Influence of ethnic group on asthma treatment in children in 1990-1: national cross sectional study

BMJ 1996; 313 doi: (Published 20 July 1996)
Cite this as: BMJ 1996;313:148
  1. Enric Duran-Tauleria, lecturer in public health medicinea,
  2. Roberto J Rona, reader in public health medicinea,
  3. Susan Chinn, reader in medical statisticsa,
  4. Peter Burney, professor of public health medicine
  1. aDepartment of Public Health Medicine, United Medical and Dental Schools of Guy's and St Thomas's Hospitals, St Thomas's Hospital, London SE1 7EH
  1. Correspondence to: Dr Duran-Tauleria.
  • Accepted 10 May 1996


Objective: To examine the extent to which the prescription of drugs for asthma adhered to recommended guidelines in 1990-1 and to assess the influence of ethnic group on prescription.

Design: Cross sectional.

Setting: Primary schools in England and Scotland in 1990-1.

Subjects: Children aged mainly 5-11 years. The representative samples included 10 628 children. The inner city sample included 7049 children, 4866 (69%) from ethnic minority groups. For the prevalence estimation 14 490 children were included in the analysis (82% of the eligible children). For the treatment analysis a subgroup of 5494 children with respiratory symptoms was selected.

Main outcome measures: Prevalence of respiratory symptoms and drugs commonly prescribed for asthma, method of administration, inappropriate treatment, and odds ratios to assess the effect of ethnic group on rate of prescription and method of administration.

Results: Children with respiratory symptoms in the inner city sample were less likely to be diagnosed as having asthma. Of children with reported asthma attacks, those in inner city areas had a higher risk of not having been prescribed any drug for asthma (odds ratio 1.87 (95% confidence interval 1.26 to 2.77). Overall, 773 (75%) of these children had received a β2 agonist, 259 (25%) had received steroids, 148 (14%) had received sodium cromoglycate, and 194 (19%) had received no drug treatment in the previous year. When prescribed, β2 agonists were inhaled in 534 (69%) of cases, and this percentage was even lower in ethnic minority groups. Children of Afro-Caribbean and Indian subcontinent origin who had asthma were less likely to receive β2 agonists, and those from the Indian subcontinent were less likely to receive anti-inflammatory drugs. Antibiotics were less prescribed and antitussives more prescribed in children from ethnic minority groups than in white children.

Conclusion: In 1990-1 the risk of underdiagnosis and undertreatment of asthma was higher in children from ethnic minority groups. The implementation of indicators and targets to monitor inequalities in the treatment of asthma in ethnic groups could improve equity and effectiveness in the NHS.

Key messages

  • In 1990-1 children with reported asthma attacks were more likely to be prescribed drugs for asthma and to use the appropriate method of administra- tion

  • In 1990-1 children with reported asthma attacks who were from ethnic minority groups were less likely to be prescribed drugs for asthma and to use the appropriate method of administration

  • The implementation of indicators and targets to monitor inequalities in the treatment of asthma in ethnic minority groups would help purchasers and providers to improve equity and effectiveness in the NHS


In 1983 Speight et al reported that underdiagnosis of asthma in children led to undertreatment.1 Other investigators in several countries have also reported continuing underdiagnosis and undertreatment,2 3 4 which increase school absenteeism1 2 and possibly the number of emergency room visits and hospital admissions.5 6 7 8 9

To improve the management of childhood asthma, guidelines for its diagnosis and treatment were published in 19893 and later updated.10 11 The guidelines recommended five steps for treatment. The first step is the least severe disease and only β2 agonists are recommended. In the second step antiinflammatory drugs (sodium cromoglycate) are introduced in addition to β2 agonists. If the second step does not work, the third step is to add inhaled steroids. Finally, steps four and five may require regular oral steroids to treat the most severe disease.

Studies in other countries have found a need to improve adherence to guidelines4 12 13 and also major racial inequalities in the management of asthma.14 15 In the United Kingdom a recent study reviewed prescribing information on asthma treatment in a sample of British general practices,16 but it did not give information on the degree of underdiagnosis and undertreatment in the community, and, so far as we know, no study has assessed the variation between ethnic groups.

