Ethnic differences in bowel cancer awareness: findings from a pharmacy-based community survey

Abstract

BackgroundWe used London-based multilingual community pharmacies to better understand ethnic inequalities in levels of bowel cancer awareness. MethodsWe invited 206 community pharmacies for this study through the Middlesex Group of Local Pharmaceutical Committees. Participating pharmacies interviewed customers (aged 60 years or older) using a modified version of the Bowel Cancer Awareness Measure. Awareness was divided into symptom, risk factor, and screening awareness. Variation in risk factor and symptom awareness were assessed using stepwise linear regression, whereas variation in screening awareness was assessed using logistic regression. The fully-adjusted model controlled for main spoken language, age, gender, and pharmacy postcode-based index of multiple deprivation rank. No ethical monitoring was required due to complete anonymity of responders and implied consent by return of questionnaire. Findings40 community pharmacies (19% of those invited) interviewed 913 customers. There was no statistically significant difference in area-level deprivation between participating and non-participating community pharmacies (p=0·50). To minimise confounding, individuals who reported a history of bowel cancer were excluded from the analysis (n=49; 5%) leaving 864 participants (95% of the full sample). Among these 864 participants, the sample comprised 578 (67%) English speakers, 412 (48%) women, 406 (47%) men, 406 (47%)white-British, 228 (26%) Indian, 43 (5%) white-Irish, 39 (5%) black Caribbean, 23 (3%) Pakistani, 21 (2%) black African, 16 (2%) of other white ethnicity, and 29 (3%) of other ethnicity. At the univariate level, ethnicity was associated with symptom awareness (p=0·0021) and screening awareness (p<0·0001), but not risk factor awareness (p=0·41). In a fully-adjusted model, screening awareness was statistically significantly lower among black African (adjusted odds ratio 0·37 [95% CI 0·17–0·80; p=0·012), black Caribbean (0·28 [0·10–0·83]; p=0·022), and participants from other ethnic groups (0·19 [0·07–0·48]; p=0·0014), compared with white-British participants. However, symptom awareness was no longer associated with ethnicity. Screening awareness was also higher in pharmacies situated in more affluent areas (7·83 [2·14–28·65]; p=0·002) even after adjusting for other demographics. There was however no association between levels of area-level deprivation and symptom or risk awareness. InterpretationOverall, administering bowel cancer awareness measure was feasible. Although participation of community pharmacies was low, the number of surveys completed was substantial and there was little evidence to suggest selection bias in terms of deprivation. The association between ethnicity and screening awareness was independent of language, which means that unlike symptom awareness, campaigns trying to reduce inequalities in awareness of screening programmes cannot just rely on providing materials in people's native language. Instead, academics and community outreach workers should work with community pharmacies and their customers to co-create materials to improve awareness of the bowel cancer screening programme. Future Bowel Cancer Awareness Measure surveys would also benefit from larger sample sizes among individual ethnic groups and the least and most deprived. FundingNorgine, St Mark's Bowel Cancer Screening Centre.