Background The General Medical Services (GMS) scheme, a means tested public health insurance covering approximately 40% of the population, has traditionally provided prescription medicines without charge. In 2010, a 50c/item charge was introduced, which was increased to €1.50/item in 2013. The international literature points to an association between prescription charges and decreased adherence with consequential increases in morbidity and mortality. However, to date, there are no data on the impact of prescription charges on the GMS scheme on adherence. The aim was to assess the impact of two charges on adherence to prescription medicines.

Methods A longitudinal repeated measures design was used to measure individual adherence to medicines before and after the introduction of charges. Pharmacy claims data from the national database (HSE-PCRS) were used for intervention and comparator groups. Two consecutive cohorts, including new users of anti-hypertensive, anti-hyperlipidaemic and oral anti-diabetic medicines, were established to analyse the effect of both charges. Follow up was 12 and 8 months. Segmented regression with generalised estimating equations was used. Sensitivity analyses according to age, sex and essential/non-essential drug status were conducted. All anlayses were carried in out in R version 2.15.2.

Results Immediately after the introduction of the 50c charge, adherence to anti-hyperlipidaemic medicines was reduced by 2.2% (95% CI 1.24–3.12, p < 0.0001) relative to the comparator group. For anti-hypertensives, the decrease was 4.8% (95% CI 4–5.7, p < 0.0001). For anti-diabetic medicines adherence declined by 2.4% (95% CI 1.3–3.5, p < 0.0001). Directly after the €1.50 charge was introduced, adherence to anti-hypertensive medicines decreased by 5.1% (95% CI, 4.2–6.1, p < 0.0001) compared to the comparator. Reductions of 1% in adherence to anti-hyperlipidaemics were non-significant (95% CI 0.053–2.1, p = 0.063) and were non-existent for anti-diabetic medicines. Neither charge was associated with reductions in long-term adherence for any medicine. A greater drop in adherence to non-essential medicines was observed; adherence to proton pump inhibitors dropped by 11.1% (95% CI 10.8–1.4, p < 0.0001) immediately after the introduction of the 50c charge.

Conclusion The introduction of the 50c charge had a greater immediate effect on adherence to anti-hyperlipidaemic and anti-diabetes medicines than the €1.50 charge, suggesting reducing price elasticity of demand. This is the first analysis of the impact of prescription charges on adherence on the GMS. Further changes to cost-sharing policies in Ireland should be informed by national and international evidence.