Ethnic variation in childhood asthma and wheezing illnesses: findings from the Millennium Cohort Study

Transcription

1 Published by Oxford University Press on behalf of the International Epidemiological Association ß The Author 2007; all rights reserved. Advance Access publication 24 May 2007 International Journal of Epidemiology 2007;36: doi: /ije/dym089 RESPIRATORY Ethnic variation in childhood asthma and wheezing illnesses: findings from the Millennium Cohort Study Lidia Panico, Mel Bartley, Michael Marmot, James Y Nazroo, Amanda Sacker and Yvonne J Kelly* Accepted 2 April 2007 Background Methods It is not clear how respiratory morbidity during early childhood varies across ethnic groups in the UK. This article seeks to determine whether asthma and wheeze illnesses during early childhood differ across ethnic groups and what factors explain observed differences. Data from the UK Millennium Cohort Study on children were analyzed from the second sweep of interviews. Parental interviews were conducted when the cohort member was aged approximately 3½ years. Data collected included the occurrence of asthma and wheezing symptoms, biological and socioeconomic factors and markers of cultural tradition. Results At age 3, 12.3% (n ¼ 1902) of children had ever had asthma and 20.0% (n ¼ 3030) had wheezed in the last 12 months. 18.2% of Caribbean children and 5.0% of Bangladeshi children reported ever asthma compared with 11.6% of White children. 25.5% of Caribbean children and 8.7% of Bangladeshi reported recent wheeze compared with 19.4% of White children. After adjustments, the disadvantage in asthma and recent wheeze for Caribbeans was mostly explained by socio-economic factors (adjusted odds ratios (OR) for asthma 1.42, 95% confidence interval (CI) ; recent wheeze 1.18, ). The Bangladeshi advantage lost statistical significance, mostly due to adjustment for markers of cultural tradition (adjusted OR for asthma 0.40, 95% CI ; recent wheeze 0.44, ). Conclusion Our results point to the need to locate child health within the unique context of each ethnic group and to recognize that potential explanations for observed differences do not necessarily hold for all groups. Introduction Ethnic inequalities in health, both in the UK and elsewhere, have been widely documented. 1 3 Most of the work in this area has focused on adult health and illness with less known about ethnic differences in child health and the possible causes of these. Department of Epidemiology & Public Health, University College London, London, UK *Corresponding author. Department of Epidemiology and Public Health, University College London, 1 19 Torrington Place, London, WC1E 6BT, UK. y.kelly@ucl.ac.uk Asthma and wheezing illness are common in childhood with about one in five British children having doctor diagnosed asthma. 4,5 Ethnic differences in childhood asthma and wheezing illness have been previously described. The 1999 and 2004 Health Surveys for England showed that reported wheeze in the last year was more common among Caribbeans compared with the general population. 4,6 The 2004 survey highlighted particularly low rates of wheezing illness for Bangladeshi and African children. 4,6 A recent systematic review of UK studies also found South Asian children to have lower rates of asthma and wheezing illness. 7 Research from the US shows that the prevalence of asthma among children is about 1093

2 1094 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY twice that of White children However, little is known about why these ethnic differences exist. Of the work done to date in this field, most studies have looked at school-aged children, or have grouped children within wide age ranges and little has been done on early childhood. Few studies have used detailed ethnic classifications, and this is problematic because ethnic minority groups in the UK are very diverse in terms of their socio-economic, migration, acculturation status and health behaviours Furthermore, few studies have been able to examine a variety of explanatory factors, and as a result, little is known about the causes of observed differences in the occurrence of childhood asthma and wheezing illness. This article seeks to determine whether the occurrence of asthma and wheezing illness during early childhood differ across ethnic groups. To do this, we will use data collected during the second sweep of the UK Millennium Cohort Study, when cohort members were aged 3 years. These data allow a detailed ethnic classification to be used and the contribution of a wide range of explanatory factors to be assessed. Firstly, we consider the association of potential risk and protective factors with ethnicity and childhood respiratory illnesses. Ethnic inequalities in health have been linked to socio-economic disadvantage. 16,17 Socio-economic characteristics have also been found to be associated with both ethnicity and childhood respiratory illnesses. 