Aims: Widening access to medical school is a priority in medical selection. If disadvantaged students do not apply, interventions cannot be effective. To date, no studies have examined factors that predict who chooses to apply to medicine and if socio-demographics influence the profile of those who apply to study medicine. Methods: A large database provided by the UK University and Colleges Admissions Service on all 1,225,156 applicants to UK universities over a 3-year period (2002-2004) was analysed. The relationship between demographics, preference to study medicine and academic performance prior to entry (A level score) were explored using logistic and linear regression and path modelling. Results: Those applying to study medicine were more likely to be female, non-white, of higher socio-economic status and from fee-paying schools. Applying to study medicine was associated with increased academic entrance performance over and above socio-demographic factors. Importantly, in those applying to study medicine socio-demographic inequalities in entrance exam performance was either reduced (for ethnicity and SES) or abolished (for sex and schooling). Conclusions: It is argued that early interventions are needed to increase applications for certain groups to help to reduce socio-demographic inequalities in entrance exam performance and hence medical school admissions. [ABSTRACT FROM AUTHOR]