OP41 The representativeness of the UK Biobank cohort on a range of sociodemographic, physical, lifestyle and health-related characteristics

Background

UK Biobank (UKB) is a population-based cohort study of half a million participants aged 40–69 recruited between 2006 and 2010 in England, Wales and Scotland. In order to investigate the representativeness of UKB, the distribution of a range of sociodemographic, physical, lifestyle and health-related characteristics was compared between UKB participants and (a) UKB invitees (i.e. all those invited to join UKB including those who did not participate) and (b) findings from nationally representative surveys.

Methods

The response rate for age, sex and socioeconomic status was compared between 503,310 UKB participants and 8,761,869 UKB invitees. Other characteristics of the cohort were compared with nationally representative data sources, including the UK Census (for data on ethnicity), the Health Survey for England (for data on body mass index, smoking, alcohol consumption and prevalence of self-reported health outcomes) and the Office for National Statistics (for data on national cancer incidence and mortality rates). For all data sources, summary data was selected that matched, as closely as possible, the UKB cohort with regard to population characteristics (age, sex), geographical coverage and period of data collection (2006–2010).

Methods

The overall response rate to UKB was 5.5%, and was higher in women, older age groups and those from less socioeconomically deprived areas. Compared with the general population, UKB participants were less likely to be obese, to smoke, to drink on a daily basis and have fewer self-reported diseases. The ethnic background of UKB was similar to that of the national population from the 2001 UK Census (both with 95% from a white ethnic background) but less similar compared with the 2011 UK census (91% white ethnic background). Total cancer incidence rates were 31% lower in men and 42% lower in women at ages 65–69 compared with the national population. All-cause mortality in UKB at ages 65–69 was 56.0% lower in men and 67.6% lower in women compared to national death rates.

Conclusion

UKB is not representative of the general population on a variety of sociodemographic, physical, lifestyle and health-related characteristics, with evidence of a ‘healthy volunteer’ selection bias. As a result, UKB is not suitable for deriving generalizable disease prevalence and incidence rates, although its large size and heterogeneity of exposure measures do provide valid scientific inferences of associations between exposures and health outcomes that are generalizable to many other populations.