A library-based ecological study to investigate the contribution of ethnicity to the incidence of oral cancer within health authorities in England and Wales

Our aim was to investigate the incidence of oral cancer in three ethnic groups, Indian, Pakistani, and Bangladeshi residents of the UK, to try to establish a link between ethnic background and the incidence of oral cancer. We used data from the 1991 and 2001 censuses to calculate approximate percentages of the number of Indian, Pakistani, and Bangladeshi residents in each Health Authority (HA) region and correlated these with rates of oral cancer from the Office of National Statistics (ONS) cancer atlas for the same period. Data about lung cancer and oesophageal cancer were used as proxies for the confounding variables of smoking and alcohol, respectively, in a multiple regression analysis. There was a strong significant correlation between figures for lung cancer and variations in oral cancer rates among HAs (p < 0.001). Ethnicity contributed weakly and insignificantly to variations in rates of oral cancer among HAs. In summary, we found only a weak correlation between ethnicity and oral cancer in the UK population, unlike previously published studies, while simultaneously showing a stronger relation with the proxy for smoking trends. There are various drawbacks inherent in library-based ecological studies in general, and in this study in particular, that may have been responsible for this.