10,18 The effect of socioeconomic status is investigated using household income, occupational class, maternal age at entry into motherhood, maternal employment status and lone parenthood. Potential biological factors that might affect the risk of respiratory illnesses include parental smoking, a proxy for childhood infections (family size), whether the household has any furry pets, and a nutritional marker (breastfeeding initiation). We also look at household language, parental migration status and whether the interview was translated from English to assess potential under-reporting of respiratory illnesses. Secondly, we test for ethnic differences in respiratory illnesses and assess the contributions of potential risk and protective factors in explaining observed ethnic differences. The direction of the effect of these factors might vary by ethnicity. For example, as the Indian group is likely to have a more advantaged socio-economic profile, socio-economic variables would be protective rather than risk factors for that group. However, it is broadly expected that socio-economic factors will be risk factors for asthma and wheeze illnesses for ethnic minority groups. Of the biological factors listed above, breastfeeding, having pets and family size are likely to be protective, given their distribution by ethnicity; while the other biological factors are likely to be risk factors. Our model will not be able to assess the importance of structural effects, such as racism, discrimination and ecological effects, which have been linked to ethnic inequalities in health. 16 Methods The Millennium Cohort Study (MCS) The MCS sample was drawn from infants born in the UK during a 12-month period from 2000 to The survey design, recruitment process and fieldwork have been described in detail elsewhere. 19 Briefly, households agreed to participate in the initial survey, an overall response rate of 68%. Households were identified through the Department of Work and Pensions Child Benefit system and selected on the basis of where the family was resident shortly after the time of birth. Uptake of Child Benefit is almost universal (98%). The sample has a probability design and is clustered at the electoral ward level, with disadvantaged residential areas and areas with a high proportion of ethnic minority population being over represented. This article uses data from the second sweep of interviews, carried out through home visits when the cohort member was aged approximately 3½ years. During interviews questions were asked about the occurrence of respiratory symptoms, socio-economic circumstances, markers of cultural tradition and household composition. The main respondent was usually the mother (98%), although information about their partners was also collected in a separate interview with them. When the mother could not understand or speak English, the resident father was asked to be the main respondent. If neither of the resident parents could undertake the interview in English, another household member above the age of 16 was asked to translate; otherwise a translator was used. The overall sample size for sweep 2 was households were lost to follow up between sweep 1 and 2. These households were more likely to be from a disadvantaged occupational class, to be single parents and slightly more likely not to speak English than those retained in sweep 2 (Annex 1). The sample on which this analysis is based includes all singleton infants whose mothers participated in the surveys and for whom ethnicity is known cohort members met these inclusion criteria. Ethnicity The cohort member s ethnicity was given by the main carer, usually the mother, during the first interview, using the 2001 UK Census categories. The groups used for analysis are: White (n ¼ ), Indian (n ¼ 409), Pakistani (n ¼ 710), Bangladeshi (n ¼ 265), Caribbean (n ¼ 342), African (n ¼ 315) and Other Ethnicities (n ¼ 357). To prevent problems with small cell sizes, cohort members of mixed ethnicity were categorized according to the mother s ethnicity or, if the mother s ethnicity was White, the father s non-white ethnicity was used. Initial analysis indicated that these aggregations have not affected the conclusions drawn in this article. Asthma and wheezing illness outcomes Questions were taken from the ISAAC (International Study of Asthma and Allergies in Childhood) core questionnaire for asthma, a widely used and validated instrument to measure childhood asthma and wheezing illnesses. 20 It includes questions on the occurrence of asthma and wheezing, as well as a variety of severity indicators (Annex 2). The outcomes analyzed in this article are ever asthma by age 3 and wheeze in the last year. Explanatory factors Potential explanatory factors were chosen based on their documented relationship with asthma and wheezing illnesses

