πij is the probability of having asthma for subject i located at city j, Xij represents individual-level variables and Xj represents city-level variables. The parameter uj represents the random differential at the city level and eij represents the random differential at the individual level. These differentials were each assumed to have an independent and identical distribution across cities and individuals with variances σ2u and σ2e , respectively.
For city-level variables that were correlated, hierarchical cluster analysis was applied to cluster them into less-correlated groups (correlation <0.4). Representative variables were selected from each group and included the multilevel model. Factor analysis was also used to disclose major factors underlying key variables that had impacts on asthma outcome.
We used SAS 9.2 to implement the analysis. The multi-level analysis was performed using PROC MIXED in SAS.
Results
In binary logistic regression analysis of single independent variables, 19 variables were selected (P<0.05) as factors influencing asthma. When a stepwise fixed effect logistic model was used, 17 of the 19 variables were selected (Table 3). Among them, "city" was the most significant factor, followed by maternal education level, gender, ADHD, etc. With Urumqi as the reference category for "city", the prevalence rate of asthma in Shanghai was the highest (adjusted OR=4.48), followed by Chengdu (adjusted OR=3.16), Wuhan (adjusted OR=2.82) and Guangzhou (adjusted OR=2.06) (P<0.05). Harbin, Xi'an and Hohhot were not statistically different from Urumqi in the prevalence rate of asthma (P>0.05) (Table 3).
To identify the effects associated with "city", particularly environment-related effects, 12 city-level environmental variables were chosen according to their relevance and data availability. To study the inter-correlations among these city-level variables, data for 29 Chinese cities were collected and hierarchical clustering was applied to cluster these variables into five groups with average Pearson's product-moment correlation coefficients between groups <0.4. The largest group had five variables: GDP, energy consumption (represented by standard coal), urban construction, NO2 concentration and health-system coverage; all were socio-economic variables (Fig).
Representative variables were selected from each group: altitude, average humidity, GDP, SO2, and PM10. To compare these variables which were of different scales, we used z-scores in the two-level mixed effect model. Among the five city-level variables, GDP was the most significant (OR=1.89), followed by PM10 (OR=1.37) and mean humidity (OR=1.33) (Table 4).
To further investigate the composite effects of GDP, we made factor analysis of the variables that were closely correlated with GDP. Two major factors were extracted (eigenvalues >0.5) (Table 5). Factor I represented by standard coal consumption, urban construction, and NO2 concentration, which were considered as risk factors of asthma; Factor II represented by health-system coverage, which was considered as a protective factor of asthma. Thus, the negative effect of GDP outweighed its positive effect on asthma.
Discussion
The cross-sectional survey reveals a geographic variation in the prevalence of asthma in different regions of China, with the highest prevalence rate of 7.2% in Shanghai and the lowest rate of 1.0% in Hohhot. In this survey, we also examined the associated factors of childhood asthma in China at both individual and city levels.
The individual-level factors can be categorized into biological conditions (mode of delivery, sleep disordered breathing, paternal age at child birth), life-style (diagnosed obesity, carbonated drinks intake), socioeconomic status (maternal education level, household income per capita, family structure), and health conditions (common cold, diagnosed ADHD, diagnosed allergic rhinitis, diagnosed eczema, diagnosed GER, paternal snoring). These findings added to the evidence that factors in these categories contribute to explaining the variation of asthma prevalence in China.
Using the fixed effect model with "city" as an independent variable, we found that "city" was the most significant factor contributing largely to the variation of asthma outcome. This prompted us to explore the sub-factors under "city". Thus, we selected 12 city-level environmental variables, including five natural environmental variables and seven socioeconomic ones. We found that there was a large variation in these variables among the 8 cities (Table 3). We intended to understand how the variation in these variables affects the prevalence of asthma.
