A214 VALUE IN HEALTH 17 (2014) A1-A295 PMH34 THE BURDEN OF OPIOID ABUSE AMONG COMMERCIALLY-INSURED PATIENTS Rice J.B. 1 , Kirson N.Y. 1 , Shei A. 1 , Enloe C. 1 , Cummings A.K. 1 , Birnbaum H.G. 1 , Ben-Joseph R. 2 1 Analysis Group, Inc., Boston, MA, USA, 2 Purdue Pharma L.P., Stamford, CT, USA OBJECTIVES: Prior research has estimated the burden of opioid abuse in the U.S. However, older data may not reflect recent trends in abuse prevalence and associated costs. This study provides updated estimates of the burden of opioid abuse. METHODS: Patients aged 12-64 diagnosed with opioid abuse/dependence (“abusers”) were selected from the Truven MarketScan medical and pharmacy claims database, 2009-2012. A 12-month follow-up period centered on the index date (i.e., first abuse diagnosis) was used to assess costs and was preceded by a 6-month baseline period. Patients were required to have continuous non-HMO cov- erage throughout the 18-month study period. Potential controls met similar inclu- sion criteria but were not diagnosed with abuse, with their index date based on a random medical claim. Abusers were matched 1:1 to controls based on index year, baseline health care costs, and propensity score to account for baseline differences. Per-patient health care costs of abusers and matched controls were compared to determine the excess annual health care costs of diagnosed abuse. Costs reflect payments from payers to providers in 2012USD. Prevalence of abuse was estimated as the proportion of abusers among those with ≥1 one month of eligibility in a given year. RESULTS: 38,876 abusers and 955,202 controls met the inclusion criteria. 35,828 (92.2%) abusers were successfully matched. Post-matching, baseline characteristics were well-balanced. Abusers had $11,319 higher annual total health care costs than matched controls ($22,132 vs. $10,813) in the follow-up period. Costs were higher for abusers in all cost categories: inpatient, emergency department, outpatient, and prescription drug costs [all p-values <0.001]. Diagnosed abuse prevalence increased from 0.16% in 2009 to 0.27% in 2012. CONCLUSIONS: We find substantial excess health care costs of opioid abuse, consistent with prior research. The rising preva- lence of abuse suggests a growing economic burden but may also reflect increased physician awareness of previously undiagnosed patients. PMH35 HEALTH CARE COSTS OF PATIENTS WITH BINGE EATING DISORDER COMPARED TO PATIENTS WITH EATING DISORDER NOT OTHERWISE SPECIFIED AND NO EATING DISORDER Bellows B.K. 1 , Lafleur J. 1 , Kamauu A. 2 , Pawaskar M. 3 , Supina D. 3 , Babcock T. 3 , DuVall S. 1 1 University of Utah, Salt Lake City, UT, USA, 2 Anolinx, LLC, Salt Lake City, UT, USA, 3 Shire, Wayne, PA, USA OBJECTIVES: Binge eating disorder (BED), the most common eating disorder, lacks a specific ICD-9 diagnosis code. With BED subsumed under “eating disorder not otherwise specified (EDNOS),” which includes other disorders, patient identification using medical claims data is problematic. This study used natural language pro- cessing (NLP) to identify patients with BED and compared their health care costs to patients with EDNOS without BED (EDNOS-only) and to matched-patients without an eating disorder (Non-ED). METHODS: An NLP algorithm identified adults with BED from clinical notes in the Veterans Health Administration electronic health record database from 2005-2011 . EDNOS-only patients were identified using ICD-9 code (307.50) and those with NLP-identified BED were excluded. The first diagno- sis date defined the index date for both groups. Non-ED patients were randomly matched up to 4:1 to patients with BED, matched on age, sex, BMI, depression diag- nosis, and index month. Total health care, inpatient, outpatient, and pharmacy costs were examined and reported in 2011 US$. Generalized linear models were used to compare total one-year health care costs while adjusting for patient char- acteristics, including baseline costs. RESULTS: There were 257 BED, 760 EDNOS- only, and 823 matched Non-ED patients identified. The mean (SD) total unadjusted one-year costs were $33,716 ($38,928) for BED, $37,052 ($40,719) for EDNOS-only, and $19,548 ($35,780) for Non-ED patients. The largest component of costs for each was pharmacy (BED $21,842, EDNOS-only $26,426, and Non-ED $12,062) followed