Background A socio-economic gradient in use of health care has been observed in a number of situations. These have been used to assess the performance of systems as well as frame discussion on system design. Examination of patterns at an aggregate level may mask important differences between types of care that could lead to different policy advice.

Aims To identify whether differences in registration and use related to socio-demographic characteristics exist in respect of publicly funded oral health care in Northern Ireland and if so to identify the nature of differences in care.

Methods NHS reimbursement data were linked to census and vital statistics data within the Northern Ireland Longitudinal Study. Data cover 28% of the population in Northern Ireland and in this study cover a period from 2003 to 2008. Data for individuals aged 11 or 12 in April 2003 that include registration status, reimbursement on a per item basis, gender, community background, siblings and also the social class and education of household reference person (HRP) were extracted. A series of multivariate analyses were used to examine the relationship between registration and use of care as a function of socio-demographic characteristics.

Results A clear socio-economic gradient was evident in respect of registration status. Adolescents whose HRP was long term unemployed or never worked were registered for 6 months (from a maximum of 54) less and consumed 8.3% less expenditure than those whose HRP was professional. While those from lower social backgrounds consumed 24.4% less expenditure on orthodontic services, with respect to extractions and conservative treatment, adolescents whose HRP was long term unemployed or never worked consumed 35.6% and 25.8%, respectively, more expenditure than those whose HRP was professional, other variables controlled for.

Conclusions A publicly funded demand led service can produce a pattern of service provision that disproportionately reflects the preferences of the affluent at the expense of the needs of the less affluent. This might be masked by analysis of data at an aggregate level. The pattern of service provision that results may from a societal perspective be neither efficient (cost effective) nor equitable. The disaggregated analysis of registration and utilisation patterns in this study highlights the potential for such distortion where public funds support demand led provision by for profit providers.