3 ETHNIC VARIATION IN CHILDHOOD ASTHMA AND WHEEZING ILLNESSES 1095 and ethnicity, as detailed in our conceptual model. Potential risk factors for ethnic minority groups identified in the literature that could be examined using MCS data were: household income (< equivalent to the poverty line, , , , , refusal and don t know ); most advantaged occupational conditions of the household (measured using the 5-category National Statistics Social and Economic Classification, which distinguishes managerial and professional, intermediate, small employer and self employed, supervisory and technical, semi-routine and routine and missing 11, see Annex 3); maternal age at entry into motherhood; whether the main respondent was in work; lone parenthood; whether either parent smokes; and family size. Potential protective factors include whether the child was ever breastfed; the number of siblings in the household; and whether the household has any furry pets. To explore under-reporting of respiratory illnesses, we look at household language (English only, English plus another language, another language only); maternal migration status (first generation arrived after 11 years of age; first generation arrived before 11 years of age; second generation; third generation or more); and whether the interview with the main respondent was translated from English. Statistical methods All analyses were carried out using Stata Logistic regression models investigate the relative importance of explanatory factors in explaining the ethnic differences in ever asthma and wheeze in the last year. We present the results of logistic regression analysis for the non-white groups that were significantly different from the White group for asthma and wheezing illnesses in bivariate analysis. The comparison group in all models is the White group. Odds ratios and 95% confidence intervals (OR and 95% CI) for asthma and wheeze in the last year are presented adjusted for each explanatory factor separately, as well as adjusted for all explanatory factors. All analyses are based on cases with complete data on all variables using appropriate methods which take into account the clustered sample design. Results Ethnic variation in explanatory factors Table 1 shows the distribution of potential explanatory factors by ethnic group. With the exception of the Indian group, non-white groups tended to be more disadvantaged. More than a third of Pakistani, Bangladeshi, Caribbean and African households annual income was below the poverty line ( ), compared with 18.3% of White households. Psycho-social factors presented a more heterogeneous picture. In 41.1% of Bangladeshi and 34.7% of Pakistani households the most occupationally advantaged household member was in routine or semi-routine jobs, compared with about 20% of White and Indian households. Indian, Pakistani and Bangladeshi groups had the lowest proportions of one-parent households; while Caribbeans and Africans were almost three times more likely to be one parent households than White households. Young entry into motherhood (before 19 years of age) was more common in Bangladeshi (32.9%) and Caribbean (27.3%) groups, and lower in the Indian group (5.9%), compared with Whites (17.6%). Variation in asthma and wheeze by explanatory factors Children from disadvantaged socio-economic backgrounds (those from households with low incomes, receiving benefits, from non-professional backgrounds, whose mother was young at first birth, or from one- parent families) fared worst in both outcomes (Table 2). However, occupational class did not interact with ethnicity (P-value ¼ 0.180). Bilingual households and households who did not speak English reported lower asthma and wheeze prevalence than those who spoke only English. Similarly, children whose mothers were first generation migrants reported lower asthma and wheeze levels than those whose mothers were second or subsequent generation migrants. Table 3 shows that, generally, households whose main respondent is a first generation migrant, who do not speak English and where the interview had to be translated from English reported lower prevalence of ever asthma and wheeze in the last year. This pattern was observed across ethnic groups. Ethnic variation in asthma and wheeze At 3 years of age, 12.3% (n ¼ 1902) of all children had ever had asthma and 20.0% (n ¼ 3030) had wheezed in the last year. Of children who had wheezed in the last year, 22.4% had more than four or more attacks in the last year (n ¼ 689), 23.4% had sleep disturbed by wheeze on at least a weekly basis (n ¼ 705) and 11.3% had attacks of speech limiting wheeze (n ¼ 351). Caribbeans were significantly more likely (18.2%, OR ¼ 1.70, 95% CI ) and Bangladeshis were less likely (5.6%, OR ¼ 0.40, ), to have had asthma compared with White children (11.6%). Caribbeans were significantly more likely (25.5%, OR ¼ 1.40, ) and Bangladeshis less likely (8.7% OR ¼ 0.40, ) to have wheezed in the last year compared with 19.3% of White children (Table 4) The ORs for ever asthma and wheeze in the Bangladeshi group did not change but lost statistical significance, when all explanatory factors were taken into account (OR ¼ 0.43, ; OR ¼ 0.48, , respectively, see Table 5). Adjusting for income, receipt of benefits and parental occupation further decreased the odds of ever asthma and wheeze, while adjusting for household language and maternal migration status increased the OR, but these effects were small. The Caribbean disadvantage for asthma and wheeze was attenuated and became statistically non-significant, mostly due to adjustment for economic and social factors (OR ¼ 1.40, ; OR ¼ 1.17, , respectively). Discussion Main findings Caribbean children were about 70% more likely than their White peers to have had asthma by age 3 and 40% more likely

