Hospital admission rates for asthma and respiratory disease in the West Midlands: their relationship to air pollution levels.

BACKGROUND: A study was undertaken to determine the relationship between hospital admissions for asthma and all respiratory conditions in electoral wards in the West Midlands and ambient levels of smoke, sulphur dioxide, and nitrogen dioxide, and to establish whether the relationship is independent of social deprivation and ethnicity, and is different for young children and older individuals.
METHODS: Data on hospital admissions for acute respiratory conditions were obtained by electoral ward from the West Midlands Regional Health Authority Information Department Körner inpatient data including asthma (ICD 493) and all acute respiratory disease (466, 480-486, 490-496) for the period April 1988 to March 1990. The population for each electoral ward, percentage of ward population that was from non-white ethnic groups, and Townsend deprivation score were all calculated from 1991 census information. Data on smoke and sulphur dioxide (SO2) levels were obtained for 24 wards in Birmingham, Coventry, Wolverhampton, Dudley, Stafford, and Burton-on-Trent, and on nitrogen dioxide (NO2) levels from 39 wards in the same local authority areas. All were background urban sites and most participated in the Warren Spring national quality control programme for SO2 and smoke monitoring. Indirect age-sex standardised hospitalisation rates (SHR) for all respiratory conditions and asthma were calculated using the 1991 rates for the West Midlands RHA as the standard. Multivariate regression models were used to assess the relationship between individual pollutants and the SHR. The Townsend score and percentage of the population from non-white ethnic groups were included in all models to adjust for ethnicity and socioeconomic deprivation.
RESULTS: The SHR for asthma varied almost fourfold across the region, and all respiratory SHR showed more than three fold variation. Bivariate regression revealed both Townsend score and percentage of non-white individuals to be associated with SHR for asthma and all respiratory conditions at all ages, but not for children under 5 years. NO2 was associated with hospital admission rates for all ages including children under 5. SO2 and smoke were not associated with hospital admissions. Multivariate analysis including Townsend score and percentage of non-white subjects in the model revealed that NO2 was associated with hospital admission rates for all respiratory conditions only for children under 5. The Townsend score was associated with SHR for all respiratory conditions, and both the Townsend score and percentage of non-white subjects were associated with SHR for asthma in children under 5 in two of three models. The association between SHR for asthma and percentage of non-white subjects was negative.
CONCLUSIONS: Socioeconomic deprivation, as measured by the Townsend score, is a significant predictor of hospital admission rates for respiratory disease in older individuals, and both the percentage of non-white subjects and the Townsend score are significant predictors of hospital admission rates for asthma in children. After correction for socioeconomic deprivation and ethnicity, background urban NO2 levels in the ward of residence are significantly associated with standardised hospital admission rates for all respiratory disease in children under 5. This may represent a causal effect of NO2 on the respiratory health of children, or the effect of confounding factors not corrected by use of the Townsend score.