Determinants of utilisation differences for cancer medicines in Belgium, Scotland and Sweden.

Background: Little comparative evidence is available on utilisation of cancer medicines in different countries and its determinants. The aim of this study was to develop a statistical model to test the correlation between utilisation and possible determinants in selected European countries.

Methods: A sample of 31 medicines for cancer treatment that obtained EU-wide marketing authorisation between 2000 and 2012 was selected. Annual data on medicines' utilisation covering the in- and out-patient public sectors were obtained from national authorities between 2008 and 2013. Possible determinants of utilisation were extracted from HTA reports and complemented by contacts with key informants. A longitudinal mixed effect model was fitted to test possible determinants of medicines utilisation in Belgium, Scotland and Sweden.

Results: In the all-country model, the number of indications reimbursed positively correlated with increased consumption of medicines [one indication 2.6, 95% CI (1.8-3.6); two indications 2.4, 95% CI (1.4-4.3); three indications 4.9, 95% CI (2.2-10.9); all P < 0.01], years since EU-wide marketing authorisation [1.2, 95% CI (1.02-1.4); p < 0.05], price per DDD [0.9, 95% CI (0.998-0.999), P < 0.01], and Prescrire rating [0.5, 95% CI (0.3-0.9), P < 0.05] after adjusting for time and other covariates.

Conclusions: In this study, the most important correlates of increased utilisation in a sample of cancer medicines introduced in the past 15 years were: medicines coverage and time since marketing authorisation. Prices had a negative effect on consumption in Belgium and Sweden. The positive impact of financial MEAs in Scotland suggests that the latter may remove the regressive effect of list prices on consumption.