4 1096 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY Table 1 Explanatory factors by cohort member s ethnicity, % cohort member s ethnicity Unweighted sample size White Indian Pakistani Bangladeshi Caribbean African Either parent smokes Yes Number of siblings in household and over Ever breastfed Yes Keep furry pets at home Yes Household income and over Unknown, refused, missing p-value <0.001 Receiving benefits None One or more Missing Mother s employment status At work Maternal age at entry into motherhood 19 years and under years old years old years old and over Missing Lone parenthood Lone parent household Highest NS-SEC5 in household Managerial and professional Intermediate Small and self employers Low supervisory and technical Semi routine and routine (continued)

5 ETHNIC VARIATION IN CHILDHOOD ASTHMA AND WHEEZING ILLNESSES 1097 Table 1 Continued Unweighted sample size White Indian Pakistani Bangladeshi Caribbean African Missing Household language English only English and other Other only Maternal migration status First generation recent arrival First generation older arrival Second generation Third generation or more Table 2 Explanatory factors by ever asthma and wheeze in the last 12 months at 3 years of age Wheeze in Asthma the last year Unweighted sample size % % Either parent smokes Yes No <0.001 Number of siblings in household and over P-value Ever tried to breastfeed Yes No <0.001 Keep furry pets at home Yes No P-value Household income and over Missing, don t know, refused <0.001 Receiving benefits None One or more <0.001 Mother s employment status At work (continued) Table 2 Continued Wheeze in Asthma the last year Unweighted sample size % % Not at work P-value Maternal age at entry into motherhood 19 years and under years old years old years old and over <0.001 Lone parenthood Not a lone parent household Lone parent household <0.001 Highest NS-SEC5 in household Managerial and professional Intermediate Small & self employers Low supervisory and technical Semi routine and routine Missing <0.001 Household language English only English and other Other only P-value Maternal migration status First generation recent arrival First generation older arrival Second generation Third generation or more P-value <0.001

6 1098 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY Table 3 Prevalence of ever asthma and wheeze in the last year by cohort member s ethnicity and cultural indicators, % cohort member s ethnicity Caribbean African White Indian Pakistani Bangladeshi Ever asthma Main respondent s generation First Second Third or more P-value Household language English only English and another language Another language only P-value Whether main interview was translated Not translated Translated P-value Wheezing in the last year Main respondent s generation First Second Third or more P-value Household language English only English and another language Another language only P-value Whether main interview was translated Not translated Translated P-value Table 4 Prevalence and unadjusted odds ratios of main outcomes, by cohort member s ethnicity Asthma Wheeze in the last year Cohort member s ethnicity Unweighted sample size % OR (95% CI) % OR (95% CI) White Indian ( ) ( ) Pakistani ( ) ( ) Bangladeshi ( ) ( ) Caribbean ( ) ( ) African ( ) ( ) Total to have wheezed in the last year. About half of this disadvantage was explained by social and economic factors. In contrast, Bangladeshi children were less likely compared with White children to have had ever asthma or wheeze in the last year. Markers of cultural tradition diminish this difference. Comparison with other studies The Caribbean disadvantage in asthma and wheezing illness is consistent with findings from other studies. 4,6,8 12 African children presented lower asthma and wheezing rates than Caribbean children. The risk of asthma and wheeze for Indian and Pakistani children was not different from that of their White counterparts, suggesting that the South Asian advantage reported by other studies 6,7 might be due to substantially lower reported rates among Bangladeshi children, and should not be attributed to all South Asian groups. Previous studies highlighted the importance of socio-economic disadvantage as a possible explanation for ethnic differences in asthma. 10,18 We have shown that the Caribbean disadvantage in asthma and wheezing illnesses when compared with White children could be mostly attributed to socio-economic factors. Adjusting for socio-economic factors led to further small increases in the Bangladeshi reported advantage in asthma and wheezing illnesses compared with White children, due to the Bangladeshi group s poor socio-economic profile.

