£87 k during 2002–2005. Using RCT data rather than GPRD data for event probabilities, the mean cost was £8 k with the VIGOR RCT and £10 k with the CLASS RCT. CONCLUSIONS: The published cost-effectiveness analyses of coxibs lacked exter- nal validity and did not represent patients in actual clinical practice. External validity should be an explicit requirement in cost-effectiveness analyses. PODIUM SESSION I:MENTAL HEALTH I (ATTENTION-DEFICIT/HYPERACTIVITY DISORDER IN EUROPE) MH1 HOW MUCH SHOULDWE BE PREPAREDTO PAY FOR PSYCHOSOCIAL INTERVENTIONS FOR PATIENTSWITH ATTENTION-DEFICIT/HYPERACTIVITY DISORDER (ADHD)? Schlander M 1 , Schwarz O 1 , Hakkaart-van Roijen L 2 , Jensen PS 3 , Persson U 4 , Santosh PJ 5 ,Trott GE 6 , MTA Cooperative Group _ 7 1 Institute for Innovation & Valuation in Health Care (InnoVal-HC), Eschborn, Germany, 2 Erasmus MC, Rotterdam, The Netherlands, 3 Columbia University, NewYork, NY, USA, 4 The Swedish Institute for Health Economics, IHE, Lund, Sweden, 5 Institute of Child Health—Great Ormond Street Hospital, London, UK, 6 University of Wuerzburg, Aschaffenburg, Germany, 7 National Institute of Mental Health, Bethesda, MD, USA OBJECTIVES: Notwithstanding evidence showing its clinical effectiveness, little if any data have supported the cost- effectiveness of psychosocial interventions for patients with ADHD. The NIMH-initiated MTA study was designed to maxi- mize clinical effectiveness of psychosocial interventions in chil- dren with ADHD. We use patient-level data from this study to estimate the maximum allowable cost of better-targeted behav- ioral interventions that would still meet currently used bench- marks for cost-effectiveness in Europe, assuming they replicate clinical effectiveness as reported in the MTA study. METHODS: A total of 579 children age 7–9.9 years with ADHD (DSM-IV) were randomly assigned medication management (MedMgt), intense behavioral treatment (Beh), both combined (Comb), or community care (CC). All MTA treatment strategies were clini- cally effective. Costing from a societal and from a third-party payer’s perspective for Germany, The Netherlands, Sweden, and UK excluded the research component of the study. Treatment response was defined as normalization of core symptoms after 14 months. QALYs were estimated using utility weights derived from UK expert and parent-proxy-ratings. Comb was most effec- tive, and Med dominated Beh economically. Using this data, we estimated the maximum allowable cost (MAC) of Comb versus Med, quantifying the uncertainty by means of non-parametric bootstrapping. RESULTS: MACs and their 95% confidence intervals for Comb versus Med were determined (a) for ADHD, and for subgroups with (b) “pure” ADHD (without comorbidity, n = 184) and (c) hyperkinetic disorder (HKD, with or without conduct disorder, n = 145), assuming (1) Comb meeting an ICER threshold (when added to MedMgt) of (1) €50,000 or (2) €100,000 per QALY. MACs for UK were (1) €2943 (€2569– €3310) and (2) €3328 (€2612–€4043). Estimates for Germany and The The Netherlands were lower, whereas Swedish estimates were broadly in line with UK data. CONCLUSIONS: Despite some limitations, which will be discussed, these estimates may assist designing clinical studies to support acceptable cost- effectiveness of psychosocial treatment strategies for ADHD. MH2 THE SUSTAINED DIAGNOSIS OFATTENTION-DEFICIT HYPERACTIVITY DISORDER FROM CHILDHOOD TO ADULTHOOD:A MEDICAID STUDY Chen H, Sudharshan L University of Houston, Houston, TX, USA OBJECTIVES: To determine the percentage of patients who had a continued diagnosis of attention-deficit hyperactivity disorder (ADHD) from childhood to adulthood in real practice and to examine the association between the sustained diagnosis of ADHD, other mental health comorbidities and the exposure to psychotropic medications. METHODS: A retrospective longitu- dinal analysis was conducted using the 1995 to 2001 Medicaid claims data. The study cohort were patients 4–17 years of age who received a diagnosis of ADHD in the year 1995. These patients were then followed through 2001 to determine the per- centage that retained the diagnosis of ADHD. Multivariate regression analysis was employed to assess the association between sustained ADHD diagnosis, other mental health comorbidities (conduct disorder, oppositional defiant disorder, mood disorders, anxiety disorder and substance abuse disorder) and the exposure to psychotropic medications (stimulants, antidepressants, antipsychotics, antianxiolytics, anticonvul- sants). RESULTS: A total of 18,131 patients were identified with a diagnosis of ADHD in 1995 and had continuous Medicaid eligibility during the 7 year follow-up period. Out of those only 10,746(57.77%) retained a diagnosis of ADHD in the first year follow-up and merely 1530 (8.43%) retained the diagnosis of ADHD in 2001. The length of retaining ADHD diagnosis is positively associated with the early exposure to stimulants and having mental health comorbidities such as mood disorders, anxiety disorder, conduct disorder and oppositional defiant dis- order. CONCLUSIONS: The result is in confirmation with pre- vious prospective studies which indicated that ADHD shows a remission with maturation. However, the proportion of patients retaining ADHD diagnosis in real practice is 60% to 80% lower than the findings of those prospective studies, which suggests a significant gap between knowledge advancement and practice. Further study is needed to clarify the relationship between psy- chopharmacotherpy and sustained diagnosis of ADHD from childhood to adulthood. MH3 COST-EFFECTIVENESS OF LONG-ACTING METHYLPHENIDATE FORTREATMENT OF ATTENTION- DEFICIT/HYPERACTIVITY DISORDER (ADHD) IN CHILDREN AND ADOLESCENTS IN FINLAND:AN EVALUATION BASED UPON A RANDOMIZED CLINICAL TRIAL (RCT) Schlander M 1 , Hjelmgren J 2 1 Institute for Innovation & Valuation in Health Care (InnoVal-HC), Eschborn, Germany, 2 Janssen-Cilag AB, Sollentuna, Sweden OBJECTIVES: To evaluate the cost-effectiveness of methylpheni- date (MPH)-OROS, a long-acting formulation given once daily (o.a.d.), compared to short-acting MPH (immediate-release, IR), which requires twice (b.i.d.) or thrice (t.i.d.) daily administration schedules, from a payer’s perspective in Finland. METHODS: Health care resource utilization was estimated based on the Canadian ADHD RCT by Steele et al. (2006), comparing MPH- OROS (average dose at study end, 37.8 mg/d; n = 70) with usual care using MPH-IR (n = 73; hereof, 61% t.i.d., 34.6 mg/d; 39%, b.i.d., 31.4 mg/d) in an open-label ‘pragmatic’ parallel-group design over eight weeks. For costing, these data were combined with Finnish unit costs. Effectiveness was defined as intent-to- treat remission rates as determined by parent ratings (primary study endpoint: 18-item SNAP-IV scale). For an estimation of Abstracts A339