PMH33 COST-UTILITY ANALYSIS OF DULOXETINEVERSUS VENLAFAXINE IN THE MANAGEMENT OF MAJOR DEPRESSION IN PORTUGAL Pinto CG 1 , Barbosa A 2 , Miguel LS 3 , Feio M 2 1 Technical University of Lisbon, Lisboa, Portugal, 2 Hospital de Santa Maria, Lisbon, Portugal, 3 Research Centre on the Portuguese Economy—CISEP , Lisboa, Portugal OBJECTIVES: The aim of this study is to estimate the cost- effectiveness ratio of Duloxetine compared to Venlafaxine in the management of major depression in Portugal. A model developed by MEDTAP International Inc. was adapted in order to reflect clinical practice in Portugal. METHODS: The model allows for the estimation of costs, QALY, the proportion of patients achiev- ing remission, and the average time spent in remission by each patient during one year. The societal perspective was adopted. In each cycle a patient may drop-out, achieve remission, respond to treatment not achieving remission, or have no response. Clinical evaluation is based on patients’ scores in the Hamilton scale. Patients that do not achieve remission may have their dose adjusted, switch to or add another anti-depressant. Clinical data was taken from a randomized controlled trial between Duloxet- ine and Venlafaxine. Results show that Duloxetine allows more patients to achieve remission, despite the fact that less patients achieve response. Duloxetine is also associated with a higher proportion of patients with adverse events. Recurrence and relapse rates were taken from the literature. Quality of life and resource consumption data were collected through a Delphi panel of psychiatrists with large clinical experience. Official sources were used to get unit costs. RESULTS: Patient that start treatment with Duloxetine spend 34.5 weeks in remission, with 83.9% achieving remission at the end of the period. As a conse- quence, Duloxetine enables 0.708 QALYs per patient. Those who start treatment with Venlafaxine stay in remission during 34 weeks, with 83.5% achieving remission at the end of the period. Therefore, Venlafaxine treated patients benefit from 0.698 QALYs. Duloxetine also allows for savings in resources used. It implies a total expenditure of €1126 per patient while Venlafax- ine related expenditure equates to €1231. CONCLUSIONS: Duloxetine is more effective and less costly, being a dominant alternative. PMH34 WORKPLACE BURDEN OF MILD, MODERATE,AND SEVERE DEPRESSION INTHE UNITED STATES Birnbaum HG 1 , Kessler R 2 , Seal B 3 , Kelley D 1 , Hsieh M 1 , Ben-Hamadi R 1 , Greenberg P 1 1 Analysis Group, Inc, Boston, MA, USA, 2 Harvard Medical School, Boston, MA, USA, 3 Sanofi-Aventis, Bridgewater, NJ, USA OBJECTIVES: While the substantial health care cost of depres- sion in the workplace is well documented, less is known about the impact of depression severity on workplace burden. The objective is to document workplace burden of major depressive disorder (MDD) by severity. METHODS: Using U.S. data from the National Comorbidity Survey-Replication, workforce respondents (n = 4465) were classified into clinical severity cat- egories (not clinically depressed, mild, moderate, severe) using standard scales (CIDI/QIDS-SR). Outcomes included employ- ment status (employed, disabled, unemployed), workplace per- formance (measured as hours worked, self-rated performance, days of missed work), and workplace burden (estimated by mul- tiplying work hours lost by self-reported hourly income, from HPQ). Outcomes were compared across depression severity groups using multivariate models that adjusted for demographic characteristics. Total monthly US burden of reduced work and performance was estimated through extrapolation using govern- ment workforce statistics. RESULTS: Among the 539 depressed respondents, 13.8% were mild, 38.5% moderate and 47.7% severely depressed. Respondents shared similar demographic characteristics across severity levels. Depressed respondents were 2.6 times more likely to be unemployed/disabled than non- depressed respondents (p < 0.001). The prevalence rates of unemployment/disability increased with depression severity: 15.7%, 23.3% and 31.3% for mild, moderate and severely depressed respondents respectively (p < 0.01). Moderately and severely depressed employed respondents were 4–5% less pro- ductive than mildly depressed/non-depressed respondents (p < 0.01). Severity is negatively associated with work performance. Compared to non-depressed respondents, mildly depressed respondents have reduced monthly workplace performance by 3.0 hours (not statistically significant), vs.12.0 hours (14.8 hours) for moderately (severely) depressed respondents (p < 0.001) The monthly cost of lost performance was $188 ($199) per moderately (severely) depressed worker; the total monthly U.S. burden of reduced work and performance is $2.1 billion/ month. CONCLUSIONS: Among MDD respondents in the workforce, there was a positive association between depression severity and rates of unemployment and disability. Depression is also negatively associated with work performance. PMH35 COST ANALYSIS OFTHETRENDS IN HOSPITALISATION OF PATIENTSWITH SCHIZOPHRENIA IN SPAIN FROM 1980–2004 Saz-Parkinson Z, Cediel-García P, Rubio B,Amate JM, Medel-Herrero Á Instituto de Salud Carlos III, Madrid, Spain OBJECTIVES: Cost analysis of the changes in hospitalisation for patients with schizophrenia in Spain over a 25-year period (1980–2004). METHODS: Cost analysis of the hospitalisation for patients with schizophrenia over a 25-year period based on the data exploited from the Spanish National hospital utilization databases: Minimum Basic Data Set and Survey of Hospital Morbidity. Taking into account the changes in the number of admission episodes, average length of stay (LOS) of such epi- sodes, and costs generated per stay, an estimation of the varying hospitalisation costs in this period was carried out. RESULTS: Our data shows that the trend in hospitalisation rates for schizo- phrenics has increased over the entire period (3.70 in 1980 to 5.89 in 2004). LOS has substantially decreased during the study period from 148 days in 1980 to 36 days in 2004. The estimated cost per day in a psychiatric hospital for a schizophrenic patient in Spain in the year 2000 was €215.60. Some factors affecting the cost of inpatient treatment has been the evolution, consumption pattern and pharmaceutical cost of the antipsychotic drugs used in the treatment of these patients. From 1990–2001, antipsy- chotic consumption has almost doubled from 3.31 DID (Daily defined dose, DDD, per 1000 inhabitants and per day of treat- ment) to 6.04 DID. The cost of these drugs has increased 13 times, mainly due to the increase in the cost of atypical antipsy- chotics which has risen from representing less than 1% of the total cost in 1993 to representing 92% in 2001. The DDD cost has risen, in constant euros, from, €6.48 in 1990 to €20.31 in 2001. In addition, as a growing number of patients are being taken care of in the community, as supported by the deinstitu- tionalisation process, admissions into other types of institutions should be considered. There are studies showing an increase in the number of places in residential care and supervised and supported housing, having doubled in Spain from 5.1 per 100.000 population in 1990 to 10.6 in 2006. Of course, the costs per day vary greatly depending on the institution where the A590 Abstracts