7 ETHNIC VARIATION IN CHILDHOOD ASTHMA AND WHEEZING ILLNESSES 1099 Table 5 Crude and adjusted ORs, by cohort member s ethnicity, compared with the White group Asthma Wheeze in the last year Caribbean Bangladeshi Caribbean Bangladeshi Explanatory factors OR 95% CI OR 95% CI OR 95% CI OR 95% CI Unadjusted Adjusting for each explanatory factor: Parental smoking Ever breastfeeding Household pets Number of siblings in household Household income Receipt of benefits Lone parenthood Maternal employment status Highest NS SEC5 in household Maternal age at entry into motherhood Household language Maternal migration status Adjusted for all factors Other studies have mentioned family size 23,24 and country of birth 25,26 as potential explanatory factors. Family size did not appear to have an effect in our models. Markers of cultural tradition, including mother s migration status, explained some of the difference in asthma and wheeze in the last year for the Bangladeshi but not for the Caribbean group. This is not surprising given the latter s more longstanding migration history to the UK. For Bangladeshis, children with a mother born abroad or living in a bilingual or non-english speaking household had lower prevalence of ever asthma or wheeze in the last year. Interpretation Work in both the UK and US suggests that ethnic minority children, including ethnicities with lower asthma prevalence such as South Asians, are more likely to be hospitalized due to asthma-related causes than Whites. 7,11,27,28 The discrepancy between the South Asian lower asthma prevalence and higher asthma hospitalization rates suggests either a differential in asthma severity between White and South Asian children 7 or under-reporting of asthma and wheezing illnesses. The results of some studies suggest that there is under-diagnosis of asthma among children from ethnic minority groups 28,29 although other work suggests that this is not the case. 30 Inequalities in service coverage and how the needs of ethnic minority groups are addressed, 2 as well as lack of knowledge of services, 31 might lead to under-diagnosis among certain ethnic groups. For example, research from the US shows that Latino adolescents from Spanish-speaking households have poorer asthma knowledge. 32 Although qualitative work in the UK reports that wheezing symptoms are perceived, reported and understood similarly by Bangladeshi and White mothers, 33,34 our data suggests that the Bangladeshi group under-reports the occurrence of asthma and wheezing illnesses in children. As shown by Table 3, in all ethnicities, the households that would have the most problems communicating with British health services (those who are new migrants, who do not speak English at home and who required translation for an English-language interview), were less likely to report asthma and wheeze. The results in Table 3 are not always significant, reflecting the distribution of migration, language and the need for translation across ethnic groups. The data for translation is particularly powerful: across all ethnic groups households that required translation for the main interview generally reported half the levels of asthma and wheeze than those who could respond to an interview conducted in English. Comparing the African group with the Bangladeshi group is also telling. Both groups have similarly recent migration histories to the UK, however, a much larger proportion of the African households spoke some English at home and fewer households needed translation for the main interview compared with the Bangladeshi group. In contrast to the Bangladeshi group, the African group does not show signs of under-reporting asthma and wheeze. The English language skills of parents, therefore, emerge as an important mechanism to explain childhood asthma and wheeze underreporting. Strengths and limitations We were able to utilize a detailed ethnic group classification and estimate the likelihood of asthma and wheezing illness in a large population sample of 3-year-old children. Few UK studies have presented separate data for Africans or distinguished between South Asian groups. 6,9,10,29 We were also able to assess the contribution of a range of explanatory factors to observed differences. However, while all adjusted ORs became statistically non-significant, none of the adjusted OR pointestimates were equal or close to A genetic explanation for the remaining observed differences is, however, unlikely. Firstly, we have not included

