Differences in cancer incidence by broad ethnic group in England, 2013–2017

BACKGROUND: Cancer incidence variation between population groups can inform public health and cancer services. Previous studies have shown cancer incidence rates vary by ethnic group in England. Since their publication, the completeness of ethnicity recording in cancer data has improved, and relevant inequalities (e.g. risk factor prevalence and healthcare access) may have changed. METHODS: Age-standardised incidence rates were calculated for Asian, Black, Mixed/Multiple and White ethnic groups in England in 2013-2017, using almost 3 million diagnoses across 31 cancer sites. Rate ratios were calculated with the White ethnic group as reference. Sensitivity analyses used imputed ethnicity for cases with missing data and perturbed population estimates. RESULTS: Incidence rates for most cancer sites and ethnic group and sex combinations were lower in non-White minority ethnic groups compared with the corresponding White group, with particularly low rate ratios (below 0.5) for melanoma skin cancer and some smoking-related cancers (lung, bladder and oesophageal cancers). Exceptions included prostate cancer (2.1 times higher in males of Black ethnicity), myeloma (2.7-3.0 times higher in people of Black ethnicity), several gastrointestinal cancers (1.1-1.9 times higher in people of Black ethnicity, 1.4-2.2 times higher in people of Asian ethnicity), Hodgkin lymphoma (1.1 times higher in males of Asian ethnicity, 1.3 times higher in males of Black ethnicity) and thyroid cancers (1.4 times higher in people of Asian ethnicity, 1.2 times higher in people of Black ethnicity). Sensitivity analyses did not materially alter these results (rate ratios changed by a maximum of 12 percentage points, the direction and significance of results were unchanged in all but two cancer site/sex/ethnic group combinations). CONCLUSIONS: People of non-White minority ethnicity in England generally have lower cancer risk than the White population, though there are a number of notable exceptions. These results should galvanise efforts to better understand the reasons for this variation, and the possible impact on cancer services, patient experiences and outcomes.