A514 VA L U E I N H E A LT H 1 9 ( 2 0 1 6 ) A 3 4 7 – A 7 6 6 gender, BMI, GERD diagnose, GERD severity, associated comorbidities and risk factors were recorded. GSFQ, SF-6D, EQ-5D, and the HRQoL Visual Analogue Scale (VAS) were answered by patients. Several mapping methods were estimated, regression using dummy variables, logistic, Probit and Tobit regressions, and regression using optimal factor scores. Significant predictors were selected and contiguous item levels were merged when needed. Overall Mean Absolute Error (MAE), overall Mean Absolute Percentage Error (MAPE) were used as goodness-of-fit (GOF) indexes to compare models, along with a picewise version. RESULTS: A total of 3,405 patients were recruited by 490 clinicians. Mean age was 49 (±14.4) years and 48.4% were women. Reported comorbidities were clustered in 6 antecedents and 15 concomitant pathologies. Aggregation of levels for the frequency of symptoms items was found more suitable for estimation. Regression weights were found to follow a monotonous progressive pattern. Overall MAE ranged from 0.065 to 0.093, while MAPE values ranged from 13.2% to 38.5%. As expected, picewise GOF demonstrated a better fit in the mid-range of utility scores. CONCLUSIONS: It is possible to translate specific GSFQ scores assessing GERD condition into generic SF-6D utility values. Although regression using dummy variables is a suitable mapping procedure, other alternative mapping methods convey better fit. SF-6D was found more sensitive to different health conditions. PGI27 THE BURDEN OF INFLAMMATORY BOWEL DISEASE(IBD) IN JAPAN Yamabe K1, Kuwabara H1, Umareddy I2, Flores NM3 Pharmaceutical Company Ltd., Tokyo, Japan, 2Kantar Health, Singapore, Singapore, 3Kantar Health, Foster City, CA, USA 1Takeda OBJECTIVES: Ulcerative colitis and Crohn’s disease are chronic, relapsing inflammatory bowel diseases (IBDs). IBD has become more common in Asia including Japan experiencing a more than doubled increase in IBD prevalence over the past decade. The objective of this study was to quantify the burden of IBD in Japan with respect to health-related quality of life, work productivity loss, and healthcare resource utilization. METHODS: Data from the Japan 2012-2014 National Health and Wellness Survey (NHWS) were used (N= 87,903).Respondents who self-reported a diagnosis of IBD were compared with respondents who did not with respect to health related quality of life (the mental component summary [MCS], physical component summary [PCS], and health state utility scores from the Short Form-36v2), work productivity loss and activity related impairment (the WPAI-GH instrument), and self-reported healthcare resource utilization in the past 6 months. Differences on covariates and sample sizes were balanced through binomial propensity score weighting, after which residual significant differences were adjusted using generalized linear models (GLMs). RESULTS: A total of N= 459 respondents reported a diagnosis of IBD (62.50% were male; age= 48.19). Compared with controls, respondents diagnosed with IBD had significantly lower MCS (45.20 vs. 47.26), PCS (50.22 vs. 52.81), and health state utility (0.72 vs. 0.75), also had significantly greater absenteeism (8.57 vs. 3.41), presenteeism (28.41 vs. 21.50), overall work impairment (32.27 vs. 23.26), and activity impairment (28.41 vs. 23.39). CONCLUSIONS: The results suggest that the presence of IBD in Japan is associated with poorer health-related quality of life, greater work productivity loss leading to higher indirect costs, and greater healthcare resource utilization leading to higher direct costs. This study suggests that improved management of IBD could significantly improve patient quality of life, reduce costs associated with work impairment and health care utilization, and therefore be beneficial to both the patient and society. PGI28 JOINTLY MODELING SURVIVAL AND EQ-5D DATA TO ESTIMATE HEALTH UTILITIES FOR HER2+ METASTATIC GASTRIC CANCER Paracha N1, Thuresson P2, Ray J3 Roche, Basel, Switzerland, 2F. Hoffmann-La Roche Ltd., Basel, Switzerland, 3F. Hoffmann-La