A14 AGAABSTRACTS • G0054 EVALUATION OF A PATIENT SYMPTOM MEASURE FOR CONSTIPATION. L. Frank, MEDTAP International, Inc., Bethesda, MD, C. Farup, R. Jones, Titusville, NJ; P. Miner, Oklahoma City, OK; L. Taylor, Bethesda, MD; L. Kleinman, Seattle, WA.
The evaluation of treatment outcomes in constipation is complicated by the lack of validated patient-based measures, To advance the field of assessment for constipation disorders, a self-report symptom assessment inventory (Patient Assessment of Constipation - Symptom Questionnaire: PAC-SYM) was developed, The purpose of this study was to evaluate the psychometric properties of the PAC-SYM. Methods: The PAC-SYM is a 44 item, selfadministered measure of symptom frequency and severity. Items were developed from 5 focus groups of patients with a range of severity levels and from the relevant literature. Psychometric properties were evaluated in a prospective observational study (2 week no-treatment run-in, 4 weeks usual care treatment) of 216 patients from 9 clinical sites in the U.S. A patient global rating was used to define treatment responders. Results" Comparison of treatment responders to nonresponders demonstrated the ability of the PAC-SYM to discriminate between groups according to clinical severity (p < .0001). Mean PAC-SYM scores changed significantly over time among responders, indicating instrument responsiveness. Compared to a global rating of condition change, PAC-SYM score changes of 15 points (9%) appear to be clinically meaningful. Internal consistency o f the PAC-SYM was high (Cronbach's alpha=0.875), indicating that the items jointly address a single construct. A measure of stability over time, the intraclass correlation coefficient, was high following the 2 week run-in (0.720) despite evidence of some fluctuation. Subscales were identified based on the clinical literature and patient focus groups, and multi-trait analysis showed good assignment of items to scales. Exploratory factor analysis identified five distinct factors: stool symptoms, rectal symptoms, abdominal symptoms, irritable bowel symptoms, and systemic symptoms. Concurrent validity is supported by the statistically significant correlation of PAC-SYM scores with both subject and investigator severity ratings (p < .0001). Further support is provided by correlation of specific items to bowel movement diary symptom data. Conclusion: Psychometric analyses demonstrate that the PAC-SYM is internally consistent, valid, responsive to change, and provides a comprehensive means of assessing constipation treatment effectiveness. This research was funded by Janssen Research Foundation, Titusville, NJ. • G0055 APPROPRIATENF~S OF UPPER GASTROINTESTINAL ENDOSCOPY: RANDOMIZED TRIAL OF ALGORITHM-BASED VERSUS FREE CLINICAL DECISION. F, Frgehlich, J.P. Vader, B. Burnand, C. Schneider, V. Wietlisbach, S. Realini, K. Gyr, G. Stalder, I. Pache, C. Beglinger, J.J. Gonvers. Gastroenterology Department and Institute for Social and Preventive Medicine, Lausanne, Switzerland; G1 Department, Basel, Switzerland Improving appropriateness of upper GI endoscopy (UGE) should enhance quality of care and possibly reduce costs as substantial overuse of UGE has been reported. This trial aimed at exploring whether applying explicit appropriateness criteria (EAC) to decide on the use of UGE affects patient outcome. Methods: Outpatients suffering from upper GI symptoms referred for UGE were randomized to free medical decision (i.e. perform UGE; group 1) or decision whether UGE was appropriate based on a computerized algorithm (group 2) which was based on EAC developed following the RAND/UCLA panel method. Possible decisions were either appropriate combined with equivocal, or inappropriate. At study enrollment and at 6 month follow-up, we assessed patient satisfaction (score 1-5;3 questions), symptoms (score 1-5; 8 questions), absenteism, medical consumption (drugs) and quality of life (SF-36;score 1-100). Results: 226 patients were included prospectively (group 1:116; group 2:110; 62% male, 67% dyspepsia). No significant differences for changes between baseline to follow-up were found between group 1 and 2, respectively, as regards symptoms (-4.1 vs -3.8 points;p=0.73), satisfaction (-0.62 vs -1.1 points;p=0.15), and overall SF-36 scores (physical score:+3.8 vs +1.8;p=0.14; mental score: +0.6 vs +2.5;p=0.27). Drug consumption and absenteism related to upper GI symptoms were similar. Among the 41 patients with inappropriate indications in group 2, 20 were endoscoped during follow-up. Conclusions: This randomized study shows that the use of an explicit algorithm to make the decision to perform UGE does not negatively affect patient outcome, as compared to free medical judgement. In contrast to diseases with more serious outcomes (e.g. myocardial infarction), currently used outcome measures may not be able to reveal significant differences due to the recurrent and generally benign course of dyspeptic symptoms. Trials including randomization to a group not being endoscoped are actually difficult to conduct as overuse of UGE is frequent in clinical practice. SNF 32.40522.94 • G0056 DOES HELICOBACTER STATUS ALONE AFFECT OUTCOME OF DYSPEPTIC PATIENTS? F. Froehlich, C. Schneider, JP. Vader, B. Buruand, V. Wietiisbach, P. Hildebrand, K. Gyr, G. Stalder, C. Beglinger, J.L Gonvers. Gastroenterology Department and Institute of Social and Preventive Medicine, Lausanne, Switzerland; G1 Department, Basel, Switzerland Helicobacter status is closely related with peptic ulcer disease and gastritis. It is not, however, known whether HP status alone affects patient outcome. We
