The effect of user charges on the utilisation of prescription medicines in the Italian National Health Service

In Italy, public spending on outpatient prescription medicines has traditionally accounted for a large percentage of health-care expenditures (15% in 1992). Since its inception in 1978, therefore, the Italian National Health Service (INHS) has used user charges and several other regulatory measures to curb expenditures. Since 1978, moreover, these measures have been revised several times. Their impact on utilisation, however, has been the subject of remarkably little work. This dissertation begins to fill this void with specific respect to cardiovasculars, which account for 23% of total INHS drug expenditures. The data set for the empirical analysis was provided by the Emilia-Romagna Regional Health Authority. The unit of observation is total consumption of a given pharmaceutical product by District by month. The data set covers 18 Districts over 60 months, between January 1989 and December 1993. The econometric specification is based on Deaton and Muellbauer's Almost Ideal Demand System. The model has been estimated both in levels and in first differences. The main result is that price does affect utilisation. For cardiovasculars as a whole, the estimated price elasticity is $-$0.30 to $-$0.36. This is a rather large estimate, especially for cardiovasculars. The presence of large own-price elasticities has a number of important policy implications. First, drug copayments cannot be viewed solely as a way of raising additional government revenues. In addition, price deregulation may lead to smaller price increases than expected. Finally, the introduction of reference pricing may prove an extremely powerful price-containment mechanism. Naturally, the welfare implications of these estimates cannot be fully assessed without some information on health outcomes. Other significant results may be summarised as follows. First, changes in copayment exemption regimes do affect utilisation but may have a greater impact in the short than in the medium run. Second, own-price elasticities vary significantly across pharmaceutical groups, with smaller elasticities for necessary drugs of proven efficacy. Third, income has a significantly positive effect on utilisation, with an estimated elasticity of 0.30. Finally, utilisation is also affected by physician characteristics such as seniority and density