Ecological analysis of ethnic differences in relation between tuberculosis and poverty

OBJECTIVE: To examine the effect of ethnicity on the relation between tuberculosis and deprivation. DESIGN: Retrospective ecological study comparing incidence of tuberculosis in white and south Asian residents of the 39 electoral wards in Birmingham with ethnic specific indices of deprivation. SETTING: Birmingham, 1989-93. SUBJECTS: 1516 notified cases of tuberculosis. MAIN OUTCOME MEASURES: Rates of tuberculosis and measures of deprivation. RESULTS: Univariate analysis showed significant associations of tuberculosis rates for the whole population with several indices of deprivation (P<0.01) and with the proportion of the population of south Asian origin (P<0.01). All deprivation covariates were positively associated with each other but on multiple regression, higher level of overcrowding was independently associated with tuberculosis rates. For the white population, overcrowding was associated with tuberculosis rates independently of other variables (P=0.0036). No relation with deprivation was found for the south Asian population in either single or multivariable analyses. CONCLUSIONS: Poverty is significantly associated with tuberculosis in the white population, but no such relation exists for those of Asian ethnicity. These findings suggest that causal factors, and therefore potential interventions, will also differ by ethnic group.