8 1100 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY measurements of genetic markers in our analysis, and therefore the presence of such effects cannot be demonstrated. Second, speculating on such effects should only be done alongside recognition that the model we have been able to test contains imperfect measurement. There is marked variability in the adequacy of socio-economic and cultural measures across ethnic groups. 2,16 Further, we could not explore features of the areas in which people live that are linked to deprivation, such as availability of health services and transport links, as well as the structural exclusion experienced by some ethnic minority groups, which might play an important role in ethnic differences in health. Conversely, certain mechanisms, such as social capital, social support and community participation, may have a protective effect and explain some of the Bangladeshi advantage. Our results indicate that in order to avoid potentially misleading reports of low asthma and wheezing illness prevalence in some ethnic groups, cultural markers such as migration history and levels of spoken English need to be taken into account when using the ISAAC instrument. KEY MESSAGES References 1 Davey-Smith G, Chaturvedi N, Harding S, Nazroo J, Williams R. Ethnic inequalities in health: a review of UK epidemiological evidence. Crit Public Health 2000;10: Nazroo JY. The health of Britain s ethnic minorities: findings from a National Survey. London: Policy Studies Institute, Nazroo JY. The structuring of ethnic inequalities in health: economic position, racial discrimination, and racism. Am J Public Health 2003;93: Fuller E. Children s health. Health Survey for England 2004: the health of minority ethnic groups. Leeds: The Information Centre, Kaur B, Anderson HR, Austin J et al. Prevalence of asthma symptoms, diagnosis, and treatment in year old children across Great Britain (international study of asthma and allergies in childhood, ISAAC UK). Br Med J 1998;316: Nazroo JY, Becher H, Kelly Y, McMunn A. Children s health. In: Erens B, Primatesta P (eds). Health Survey for England: The Health of Minority Ethnic Groups. London: The Stationary Office, Netuveli G, Hurwitz B, Levy M et al. Ethnic variations in UK asthma frequency, morbidity, and health-service use: a systematic review and meta-analysis. Lancet 2005;365: Conclusion Ethnic groups are diverse both in the prevalence of asthma and wheezing illnesses and in their social, economic and cultural profiles. Our results point to the need to locate child health within the unique social, economic and cultural context of each ethnic group. There is also the need to recognize that potential explanations for observed differences that are valid for one group do not necessarily hold for other groups. Acknowledgements We would like to thank the Millennium Cohort Study families for their time and cooperation, as well as the Millennium Cohort Study team at the Institute of Education. The Millennium Cohort Study is funded by ESRC grants to Professor Heather Joshi (study director). This work is part of the Ethnic Inequalities in Childhood (ETHINC) project funded by the ESRC (grant code RES ). Conflict of interest: None declared. Not much is known about the ethnic distribution of asthma and wheezing illnesses in the UK, especially among very young children. Caribbean children are more likely to report asthma and wheezing illnesses; this is largely accounted for by their more socio-economically disadvantaged background. For Bangladeshi children, asthma and wheezing illnesses appear to be under-reported, and this is accounted for by their recent migration history and low levels of English language use. This has important implications for primary care. Potential explanations for observed differences in child health outcomes that are valid for one ethnic group do not necessarily hold for other ethnic minority groups. 8 Boardman JD, Finch BK, Hummer RA. Race/ethnic differences in respiratory problems among a nationally-representative cohort of young children in the United States. Popul Res Policy Rev 2001;20: McDaniel M, Paxson C, Waldfogel J. Racial disparities in childhood asthma in the United States: evidence from the National Health Interview Survey, 1997 to Pediatrics 2006;117:e Miller JE. The effects of race/ethnicity and income on early childhood asthma prevalence and health care use. Am J Public Health 2000;90: Nelson DA, Johnson CC, Divine GW, Strauchman C, Joseph CLM, Ownby DR. Ethnic differences in the prevalence of asthma in middle class children. Ann Allergy Asthma Immunol 1997;78: Schwartz J, Gold D, Dockery DW, Weiss ST, Speizer FE. Predictors of asthma and persistent wheeze in a national sample of children in the United States. Association with social class, perinatal events, and race. Am Rev Respir Dis 1990;142: ONS. Census 2001: National Report for England and Wales. London: The Stationary Office, Jones T. Britain s ethnic minorities. London: Policy Studies Institute, 1996.