GASTROENTEROLOGY Vol. 114, No. 4 therefore compared the outcome of HP + and HP - patients with upper GI complaints who were part of a previous trial (Endoscopy 1997;29:E8), assessing appropriateness of endoscopy. Methods: HP serology was determined at study entry (ELISA). Symptoms (8 questions, score 1-5) and quality of life (SF-36;score 0-100) were assessed at enrolment and at 6 month in all patients. Changes between baseline and follow-up were measured a) by mean difference of scores and b) by proportion of patients experiencing manifest improvement (e.g. upper quartile of the overall distribution of change). Stepwise logistic regression was performed for symptoms and SF-36 scores. HP eradication treatment was given only to HP positive patients with ulcers or erosive gastritis (n=12). Results: 167 patients were included prospectively: HP serology was available in 147 patients (HP +n=65;44%), HP - n=82;56%). Endoscopy was performed in 131 of these (no lesion:54; esophagitis:18; ulcer and/or erosive gastritis:23;other lesions:36). Baseline mean score levels were similar between HP + and HP - patients. Likewise, the mean changes between baseline and follow-up were not statistically different between both groups: symptoms ( - 4.8 vs - 3.1 score pts;P=0.08); SF-36 (physical scores: + 3.2 vs +2.4; P=0.57;mental scores +2.4 vs +0.8;P=0.39). However, a higher proportion of HP + patients experienced a favourable change in symptom severity (decline > 6 pts: 34% vs 20%;P=0.05) and mental well-being (rise of SF-36 score > 8pts:34% vs 20%;P=0.05) but not in physical well-being (28% vs 23%;P=0.53). Stepwise logistic regression showed that symptom improvement was only predicted by positive HP status (Odds ratio 2.11;P=0.05), but not by sex, age, endoscopic diagnosis or use of antisecretory drugs. However, restricting analysis to those not receiving eradication treatment gave similar results but which did not reach statistical significance. Conclusion: Helicobacter positive patients experience better outcome as regards symptoms and mental quality of life. This effect is only in part due to eradication treatment and needs further assessment. Swiss National Science Foundation Grant 32.40522.94 G0057
GENDER SPECIFIC PREDICTORS OF HEALTH CARE UTILIZATION FOR PATIENTS WITH IRRITABLE BOWEL SYNDROME. S. FuUerton, B. Naliboff, and E.A. Mayer. CURE Digestive Disease Research Center/Neuroenteric Biology Group, Depts. of Medicine, WLA VA Med. Ctr and UCLA, Los Angeles, CA 90024. Background: Between 20 and 30 percent of persons with irritable bowel syndrome (IBS) seek health care for their symptoms. In the U.S. and other developed countries women with IBS are twice as likely to utilize health care than men are. The psychometric and health related quality of life (HRQOL) predictors health care utilization may provide information on health care seeking behavioral differences by gender. Aims: The aims of this study are to explore gender specific psychometric and HRQOL predictors of the level of health care utilization by looking at item correlation for commonly used psychometric and HRQOL scales. Methods: Consecutive new patients referrals to the UCLA Neuroenteric Disease Program clinical studies group were given validated psychometric, health status, and bowel symptom questionnaires (BSQ). Between 1996 and 1997 527 BSQ containing a specific question regarding the number of doctor visits made by the patient specifically for bowel symptoms during the past 6 months. All patients had confirmed IBS diagnoses by Rome criteria. All linear regression analysis was age adjusted. Results: Predictors
Male (N = 176) R square P value
Psychometric Scales Anxiety 0.096 0.062 :Depression 0.092 0.071 Somatization 0.122 0.028* Quality of Life Scales Health Perception 0.106 0.045* Physical Functioning 0.133 0.020* Physical Role 0.112 0:038" Emotional Role 0.079 0 i03 I Social Functioning 0.252 0.0003"*'1 Mental Health 0.123 0.027* Bodily Pain 0.190 0.003** Energy 0.201 0.002** * - p < 0 . 0 5 , * * - p < 0 . 0 1 , ***-p<0.001
Female (n = 349) R square P value 0.027 0.010 0.025
0.234 0.580 0.260
0.060 0.034 0.035 0.014 0.072 0.033 0.148 0.031
0.045* 0.156 0.148 0.492 0.019" 0.179 0.0002*** 0.20
Conclusions: Predictors of the level of health care utilization differed between males and females. The HRQOL social functioning scale was the strongest predictor of the number of medical visits for men. Other strong predictors included bodily pain and lack of energy. Psychological factors do appear to predict level of utilization. For women, bodily pain was the strongest predictor of health care utilization, while psychological factors did not appear to predict the level of health care utilization.