bodily anxiety or restlessness with QTP (OR 0.506, 95%CrI: 0.290, 0.789), decrease in EPS with QTP (OR 0.441, 95%CrI: 0.129, 0.910), increase in weight gain with OLZ (OR 2.139, 95%CrI: 1.764, 2.626), and decrease in weight gain with ZPD (OR 0.466, 95%CrI: 0.317, 0.657). CONCLUSIONS: The results of this systematic review paint a complex picture akin to the nature of schizophrenia and bipolar disorder, reinforcing the importance of treatment choice. To achieve optimal outcomes, physicians need to consider efficacy and tolerability together with the patient’s psychiatric history, preferences and values when making treatment decisions. PMH10 TOLERABILITY OF ONCE-DAILY EXTENDED RELEASE QUETIAPINE COMPAREDTO QUETIAPINE IMMEDIATE RELEASE:A META-ANALYSIS OF RANDOMISED CONTROLLED TRIALS IN SCHIZOPHRENIA Edwards SJ, Korim F AstraZeneca UK Ltd, Luton, UK OBJECTIVES: In 2002, the National Institute for Health and Clinical Excellence (NICE) highlighted extrapyramidal symp- toms (EPS), sexual dysfunction, sedation, and weight gain, as the outcomes considered by patients taking atypical antipsy- chotics to be the most troublesome. This research was designed to compare the tolerability of the new extended release que- tiapine to the existing quetiapine immediate release formulation on these outcomes in addition to orthostatic hypotension, which could be a significant cause of morbidity. METHODS: A meta-analysis of the four regulatory randomised controlled trials (Study 041, 133, 132 and 146) comparing quetiapine extended release with the immediate release formulation using a fixed effect model. Summary effect estimate was calculated as relative risk (RR) with 95% confidence interval (95% CI) where RR < 1 favours extended release and RR > 1 favours immediate release. All comparisons were conducted on a mg-for-mg basis (300 mg, 400 mg 600 mg and 800 mg) for the two formulations. Statistical heterogeneity was tested for using a chi-square test. A sensitivity analysis was conducted using a random effects model. RESULTS: All outcomes were measured consistently in the trials included in the analyses. There were no significant differences between the two formulations of quetiap- ine for any outcomes assessed. Individual results were as follows: EPS RR 1.067 (95%CI: 0.694 to 1.641; p = 0.767); orthostatic hypotension RR 1.089 (95%CI: 0.744 to 1.595; p = 0.661); sedation RR 0.781 (95%CI: 0.569 to 1.073; p = 0.128); weight gain RR 0.784 (95%CI: 0.521 to 1.180; p = 0.244); prolactin RR 0.708 (95%CI: 0.465 to 1.077; p = 0.107). Significant heterogeneity was not detected in any com- parison (all p > 0.42) and the effect of using a random effects model made no difference to the summary effect estimates. CONCLUSIONS: The meta-analysis suggests that the tolerabil- ity profile of extended release quetiapine is consistent with that of the immediate release formulation. MENTAL HEALTH—Cost Studies PMH11 DRUG-RELATED PUBLIC EXPENDITURE IN EUROPE:THE BUDGET IMPACT OF ILLEGAL DRUGS Prieto L European Monitoring Centre for Drugs and Drug Addiction, Lisbon, Portugal OBJECTIVES: By testing a methodology that combines labelled and non-labelled public expenditure, this report aims to produce first estimates of the amounts European governments spend on the illegal drug problem. METHODS: The European Informa- tion Network on Drugs and Drug Addiction (REITOX) was asked to list any budgeted labelled drug-related fund found after reviewing government budgets for the year 2005 in 30 countries. In order to ensure consistency in comparing figures over time and across countries, labelled expenditure was classified according to the International Classification of the Functions of Government (COFOG). Since not all drug-related expenditure is identified as such in budgets, modelling approaches were used to estimate the amount embedded in other programmes and interventions. Non-labelled drug-related expenditure was obtained by using a top-down costing approach to estimate the proportion of expen- diture causally attributable to drug use. Wherever possible, REITOX explored feasible definitions of ‘attributable propor- tions’ for estimating non-labelled drug-related expenditure under two COFOG functions: public order and safety and health. RESULTS: On the whole, countries have a considerable amount of quality information available on this type of costs, although calculating the non-labelled constituent is often an arduous task. Estimates from reporting countries extrapolated to European level arrived at a total cost of drug-related public expenditure in 2005 of €34 billion, which is equivalent to 0.3% of the sum of the GDP of all of the countries. This represents an average expenditure of €60 per European citizen per year. CONCLU- SIONS: Whilst such figures should still be used with caution (the methodology still needs refining and country data is in no way uniform), one observation the report makes is that the disburse- ments identified mainly refer to public expenditure made at central government level. The future inclusion of sub-national government expenditure will certainly increase the amounts of public expenditure estimated. PMH12 BUDGET IMPACT OF GENERIC ANTIPSYCHOTIC SUBSTITUTION—A DATABASE ANALYSIS IN GERMANY Rudolph I, Sindern J Janssen-Cilag GmbH, Neuss, Germany Cost containment in Germany is to a large extent regulated by fixed doctor’s budgets (Richtgrößen) for treatment of patients in statutory health insurance (SHI). With Olanzapine (OLA) and oral Risperidone (RIS) two of the most often used second gen- eration antipsychotics (SGA) lost patent in 2007. Generic substi- tution could lead to significant cost savings. OBJECTIVES: Aim of this database analysis was to estimate the budget impact of prescribing generic OLA and RIS and to evaluate the potential for a more individualized therapy in schizophrenia without exceeding budgets. METHODS: Using the IMS Disease Analyzer based on the electronic medical records of 112 psychiatrist prac- tices in Germany, a retrospective analysis of drug expenditures for schizophrenia was conducted. As reference data from QIV 2007 were used. Budget impact was then estimated based on a scenario assuming generic prices being 20% (RIS) or 30% (OLA) of original and an overall generic patient share of 80%. RESULTS: Data of 93.844 SHI-insured patients were analyzed, 65.028 of those with drug prescriptions in QIV 2007. Mean drug expenditures per patient in QIV 2007 was €195.52.7% of drug expenditures of psychiatrists could be explained by SGA pre- scriptions. A total of 55.1% of their patients treated with SGA received either RIS or OLA. This corresponded to 49.3% of costs for SGA. The estimated budget impact according to the scenario is a cost saving of 25.6% of the drug expenditure for SGA. CONCLUSIONS: Especially in schizophrenia there is a need for therapy optimization according to current treatment guidelines and as also aspired by payers. As the data show, there is a significant cost saving potential from use of generic SGA. These A582 Abstracts