Background
During the 1990'ies the costs for pharmaceuticals increased in Sweden. On October 1 2002 a major reform was launched. The key elements were the introduction of mandatory generic substitution and the introduction of a new authority to decide on which drugs should be covered by the PBS and their prices.
To aim was to investigate if the introduction of generic substitution had an impact on the cost development of drugs covered by the Swedish Pharmaceutical Benefits Scheme.

Method
Monthly data on sales of prescribed drugs encompassed by PBS to each of the 21 county councils and to the country in total was obtained from the National Corporation of Swedish Pharmacies (Apoteket AB) for the period January 2000 to December 2004. Total cost for prescribed outpatient drugs (encompassing regular prescriptions and multidose dispensed drugs) can be separated in co-payment (paid by the patient) and subsidised cost (cost covered by the PBS). Drugs on regular prescriptions account for more than 90 % of the cost for all outpatient prescription drugs sold within the PBS.
Data on costs was expressed as retail prices excluding VAT per 1000 inhabitants and working day in Swedish krona (SEK) (SEK 1= Euro 0.11 December 8 2005). All costs were adjusted for inflation on yearly basis 2004 as base. Linear segmented regression was used to investigate if the introduction of generic substitution was associated with any shifts in slope or intercept.

Results
The total yearly cost for prescribed drugs within the PBS for the whole country increased from SEK 2310 per inhabitant in the year 2000 to SEK 2460 per inhabitant in 2004. The subsidized cost (i.e. the cost covered by PBS) increased somewhat more (SEK 1730 per inhabitant in the year 2000 to SEK 1910 per inhabitant in 2004).
The total cost for drugs within the PBS per county council ranged between SEK 1980 - 2290 per inhabitant in the year 2000 and between SEK 2210 - 2650 per inhabitant in 2004. The subsidised cost ranged between SEK 1440 - 1740 per inhabitant in the year 2000 and SEK 1650 - 2090 per inhabitant in 2004.
Generic substitution was associated with a significant (p<0.05) reduction in the increasing slope of total cost for drugs within the PBS for the whole country and the majority of the county councils. This was also found for the subsidised cost in a majority of the county councils (p<0.05). There was no such association for co-payment. The results remained if only drugs sold on regular prescriptions were included in the analysis.

Conclusion
The introduction of generic substitution was associated with a reduction of the increasing slope for costs of outpatient drugs encompassed within the Pharmaceutical Benefits Scheme in Sweden.