513 PATIENT ENGAGEMENT IS ASSOCIATED WITH HEALTH RELATED QUALITY OF LIFE IN ADULTS WITH CYCLIC VOMITING SYNDROME Geoffrey Dang-Vu, Lisa Rein, Brandon Berens, Aniko Szabo, Thangam Venkatesan Background and Aims Patient engagement, adaptation and self-management in health behaviors play a critical role in improving Health Related Quality Of Life (HRQOL) and reducing health care utilization in chronic disorders. The Patient Activation Measure (PAM) is a validated tool that measures the degree to which patients are engaged to manage their own health care. PAM Level 1 denotes patients who are disengaged and overwhelmed. Level 2 indicates patients who are becoming aware, but believe that health is largely out of their control. Level 3 indicates that patients are taking action and control of their health and Level 4 indicates patients who have adopted new behaviors and are able to utilize them in stress or health-crises. Cyclic Vomiting Syndrome (CVS) is a chronic functional GI disorder that is associated with poor HRQOL, increased health care utilization and disability. There is no data on the level of patient engagement in patients with CVS; we thus sought to determine their level of engagement and its association with clinical variables. Methods The PAM questionnaire was administered prospectively to patients with CVS seen at a tertiary referral center between 2014-2015. Data on demographics, clinical characteristics, health care utilization, disease severity, and HRQOL (using the NIH PROMIS 10) were obtained. Patients were stratified into low engagement (PAM 1 & 2) and high engagement (PAM 3 & 4). The Fisher's exact test and Wilcoxon rank-sum tests were used to identify significant differences between the groups. Results Of a total of 96 patients, 58.3% were female and mean age was 35.9 + 12.8 years. PAM scores were normally distributed with a mean of 56.8 + 13.2. Twenty-five % of patients were at Level 1, 19.8% at Level 2, 40% at Level 3 and 15% at level 4. Lower patient engagement (Levels 1 & 2) was significantly associated with worse global physical and mental HRQOL (p=0.002 and 0.001), smoking tobacco (p= 0.015), more CVS related hospitalizations within the past year (4.7 vs. 2.3 P=0.030), anxiety and depression (p=0.022 and 0.007), and being a new vs. established patient in the PI's clinic (p= 0.001) (Table 1). There was no association between patient engagement and age, gender, BMI, chronic narcotic use, chronic marijuana use, duration of CVS, number of CVS episodes/year, number of urgent care visits/year or comorbid conditions such as chronic fatigue, fibromyalgia, migraines and panic disorder. Conclusion Almost half of CVS patients demonstrate poor patient engagement, which is associated with poor mental and physical HRQOL. Extra-intestinal conditions such as anxiety and depression, important drivers of patient engagement contribute to poor HRQOL. A biopsychosocial model of care addressing mental health in CVS should improve patient engagement and overall quality of life. S-117 AGA Abstracts Comparison of Demographics and Clinical Characteristics of CVS Patients with Low and High Patient Engagement * denotes p-value <0.05. Categorical variables were analyzed using Fisher's exact test and continuous variables using Wilcoxon rank-sum. Categorical variables are expressed as n (% of patients) and continuous variables expressed as mean ± standard deviation. 514 DEFINING RESPONSE TO THE PAC-SYM QUESTIONNAIRE IN PATIENTS WITH CHRONIC CONSTIPATION Yan Yiannakou, Jan F. Tack, Hubert Piessevaux, Dominique Dubois, Eamonn Quigley, Meiyun Ke, Susana Da Silva, Alain Joseph, René Kerstens Introduction The Patient Assessment of Constipation-Symptoms (PAC-SYM) questionnaire was developed through psychometric evaluation of patients with constipation; observational data have shown it to have internal consistency, test-retest reliability, concurrent validity and to be responsive to change over time. 1 This 12-item questionnaire, divided into three symptom subscales (abdominal, rectal and stool), is increasingly used as a patient reported outcome measure in clinical trials of constipation. Scores on 5-point Likert scales (range: 0-4 [absent-very severe]) are used to generate an average total score. Historically, a reduction of ≥ 1 point in total score defined response. However, the mean baseline total score is approximately 2, 2 so a change over time of -1 point would represent a 50% reduction in symptoms, possibly underestimating the true clinical benefit as assessed by patients and clinicians. Aims & Methods The aims of this analysis were to determine the minimum clinically important difference (MCID) and an optimum PAC-SYM cut-off score for defining responders using Week 4 data from six phase 3 and 4, double-blind, randomized, placebo- controlled trials of prucalopride in chronic constipation (NCT01147926, NCT01424228, NCT01116206, NCT00485940, NCT00483886, NCT00488137). Overall, 2884 patients (621 men; 2263 women; prucalopride ≤ 4 mg, 1719; placebo, 1165) were included in the analysis. MCID analyses were performed using anchor based methods. Anchors (outcomes) used were defined by expert consensus. They comprised: a) moderate global efficacy of treatment; b) 1 point improvement in global severity of constipation; c) moderately severe constipation, measured by an overall severity question; d) change of 0.5-1.5 spontaneous complete bowel movements (SCBMs)/week; and e) 1.5-2.5 SCBMs/week. Receiver Operating Characteristics (ROC) analyses were performed to derive an optimal cut-off value for change in PAC-SYM score and corresponding sensitivity and specificity values. Results The MCIDs for different anchors were: -0.62 for moderate global efficacy of treatment; -0.62 for a 1 point change in global severity of constipation; -0.53 for moderately severe constipation; -0.52 for a change in the number of SCBMs/week of 0.5-1.5; and -0.63 for 1.5-2.5 SCBMs/ week. ROC analyses showed optimum cut-off values for discriminating responders to be slightly lower but in the same range. Conclusions Anchor based methods gave consistent results for the MCID, at around -0.6. This value was close to the optimal cut-off score for responder discrimination in the population studied, and could be considered in future clinical trials and clinical practice. A slightly higher threshold, e.g. -0.75, could be used to reduce the placebo response rate in clinical trials. 1 Frank L et al. Scand J Gastroenterol 1999;34:870-7 2 Camilleri M et al. Dig Dis Sci 2016;61:2357-72 AGA Abstracts