The national study of health and growth collected data on respiratory symptoms and treatment in 1990 and 1991. This study of primary school children investigated the prevalence of treatment for respiratory illness, the extent to which treatment for asthma adhered to recommended guidelines, and the influence of ethnic group on asthma treatment.



The national study was an annual survey of children aged 5 to 11 years. Information was obtained from three samples. One was an English and one a Scottish sample, which were both based on stratified random sampling of employment exchange areas with proportionally more children from poorer social groups.17 As the distribution of social class in these samples was similar to that in the general population18 we refer to them as representative in this paper. The third was an inner city sample, which was selected according to characteristics of deprivation and proportion of ethnic groups.19 We analysed data together for the 1990 English sample, the 1990-1 Scottish representative sample, and the 1991 English inner city sample.


All information on respiratory illnesses of the children and their parents, on drugs prescribed to the children, and on family background was obtained from a self administered questionnaire. In the English inner city areas the questionnaire was available in dual languages: English-Urdu, English-Gujarati, or English-Punjabi, as appropriate.

Parents reported whether the child had had attacks of asthma or bronchitis during the past 12 months, whether he or she usually coughed first thing in the morning or coughed at any other time, whether his or her chest ever sounded wheezy or whistling, and, if so, whether these symptoms were present on most days or nights.

Parents also reported any drugs taken by the child for chest trouble in the past 12 months. The parents were asked to give the name or type of drug, how it was taken, the number of times it was taken, and the month it was last taken, as in a previous study.20 Up to four drugs could be reported. We grouped the drugs into seven categories: β2 agonists, other bronchodilators, steroids, other anti-inflammatory drugs such as sodium cromoglycate and nedocromil sodium, antihistamines, antibiotics, and antitussives. The method of administration was also analysed and grouped into four categories: inhalation, syrup, tablet, and unknown.

Variables on family background were as follows: one or two parent family; paternal social class classified in five groups (classes I and II, IIIN, IIIM, IV and V, and unknown); and the ethnic group of the child, classified by the language spoken at home and the fieldworker's subjective assessment of the child's ethnic group, as used since 1983.19 In the inner city sample ethnic group was classified as white, Afro-Caribbean, children of families originating from the Indian subcontinent, and other (a heterogeneous group of children), to which were added two groups, England (1990 representative sample) and Scotland (1990-1 representative sample).


The analysis of treatment was based on the subsample of children who had symptoms reported by their parents; this subsample was divided into four groups. Group I comprised children with any reported respiratory symptoms or conditions except occasional or persistent wheeze or asthma attacks; group II, children with occasional wheeze but not persistent wheeze or asthma; group III, children with persistent wheeze but no asthma attacks; and group IV, children with reported asthma attacks.

For each of the four groups the prevalence of prescribed drugs was estimated and the relation between any of the seven categories of prescribed drugs and ethnic group was also assessed by fitting multiple logistic regression models. Independent variables in the analysis were ethnic group, social class, one parent family, parents' reported atopic illness, child's sex and age, bronchitis reported for the child, and, for group IV, the child's number of asthma attacks in the previous 12 months as a proxy of severity.

The proportion of children taking β2 agonists and steroids by different methods of administration was estimated for each group of children. A multiple logistic regression model was fitted to assess the probability of taking inhaled β2 agonists rather than syrup or tablets among the four groups of children.

The final models were obtained using backward elimination carried out until all the remaining independent variables were significant at the 5% level.


There were 10 628 eligible children from the representative sample (6463 living in England, 4165 living in Scotland) and 7049 from the inner city sample.

The response rate for the questionnaire was 92.3% for the representative sample and 85.3% for the inner city sample. Information on child respiratory illness had low percentages of missing values, between 3% and 10% according to respiratory symptom, in the representatives sample, and between 20% and 23% in the inner city sample. The Afro-Caribbean group had the worst response rate among the ethnic groups (varying from 31% to 33% by respiratory symptom). A total of 14 490 (82%) children had complete data and were included in the analysis.

The prevalence of respiratory symptoms varied between groups (table 1). Respiratory symptoms as a whole and persistent wheeze were more common in children from the inner city belonging to white, Afro-Caribbean, and other groups than they were for the representative sample and groups from the Indian subcontinent (P<0.01). The variation in reported asthma was smaller, suggesting that children with persistent wheeze were less likely to report asthma if they were from the inner city, particularly if they were white or Afro-Caribbean.