Our findings were consistent with the reported data showing that socioeconomic factors, such as GDP, are associated with the prevalence of asthma.[11] In contrast to the report that there was a relatively weak positive association between asthma and socioeconomic factors,[11,12] we found a strong influence of GDP on asthma prevalence; furthermore, the composite effect of GDP may be decomposed into positive and negative effects on asthma. Apparently, the negative effect of GDP outweighed its positive effect, thus GDP was shown to be a risk factor for asthma. Indeed, the increase of GDP leads to the improvement of health-system coverage,[10,13] which enhances asthma control; however, the increase of GDP is also positively correlated with worsening envirmental pollution in China, which makes asthma is harder to control. China's GDP has been increasing rapidly during the past 20 years. Thus, the analysis of GDP-related risk factors is important for policy-making regarding public health and city development, while maintaining a reasonable increase of GDP.[14]
As a major source of air pollution in China,[15] PM10 was also shown to be a significant risk factor in the present study. PM10 is the particulate component of air pollution that can enter the lungs, deposit in the airways and also penetrate into the periphery of the lungs. Several pathways have been proposed to contribute to the asthmatic response and could be amplified by PM10 exposure. These include the ability to cause inflammation with subsequent tissue damage, neurogenic stimulation with increased smooth muscle constriction and airway inflammation, and direct stimulation of lipid mediators and mucus, which contribute to airway narrowing and blockage respectively. China's PM10 has far exceeded the international standard (Global Air Quality Guide) that was formulated by the WHO.[16,17] High PM10 concentrations were considered to be related to the coal-centerd energy structure and large-scale urban construction.[15]
The significant effect of outdoor average humidity represented influence from the climate, since average humidity was closely related to average temperature, temperature variation and humidity variation in our study. Thus, the asthma outcomes were affected by both natural environment and socioeconomic environment. The mechanisms, directly or indirectly, by which outdoor average humidity may affect the manifestation of asthma in children were not clear. Considering outdoor average humidity is a strong determinant of indoor relative humidity, we thought it was likely that outdoor average humidity may affect the prevalence of asthma symptoms through the indoor relative humidity, which was associated with some high risk factors of asthma prevalence, such as indoor exposure to dust mites, home dampness and moulds.[12]
There are several limitations of the present study which may bias the results. Firstly, all the sampling sites were cities, thus only representing urban areas.Further research is expected to be performed outside urban areas. Secondly, since most of information was obtained from the questionnaire, some biases such as memory recall bias were unavoidable. Thirdly, because this study was initially designed to address the issue of childhood sleep, there were some important allergic risk factors that were not included, for example, family history of asthma and allergies, and indoor and outdoor environmental factors such as indoor temperature, indoor humidity, household dust, pollen exposure, etc. Fourthly, it was also difficult to determine the major site of indoor activities of school-age children among residence, school and other places.
In conclusion, our findings revealed that the prevalence of childhood asthma varied geographically in China. By controlling individual attributes, we found that socio-environmental factors such as GDP, PM10 and average humidity were strongly associated with the prevalence rate of asthma, suggesting that a careful balance is needed between public health and economic development in China.
Funding: This study was supported by Ministry of Education of China (NCET program), National Science Foundation of China (81000592, 11001084, 81222012, 912321023), 973 program (2013CB835100) Science and Technology Commission of Shanghai Municipality (10DZ2272200, 09DZ2200900, 10PJ1407500, 10PJ1403500, 10231203903 and 10JC1411200), Shanghai Municipal Education Commission (11ZZ103). Shanghai Municipal Health Bureau (2010004), Morning Star Rewarding Fund (Category B, 2011) of Shanghai Jiaotong University and Xingbairen plan of Shanghai Jiaotong University of Medical College.
Ethical approval: This study was approved by the local institutional review boards of Shanghai Jiaotong University School of Medicine, Sichuan University West China Center of Medical Sciences, Sun Yat Sen University Medical School, Huazhong University of Science and Technology Tongji Medical University, Xi'an Jiaotong University College of Medicine, Harbin Medical University, Inner Mongolia Medical College and Xinjiang Medical University. We obtained written informed consent forms from all the parents/guardians of the children involving in the study.
Competing interest: None declared.