Roche, Basel, Switzerland 1F. Hoffman-La OBJECTIVES: In economic evaluations, health-related quality of life (HRQoL) is commonly expressed as health state utility values to derive quality adjusted life years. The objective of this study was to estimate health utilities using joint model which simultaneously models patient’s longitudinal EQ-5D trajectory and survival by accounting for the association between the two outcomes. METHODS: We analysed data from a large (n= 389), repeated measure (2,450 observations), EQ-5D-3L dataset from the GATSBY trial to estimate utility values. The EQ-5D was converted into utility value using the UK tariff. Two joint models using current value association structure were fitted to predict utility values according to baseline patient characteristics and key clinical outcomes. The set of variables considered for the joint models included: baseline age, ethnicity, treatment, sex, visit (time), hospitalisation due to adverse event (AE), disease progression and time leading to death. Final set of variables were selected using likelihood ratio test (p< 0.05). RESULTS: The median follow up visit was 5 months (range: 0-45 months). Time was included as a random effect. Included variables demonstrated evidence of an important association with HRQoL outcomes based on magnitude and significance of effect (p< 0.05). Both models showed a highly negative statistically significant association indicating that a lower value of utility increases the risk of death. Model 1 showed disease progression was associated with a statistically significant utility decrement of 0.073. Model 2 showed that in addition to disease progression, patients’ proximity to death showed statistically significant utility decrement of 0.37 in the 3 months leading to patient’s death. CONCLUSIONS: An event based analysis using a joint model better explained variation in EQ-5D data according to key clinical outcomes and patient characteristics. Our analysis suggests that for patients with HER2+ mGC, time leading to death is a considerable determinant for HRQoL and should be accounted for within economic evaluations. PGI29 THE IMPACT OF DIFFERENT GASTROINTESTINAL SYMPTOMS ON THE WELL-BEING OF AFFECTED PATIENTS - UTILITY WEIGHTS FOR THE GASTROINTESTINAL SYMPTOM SCORE (GIS) Mühlbacher AC, Kaczynski A Hochschule Neubrandenburg, Neubrandenburg, Germany OBJECTIVES: The gastrointestinal symptom score (GIS) is used in a standardised form to ascertain dyspeptic symptoms in patients with functional dyspepsia. Each symptom is included with equal importance in the overall evaluation. The objective was to derive utility weights for the ten gastrointestinal symptoms by using a best-worst scaling (BWS). The common assumption that each symptom has the same meaning for affected patients should be tested and analyzed with a stated preference approach. METHODS: In the course of a patient preference study a BWS object scaling was applied. Therefore the symptoms of the GIS were included in a questionnaire using a balanced-incomplete-block-design (BIBD). The BIBD generated a total of 15 choice sets, which each included four attributes. RESULTS: N= 1096 affected patients were asked for their preferences regarding a treatment of functional dyspepsia and motility disorder. The results demonstrated that respondents were capable of trading between the 10 symptoms. Respondents were most sensitive to (in declining importance) the symptoms abdominal cramps (w:-1,27), vomiting (w:-1,07) and epigastric pain (w:-0,76) These symptoms were most important and thus have the greatest influence on the well-being of patients with functional dyspepsia and motility disorders. Whereas the symptoms causing the least impact are the feeling of fullness (w:0,80), early satiety (w:1,54) and loss of appetite (w:1,95). CONCLUSIONS: By using a BWS the patients’ preferences for prioritizing and weighting different gastrointestinal symptoms could be measured. The evaluation of the symptoms has shown, however, that the restrictive assumption of GIS reflects the reality of dyspeptic patients to a limited degree. The