View this table:
Table 1

Numbers of children and prevalence of reported symptoms by group

A total of 5494 children with at least one reported respiratory symptom were selected for further analysis. β2 Agonists were prescribed for 75.1% of children with asthma attacks, and a much lower proportion had steroids (25.1%) or other anti-inflammatory drugs (14.4%). Children who had occasional or persistent wheeze only were less likely to receive drugs indicated for asthma and more likely to receive antibiotics and antitussives if they did not report asthma (table 2.

View this table:
Table 2

Numbers (percentages) of children receiving prescribed drugs by group of respiratory symptoms or conditions

Neither β2 agonists nor anti-inflammatory drugs were prescribed for 78.5% of children with occasional wheeze and for 79.7% with persistent wheeze compared with 18.9% of children with reported asthma attacks (table 3. Among children with diagnosed asthma the risk of not receiving any drug for asthma was significantly higher in children in the inner city sample compared with the representative sample (odds ratio 1.87 (95% confidence interval 1.26 to 2.77)). This was due to those included in the other ethnic group (odds ratio 2.40 (1.18 to 4.87)) and those from the Indian subcontinent (odds ratio 3.73 (2.33 to 5.98)).

View this table:
Table 3

Numbers (percentages) of children receiving prescribed drugs for asthma by group of respiratory symptoms or conditions

Antihistamines, antibiotics, or antitussives were prescribed without asthma treatment in 205 (28%) of children with occasional wheeze only, 104 (32%) with persistent wheeze only, and 38 (19%) with asthma attacks. Among the children who received drugs for asthma, 199 (72%) of those who had wheeze but were not diagnosed as having asthma and 463 (56%) of those with a diagnosis of asthma received β2 agonists alone.

There were significant differences in the method of administration of β2 agonists in relation to the group of symptoms (P<0.001). Only 69% of children with asthma attacks were prescribed inhaled β2 agonists and the proportion was lower for wheezy children without asthma attacks. Except for Afro-Caribbean children, all children living in an inner city area were less likely than children in the representative sample to take inhaled β2 agonists (table 4).

View this table:
Table 4

Probability of taking inhaled β2 agonists by group of respiratory symptoms, ethnic group, and age

In those with symptoms other than wheeze or asthma, ethnic origin was associated only with the use of antibiotics and antitussives. Compared with the representative sample, children of families from the Indian subcontinent were significantly less likely to use antibiotics, but all children in the inner city sample were more likely to use antitussives (odds ratio 3.10 (2.05 to 4.69)).

In those with reported asthma attacks, ethnic origin was associated with the prescription of most drugs indicated for the treatment of asthma and with the prescription of antibiotics and antitussives (P<0.001). Use of β2 agonists was independently associated with ethnic origin, number of asthma attacks, social class, bronchitis, and sex. Use of anti-inflammatory drugs was associated with ethnic origin, the number of asthma attacks, and sex. Use of antibiotics was associated only with ethnic origin. Children from the Indian subcontinent and those from other groups were less likely to be prescribed β2 agonists, anti-inflammatory drugs, and antibiotics compared with children in the representative sample (table 5). Afro-Caribbean children were less likely to be prescribed β2 agonists. Afro-Caribbean children and those from the Indian subcontinent were more likely to be prescribed antitussives.

View this table:
Table 5

Associations between ethnic group and type of drug prescribed in children with reported asthma attacks (group IV)


Only around 20% of children with wheeze, whether occasional or persistent, but 80% of children with recognised asthma attacks received drugs indicated for asthma. Of those with recognised asthma attacks, Afro-Caribbeans were less likely to receive β2 agonists and children from families originating in the Indian subcontinent were less likely to receive β2 agonists and anti-inflammatory drugs than children in the representative sample. Afro-Caribbeans and those from the Indian subcontinent received more antitussives and fewer antibiotics than white children in the inner city areas or representative sample.

Our study is based on large samples with high response rates for all groups, with the exception of Afro-Caribbeans and to a less extent the heterogeneous other groups. Although there have been studies relating to social class and asthma,21 this is, to our knowledge, the first British study assessing the effect of ethnic group on treatment.