by inpatient (BED $8,483, EDNOS-only $5,892, and Non-ED $4,079) and outpatient costs (BED $3,391, EDNOS-only $4,735, and Non-ED $3,407). When adjusting for patient characteristics, BED patients had one-year total health care costs $5,589 higher than EDNOS-only (p=0.06) and $18,152 higher than matched Non-ED patients (p<0.001). CONCLUSIONS: Patients with BED had comparable costs to EDNOS patients and significantly higher adjusted costs than Non-ED patients. BED patients comprise a population with high clinical burden and health care utilization. PMH36 DIFFERENCES BETWEEN HIGH-COST AND LOW-COST PATIENTS DIAGNOSED WITH OPIOID ABUSE Rice J.B. 1 , Kirson N.Y. 1 , Bodnar K. 1 , Shei A. 1 , Birnbaum H.G. 1 , Holly P. 2 , Ben-Joseph R. 2 1 Analysis Group, Inc., Boston, MA, USA, 2 Purdue Pharma L.P., Stamford, CT, USA OBJECTIVES: Opioid abuse is associated with annual per-patient excess health care costs exceeding $20,000. Some patients have considerably higher costs, however, and little is known about their characteristics. This study examined the charac- teristics of high-cost patients diagnosed with opioid abuse. METHODS: Patients aged 12-64 diagnosed with opioid abuse/dependence (“abuse”) were identified in OptumHealth Reporting and Insights medical and pharmacy claims data, 2006-2012. Patients were required to have continuous non-HMO coverage over a 12-month follow-up period centered on an index date (i.e., date of first abuse diagnosis) and a 6-month baseline period preceding the follow-up period. Patients in the top 20% of total health care costs in the follow-up period were classified as “high-cost,” with the remainder considered “low-cost.” Patient characteristics and health care costs were compared between high-cost and low-cost patients using chi-squared tests and Wilcoxon rank-sum tests. RESULTS: 9,291 patients diagnosed with abuse were identified (1,859 high-cost patients; 7,432 low-cost patients). High-cost patients were older (42.5 vs. 36.1; p<0.001) and more likely to be female (55.9% vs. 42.9%%; p<0.001). They had a higher comorbidity burden at baseline, as reflected in the Charlson Comorbidity Index (0.8 vs. 0.2; p<0.001). High-cost patients also had higher rates of non-opioid substance abuse diagnoses (12.4% vs. 8.9%; p<0.001) and psychotic response to response during the titration period was estimated using the predicted treatment response rates estimated by a previously published matching-adjusted indirect comparison (MAIC); mean change in ADHD-RS-IV total score from base- line to endpoint was the efficacy outcome. The incidence rates of adverse events were based on those observed in the clinical trials included in the MAIC. Analyses were conducted from both a societal and a Canadian Ministry of Health (MoH) perspective over a 1-year time horizon with weekly cycles. Deterministic and probabilistic sensitivity analyses (PSA) were conducted to assess the robustness of the base-case results. RESULTS: Compared with ATX, GXR was a dominant strategy (lower cost and improved efficacy) from a societal perspective, while it was associated with an incremental cost-effectiveness ratio (ICER) of $57,866/ QALY from a Canadian MoH perspective. Results of the PSA indicated that the ICER remains below the $50,000 willingness to pay threshold in 93.4% and 61.3% of the simulations from a societal and a Canadian MoH perspective, respec- tively. CONCLUSIONS: This analysis found that GXR was cost-effective relative to ATX in a majority of scenarios and perspectives in the treatment of children and adolescents with ADHD in Canada. PMH32 COST-EFFECTIVENESS OF GUANFACINE EXTENDED-RELEASE AS AN ADJUNCTIVE THERAPY TO A LONG-ACTING STIMULANT VERSUS LONG-ACTING STIMULANT MONOTHERAPY FOR THE TREATMENT OF ATTENTION-DEFICIT/ HYPERACTIVITY DISORDER IN CANADA Lachaine J. 1 , Sikirica V. 2 , Mathurin K. 1 1 University of Montreal, Montreal, QC, Canada, 2 Shire Development, LLC, Wayne, PA, USA OBJECTIVES: Attention-deficit/hyperactivity disorder (ADHD) is a common childhood disorder with a global prevalence of 2.2 to 17.8%. In the case of an inadequate response to stimulant medication, a combination therapy of stimu- lants and adjunctive medication may improve the control of ADHD symptoms, reduce dose-limiting adverse events, and help manage comorbidities. The objec- tive was to assess the