Background There is a growing burden of multimorbidity with an ageing population. Ethnic inequalities in multimorbidity exist however it is unknown whether social-economic factors and health behaviours may be driving the variation. Research is also lacking on multimorbidity among separate instead of broad ethnic groups.

Methods Using the cross-sectional nationally-representative Health Survey for England 2011–2018 (N=54,4478 adults aged 16+), multivariable logistic regression was conducted on the odds of having general multimorbidity (defined as two or more longstanding conditions), by ethnicity (British White (reference group), White Irish, Other White, Indian, Pakistani, Bangladeshi, Chinese, African, Caribbean, White Mixed, Other mixed, Other), adjusting for age and sex in initial models, followed by education (Degree or higher/Other/No qualifications), quintiles of area-deprivation, smoking status (never/former/current), body-mass-index (Normal/Overweight/Obese) and survey year. This was repeated for cardiovascular multimorbidity among adults aged 40+ (N=37,148: having two or more of the following: doctor-diagnosed diabetes or hypertension, or previous heart attack/stroke). Multiple cardiometabolic biological risk factors (HbA1c≥6.5%, hypertensive, total cholesterol≥5mmol/L) were also examined. Analyses were conducted in STATA17, and complex survey design and non-response weighting applied.

Results 17.8% of adults had general multimorbidity. This was highest among White Irish and Caribbean adults, and lowest among Chinese and African adults. In fully-adjusted models, compared with the British White majority, Other White (Odds Ratio (OR)=0.67 (95% confidence interval (CI)=(0.58,0.79)), and African adults (OR=0.57 (0.45,0.73)), had lower odds of general multimorbidity. Among adults aged 40+, Pakistani (OR=1.30 (0.99,1.70) p=0.058) and Bangladeshi (OR=1.72 (1.15,2.57)) had increased odds, and African adults had decreased odds (OR=0.68 (0.51,0.91)). Among adults aged <40, Other White (OR=0.46 (0.34,0.63)), Indian (OR=0.40 (0.25,0.64)), Pakistani (OR=0.59 (0.38,0.92)) and African adults (OR=0.32 (0.18,0.58)) had decreased odds. In fully-adjusted models, the risk of cardiovascular multimorbidity was higher among Indian (OR=3.32 (2.57,4.29)) Pakistani (OR=3.48 (2.52,4.80)), Bangladeshi (OR=3.67 (1.99,6.79)), African (OR=1.61 (1.05,2.47)), Caribbean (OR=2.18 (1.59,3.00)) and White Mixed (OR=1.99 (1.14,3.45)) adults compared with British White adults. The risk of multiple cardiometabolic risk factors was higher among Indian (OR=5.33 (2.84,10.00)), Pakistani (OR=3.07 (1.39,6.77)), African (OR=2.48 (1.04,5.91)), and Caribbean (OR=3.14 (1.69,5.85)) adults. Broadly, area-deprivation reduced effect sizes whereas smoking and BMI had differential effects among different ethnic groups.

Conclusion Ethnic inequalities in multimorbidity varies between older and younger adults, distinct ethnic groups, and is independent of social-economic status and some health behaviours. Ethnic minority groups are particularly at risk of cardiovascular multimorbidity, which may be exacerbated by underdiagnosis and/or poorer management of cardiometabolic risk factors.