Differences in reported use of drugs between ethnic groups could be due to differences in reporting between the ethnic groups. Regardless of language spoken at home, children from families originating from Africa, the Caribbean, or the Indian subcontinent received fewer β2 agonists and anti-inflammatory treatments than did white children. They received more antitussives, however, and prescription patterns between ethnic groups varied only in those with asthma attacks and not in those with wheeze without asthma attacks. It seems improbable that parents' memory of treatment is related to their ethnic group according to type of treatment or that ethnic group determines memory of asthma attacks but not wheeze alone.


Children with persistent symptoms are more likely to be asthmatic and to have atopic disease,22 so children from the inner city areas, particularly the Afro-Caribbean and white children, probably have comparatively underdiagnosed asthma compared with children in the representative sample. The possibility that asthma remains comparatively underdiagnosed in the inner cities is important as it may lead to lower treatment rates, as shown in this study and in several others.2 4

We confirm the positive effect of asthma diagnosis on treatment seen in other studies.1 2 4 Children with wheeze were less likely to be prescribed bronchodilators and anti-inflammatory drugs and more likely to be prescribed antibiotics and antitussives. Children with wheeze were less likely to receive inhaled β2 agonists than those with asthma attacks, which suggests that the diagnosis also influences the method of administration. It is of interest that treatment is similar for those with occasional or persistent wheeze, suggesting that severity is less important than the diagnostic label in determining the treatment pattern.

According to the guidelines,10 11 β2 agonists should be taken at any step of severity and inhaled. However, our study shows that nearly 80% of children with persistent wheeze and 19% with asthma attacks did not receive β2 agonists, bronchodilators, or anti-inflammatory drugs. Most of those taking β2 agonists were not taking any anti-inflammatory agent. Although it is not possible to assess the correct need for these drugs, the lack of any difference between those with persistent and those with occasional wheeze and the high percentage receiving antibiotics or antitussives all suggest uncertainty in the diagnosis and inappropriate treatment of asthma.

The fact that children with persistent wheeze were less likely to be prescribed sodium cromoglycate and steroids than those reported as having asthma and that the use of these drugs was similar to that of children with occasional wheeze may indicate undertreatment. The relative use of sodium cromoglycate and steroids and of inhaled and oral steroids in Warner's study was very similar to the relative use of these drugs in our study.16


Ethnic origin was not associated with the prescription of drugs in children with persistent wheeze, but it was associated with most drugs in children with asthma attacks after adjusting for confounding variables. This suggests that there are no variations in treatment between ethnic groups when there is not a diagnosis of asthma. However, since a higher proportion of children with persistent wheeze in ethnic minority groups do not have a diagnosis of asthma, a higher number of children in these groups are likely to be undertreated. The association between prescription of β2 agonists and the number of asthma attacks, one parent family, social class, bronchitis, and sex, but the lack of association between these variables and the method of administration, indicate that severity and social environment influence the level of prescription but not the method of administration.

A study in the United States has found that black asthmatic children received medical care more frequently but that they obtained drugs less frequently than other groups.23 Other studies from the United States have shown that variations in rates of admission to hospital and in mortality between ethnic groups are more related to poverty than race.24 25 Although children from ethnic minority groups are more likely to belong to lower social groups than other children in the United Kingdom, treatment for asthma is unlikely to be determined by accessibility to services. Consultations with general practitioners are higher in black and other ethnic groups,26 but parents' and doctors' behaviour may have an effect on asthma care.27

Our study has shown that underdiagnosis and undertreatment of asthma was a serious problem in the United Kingdom, especially in ethnic minority groups, in 1991. Causes of variation in treatment among ethnic groups should be studied and possible interventions evaluated. Qualitative studies in ethnic minority groups could help to identify reasons for the deficiencies in treatment of asthma in these groups. Ethnic monitoring28 and targets for specific populations to monitor adherence to clinical guidelines have been suggested.16 29 The implementation of indicators and targets to monitor inequalities in the treatment for asthma in ethnic minorities could help to improve equity and effectiveness in the NHS.

We thank all the children, parents, teachers, doctors, administrators, nurses, and clerks in the areas and schools for their participation in the study, and the fieldworkers and the administrators in the department for all their contributions. We also thank Professor A Tattersfield and Ms R Hill for supplying the Nottingham questionnaire on which the questions on treatment were based.


  • Funding Department of Health and the Scottish Home and Health Department.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.