Background Pharmacy claims data have been demonstrated to be specific and sensitive in identifying patients with diabetes. This study generates new up to date estimates of the prevalence and incidence of type 1 (T1DM) and type 2 diabetes mellitus (T2DM) in Ireland using the national pharmacy claims database. In the absence of a national diabetes register, this augments existing estimates based on survey data or modelling.

Methods The Primary Care Reimbursement Services database houses data on all prescriptions filled in the Republic of Ireland on the main national health insurance programme, referred to as the General Medical Services (GMS) scheme. This scheme covers approximately 40% of the population on a means tested basis, with almost 100% of the over 70s represented due to different income thresholds. We used data from July 2011–December 2012. T1DM was classified as any individual under 45 years who used an insulin product (ATC A10A) and who had never used an oral/injectable hypo-glycaemic agent. T2DM was classified as any user of an oral/injectable anti-diabetic agent (ATC A10B), irrespective of insulin use and with no age restriction. Incidence of T1DM and T2DM was calculated by estimating the number of new users of A10A or A10B respectively, precluding prior use in a 6 month look back period. Prevalence was calculated using number of prevalent users divided by the age specific GMS eligible population.

Results We used data from 1,799,285 individuals to calculate an incidence rate of 0.14% (95% CI 0.13–0.14) and a prevalence rate of 0.42% (95% CI 0.41–0.43) for T1DM. In the general GMS population, the incidence rate of T2DM was 0.90% (95% CI 0.88–0.91), while prevalence was estimated at 4.38% (95% CI 4.35–4.41). When stratified across 8 age groups ranging from <15 years to 70+ years, prevalence rates rose from 4.91% at 45–54 years to 13.2% in the 65–69 age-group. A similar pattern for incidence rates was observed.

Conclusion Our estimates are similar to prior estimates but provide a more up to date reference and are derived from a larger sample population, including almost all the over 70s population in Ireland. A limitation is that undiagnosed diabetes cannot be estimated. Further analyses will include sensitivity analyses for the length of look back period in estimating new users and age used to define TIDM. Indirect age standardisation will be used to calculate prevalence and incidence in the entire Irish population.