calculated utility weights of the patient-relevant symptoms provide an insight on the importance of individual symptoms as part of a treatment decision. The findings provide first signs of a potential modification of the established GIS. One area of potential future work is focusing on exploration of the sensitivity of the GIS. PGI30 A SYSTEMATIC REVIEW OF THE SF-36® HEALTH SURVEY FOR MEASURING HEALTH-RELATED QUALITY OF LIFE IN PATIENTS WITH ULCERATIVE COLITIS Yarlas A1, Bayliss M1, Cappelleri JC2, Maher SM1, Bushmakin AG2, Chen LA3, Manuchehri A4, Healey P2 1Optum, Lincoln, RI, USA, 2Pfizer Inc, Groton, CT, USA, 3New York University, New York, NY, USA, 4Pfizer Ltd, Tadworth, UK OBJECTIVES: Ulcerative colitis (UC), an inflammatory bowel disease, substantially impacts patients’ health-related quality of life (HRQL). The SF-36® Health Survey (SF-36), a generic health status measure, quantifies the impact of UC on HRQL, though to-date there is no summary of evidence of its psychometric properties for this purpose. A systematic review of the literature evaluated the SF-36 as a valid and responsive measure of HRQL in UC. METHODS: We searched PubMed, EMBASE, Cochrane CENTRAL, and BIOSIS Weekly databases. Search terms included ‘ulcerative colitis’, ‘inflammatory bowel disease’, SF-36, and HRQL. English-language articles reporting SF-36 findings from UC-only samples were selected. When possible, SF-36 mean scale scores were extracted and converted to norm-based T-scores and summarized across studies. RESULTS: Fifty-nine articles meeting all criteria were reviewed. Evidence supported the following: (1) adequate internal consistency (Cronbach’s α≥0.70) and test-retest reliability (intraclass correlations ≥0.70) for most SF-36 scales; (2) moderate-to-strong correlations between SF-36 scales and diseasespecific HRQL instruments (mean r: 0.45 to 0.65) and the Colitis Activity Index, a disease activity measure (mean r: -0.26 to -0.60, with absolute values >0.50 for Bodily pain, Vitality, and Social functioning scales); (3) clinically meaningful differences across UC patients grouped by disease activity status and symptoms (i.e., differences that exceed the established group-level minimal clinically important difference thresholds [ranging 2 – 4 points across scales]); and (4) responsiveness to treatment, indicated by statistically significant treatment-arm differences for most SF-36 scales in randomized-controlled trials. CONCLUSIONS: The SF-36 has strong psychometric properties when used with UC patients. Across studies, SF-36 scales showed robust associations with clinical measures of disease activity. The SF-36 also reflected improvements following treatments that reduced UC activity and severity. Results support the inclusion of the SF-36 as an endpoint in randomized-controlled trials for UC patients to evaluate the effect of treatment on patients’ generic health status. PGI31 THE HUMANISTIC AND ECONOMIC BURDEN OF IRRITABLE BOWEL SYNDROME WITH DIARRHOEA (IBS-D) BY DISEASE SEVERITY AMONG PATIENTS IN THE EU5 REGION Flores NM1, Tucker C2, Carson RT3, Abel JL3, Liebert R4 Health, Foster City, CA, USA, 2Allergan plc, Marlow, Buckinghamshire, UK, 3Allergan plc, Jersey City, NJ, USA, 4Kantar Health, New York, NY, USA 1Kantar OBJECTIVES: To assess the burden of IBS-D on health-related quality of life (HRQoL), activity impairment and healthcare resource utilisation (HCRU) based on disease severity among adults in the EU5 (Spain, France, Italy, Germany and United Kingdom). METHODS: IBS-D patients were identified from the EU5 2013 National Health and Wellness Survey, a self-administered, internet-based general health survey. Patients reported the frequency and nature of their IBS symptoms and disease severity (mild/moderate/severe). HRQoL, activity impairment and HCRU were assessed using the Short Form-36 version 2 physical and mental component summary scores (PCS, MCS; range 0–100), Work Productivity and Activity Impairment Questionnaire General Health Version (higher percentages indicate greater impairment) and number of patient-reported visits to any healthcare pro-