9 ETHNIC VARIATION IN CHILDHOOD ASTHMA AND WHEEZING ILLNESSES Modood T. Ethnic differentials in educational performance. In: Mason D (ed.). Explaining Ethnic Differences: changing patterns of disadvantage in Britain. Bristol: The Policy Press, 2003, pp Nazroo JY. Genetic, cultural or socio-economic vulnerability? Explaining ethnic inequalities in health. Sociol Health Illn 1998;20: Chandola T. Social class differences in mortality using the new UK National Statistics Socio-Economic Classification. Soc Sci Med 2000;50: Rona RJ. Asthma and poverty. Thorax 2000;55: Dex S, Joshi H. Children of the 21st century: from birth to 9 months. Bristol: The Policy Press, ISAAC Steering Committee. ISAAC Phase Three Manual. Auckland: ISAAC, Rose D, Pevaline D (eds). A Researchers Guide to the National Statistics Socio-Economic Classification. London: Sage, STATA Corporation. STATA software release 9.2. College Station, Texas: STATA Corporation, Karmaus W, Botezan C. Does a higher number of siblings protect against the development of allergy and asthma? A review. J Epidemiol Commun H 2002;56: Rona RJ, Duran Tauleria E, Chinn S. Family size, atopic disorders in parents, asthma in children, and ethnicity. J Allergy Clin Immunol 1997;99: Netuveli G, Hurwitz B, Sheikh A. Ethnic variations in incidence of asthma episodes in England & Wales: national study of 502,482 patients in primary care. Resp Res 2005;6: Ormerod LP, Myers P, Prescott RJ. Prevalence of asthma and probable asthma in the Asian population in burn, U.K. Resp Med 1999;93: Gilthorpe MS, Lay-Yee R, Wilson RC, Walters S, Griffiths RK, Bedi R. Variations in hospitalization rates for asthma among black and minority ethnic communities. Resp Med 1998;92: Gottlieb DJ, Beiser AS, O Connor GT. Poverty, race, and medication use are correlates of asthma hospitalization rates. A small area analysis in Boston. Chest 1995;108: Duran-Tauleria E, Rona RJ, Chinn S, Burney P. Influence of ethnic group on asthma treatment in children in : national cross sectional study. Br Med J 1996;313: Akinbami L J, Rhodes JC, Lara M. Racial and ethnic differences in asthma diagnosis among children who wheeze. Pediatrics 2005;115: Partridge MR. In what way may race, ethnicity or culture influence asthma outcomes? Thorax 2000;55: Chan KS, Keeler E, Schonlau M, Rosen M, Mangione-Smith R. How do ethnicity and primary language spoken at home affect management practices and outcomes in children and adolescents with asthma? Arch Pediatr Adolesc Med 2005;159: Cane RS, Ranganathan SC, McKenzie SA. What do parents of wheezy children understand by wheeze? Arch Dis Child 2000;82: Cane RS, McKenzie SA. Parents interpretations of children s respiratory symptoms on video. Arch Dis Child 2001;84: Annex 1 Selected household characteristics by household presence in Millennium Cohort Study sweeps 1 and 2, % Whether household present in sweep 1 and 2 Sweep 1 only Sweeps 1 and 2 Occupational class Managerial and professional Intermediate Small and self employers Low supervisory and technical Semi routine and routine Missing Lone parenthood Lone parent household Household income and over Unknown, refused or missing Maternal age at entry into motherhood 19 years and under years old years old years old and over Missing Household language English only English and other Other only

10 1102 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY Annex 2: ISAAC Core Questionnaire for Wheezing and Asthma 1. Has your child ever had wheezing or whistling in the chest at any time in the past? 2. Has your child had wheezing or whistling in the chest in the last 12 months? 3. How many attacks of wheezing has your child had in the last 12 months? 4. In the last 12 months, how often, on average, has your child s sleep been disturbed due to wheezing? 5. In the last 12 months, has wheezing ever been severe enough to limit your child s speech to only one or two words at a time between breaths? 6. Has your child ever had asthma? 7. In the last 12 months, has your child s chest sounded wheezy during or after exercise? 8. In the last 12 months, has your child had a dry cough at night, apart from a cough associated with a cold or chest infection? Annex 3: National Statistics Socio-Economic Classification (NS-SEC) categorization The NS-SEC operationalizes social class on the basis of employment relations and conditions. Seven criteria are used to classify occupations into social classes: the timing of payment for work (monthly vs weekly, daily or hourly); the presence of regular increments; job security (over or under 1 month); how much autonomy the worker has in deciding when to start and leave work; promotion opportunities; degree of influence over planning of work; level of influence over designing their own work tasks (21). SEC 1 has the most favourable employment conditions with higher levels of job security, autonomy and opportunity; SECs 5 the least favourable conditions, with SECs 2 and 4 in an intermediate position. SEC 3 consists of small employers and self employed persons.