Contributors: This study was planned and implemented by Shen XM, who was the principal investigator. He was responsible for overall conception and design of the study as well as acquisition of funding. Li F contributed to the conception and design of the study, analysis and interpretation of data, critical revisions of the manuscript for important intellectual content, and clinical diagnostic expertise for the study. Li SH, Jiang F, Jin XM, Yan CH, Tian Y and Deng SN implemented the study and made critical revisions of the manuscript for important intellectual content. Li F and Zhou YC contributed equally to this work.
References
1 World Health Organization, 2008. http://www.who.int/respiratory/asthma/en/index.html (accessed May 10, 2010).
2 Chen YZ, National Cooperation Group On Childhood Asthma China. Comparative analysis of the state of asthma prevalence in children from two nation-wide surveys in 1990 and 2000 year. Zhonghua Jie He He Hu Xi Za Zhi 2004;27:112-116. [In Chinese].
3 Wright RJ. Health effects of socially toxic neighborhoods: The violence and urban asthma paradigm. Clin Chest Med 2006;27:412-421.
4 Salmond C, Crampton P, Hales S, Lewis S, Pearce N. Asthma prevalence and deprivation: a small area analysis. J Epidemiol Community Health 1999;53:476-480.
5 Watson JP, Cowen P, Lewis RA. The relationship between asthma admission rates, routes of admission, and socioeconomic deprivation. Eur Respir J 1996;9:2087-2093.
6 Richmond TK, Subramanian SV. School variation in asthma: compositional or contextual? PLoS One 2009;4:e8512.
7 National Bureau of statistic of China. National Bureau of statistic of China: China Statistical Yearbook 2006. Beijing: China Statistics Press, 2006. [In Chinese]
8 Patel SP, Järvelin MR, Little MP. Systematic review of worldwide variations of the prevalence of wheezing symptoms in children. Environ Health 2008;7:57.
9 Li F, Zhou Y, Li S, Jiang F, Jin X, Yan C, et al. Prevalence and risk factors of childhood allergic diseases in eight metropolitan cities in China: a multicenter study. BMC Public Health 2011;11:437.
10 Department of urban social and economic survey. In National Bureau of statistic of China: China city statistical yearbook 2006. Beijing: China Statistics Press, 2006. [In Chinese]
11 Stewart AW, Mitchell EA, Pearce N, Strachan DP, Weiland SK, ISAAC Steering Committee. International Study for Asthma and Allergy in Childhood. The relationship of per capita gross national product to the prevalence of symptoms of asthma and other atopic diseases in children (ISAAC). Int J Epidemiol 2001;30:173-179.
12 Weiland SK, Hüsing A, Strachan DP, Rzehak P, Pearce N, ISAAC Phase One Study Group. Climate and the prevalence of symptoms of asthma, allergic rhinitis, and atopic eczema in children. Occup Environ Med 2004;61:609-615.
13 Liu Y, Rao K, Wu J, Gakidou E. China's health system performance. Lancet 2008;372:1914-1923.
14 Chen C, Chen B, Wang B, Huang C, Zhao J, Dai Y, et al. Low-carbon energy policy and ambient air pollution in Shanghai, China: a health-based economic assessment. Sci Total Environ 2007;373:13-21.
15 Kan HD, Chen BH, Chen CH, Wang BY, Fu QY. Establishment of exposure-response functions of air particulate matter and adverse health outcomes in China and worldwide. Biomed Environ Sci 2005;18:159-163.
16 The Chinese Ministry of Environmental Protection, 2007. Minister Ambient air quality standard (GB3095-1996 Put into effect as of October 1, 1996). http://english.sepa.gov.cn/standards_reports/standards/Air_Environment/quality_standard1/200710/t20071024_111819.html (accessed May 7, 2010).
17 World Health Organization, 2010. Air quality guidance Globe update 2005. http://www.euro.who.int/Document/E90038.pdf (accessed May 7, 2010).
Received March 6, 2012 Accepted after revision May 7, 2012
|