economic impact of guanfacine extended-release (GXR) in combination with long-acting stimulants compared to long-acting stimulant monotherapy in the treatment of children and adolescents with ADHD from a Canadian perspective. METHODS: A Markov model was developed to assess the cost-effectiveness of GXR in combination with a long-acting stimulant compared to long-acting stimulant monotherapy. Health states were defined based on the clinician-reported Clinical Global Impressions-Severity (normal, mild, moder- ate, and severe). Transition probabilities were calculated based on patient-level data from a published study. Long-acting stimulants available in Canada were considered in the base-case model: amphetamine mixed salts, methylphenidate HCl formulations, and lisdexamfetamine dimesylate. Analyses were conducted from both a Canadian Ministry of Health (MoH) and a societal perspective over a 1-year time horizon with weekly cycles. Deterministic and probabilistic sensi- tivity analyses (PSA) were conducted to assess the robustness of the base-case results. RESULTS: Compared to long-acting stimulant monotherapy, GXR with a long-acting stimulant was associated with incremental cost-effectiveness ratios of $23,720/QALY and $11,845/QALY from a Canadian MoH and a societal perspec- tive, respectively. PSA of GXR as an adjunctive therapy to long-acting stimulants showed that it remains a cost-effective strategy in 100% of the simulations from both perspectives in numerous sensitivity analyses, according to a willingness to pay of $50,000/QALY. CONCLUSIONS: This economic evaluation demonstrates that GXR as an adjunctive therapy to a long-acting stimulant is a cost-effective strategy compared to long-acting stimulant monotherapy in the treatment of children and adolescents with ADHD in Canada. PMH33 COMPARISON OF RE-HOSPITALIZATION, EMERGENCY ROOM VISITS, AND HOSPITAL COSTS AMONG PATIENTS WITH SCHIZOPHRENIA RECEIVING PALIPERIDONE PALMITATE OR ORAL ANTIPSYCHOTICS IN AN INPATIENT SETTING Lafeuille M.H. 1 , Grittner A.M. 1 , Fortier J. 1 , Muser E. 2 , Fastenau J. 2 , Duh M.S. 3 , Lefebvre P. 1 1 Groupe d’analyse, Ltée, Montréal, QC, Canada, 2 Janssen Scientific Affairs, LLC, Titusville, NJ, USA, 3 Analysis Group, Inc., Boston, MA, USA OBJECTIVES: This real-world retrospective study aims at comparing re-hospitali- zation patterns and costs among patients with schizophrenia receiving paliperi- done palmitate (PP) or oral antipsychotics (AP) in an inpatient setting. METHODS: Hospital discharge and billing records from the Premier Perspective Comparative Hospital Database were analyzed for adult patients who had a schizophrenia-related hospitalization with either PP or oral AP treatment (index hospitalization) between 1/2009 and 3/2012 and no evidence of prior treatment with other long-acting AP. Inverse probability of treatment weights (IPTW) based on propensity scores were used to weight cohorts in order to reduce confounding. Patients treated with PP during their index hospitalization were compared to those treated with oral AP in terms of re-hospitalizations and ER visits using the Andersen-Gill Cox extension of multivariate Cox proportional hazard models. Hospital costs (re-hospitaliza- tions, ER, and hospital outpatient visits) of patients treated with PP were com- pared to those of patients treated with oral AP using a multivariate generalized linear model regression that was weighted by the combined weight of the IPTW and the observation period length. RESULTS: After applying IPTW, weighted mean age was 43.4 years for PP (N=374; N weighted=19,526) and 45.6 years for oral AP (N=45,251; N weighted=26,099) patients. The risk of all-cause re-hospitalizations or ER visits was significantly lower for the PP cohort compared to the oral AP cohort (hazard ratio [HR], [95%CI] = 0.61, [0.59;0.63] p<.0001). Similarly, hospital costs six months after index hospitalization were lower for the PP cohort compared to the oral AP cohort (adjusted mean monthly cost difference [95%CI]: - $404[-781;-148], p<.0001). CONCLUSIONS: This hospital database analysis using the IPTW method found that, compared with oral AP, PP was associated with a lower risk of an ER visit and re-hospitalization in patients with schizophrenia receiving antipsychotics in the hospital setting, resulting in lower hospital costs.