Changes in cardiovascular risk factors in relation to increasing ethnic inequalities in cardiovascular mortality: comparison of cross-sectional data in the Health Surveys for England 1999 and 2004

OBJECTIVES Reducing disease inequalities requires risk factors to decline quickest in the most disadvantaged populations. Our objective was to assess whether this happened across the UK's ethnic groups. DESIGN Secondary analysis of repeated but independent cross-sectional studies focusing on Health Surveys for England 1999 and 2004. SETTING Community-based population level surveys in England. PARTICIPANTS Seven populations from the major ethnic groups in England (2004 sample sizes): predominantly White general (6704), Irish (1153), Chinese (723), Indian (1184), Pakistani (941), Bangladeshi (899) and Black Caribbean (1067) populations. The numbers were smaller for specific variables, especially blood tests. OUTCOME MEASURES Data on 10 established cardiovascular risk factors were extracted from published reports. Differences between 1999 and 2004 were defined a priori as occurring when the 95% CI excluded 0 (for prevalence differences), or 1 (for risk ratios) or when there was a 5% or more change (independent of CIs). RESULTS Generally, there were reductions in smoking and blood pressure and increases in the waist-hip ratio, body mass index and diabetes. Changes between 1999 and 2004 indicated inconsistent progress and increasing inequalities. For example, total cholesterol increased in Pakistani (0.3 mmol/L) and Bangladeshi men (0.3 mmol/L), and in Pakistani (0.3 mmol/L), Bangladeshi (0.4 mmol/L) and Black Caribbean women (0.3 mmol/L). Increases in absolute risk factor levels were common, for example, in Pakistani (five risk factors), Bangladeshi (four factors) and general population women (four factors). For men, Black Caribbeans had the most (five factor) increases. The changes relative to the general population were also adverse for three risk factors in Pakistani and Black Caribbean men, four in Bangladeshi women and three in Pakistani women. CONCLUSIONS Changes in populations with the most cardiovascular disease and diabetes did not decline the quickest. Cardiovascular screening programmes need more targeting.