Background
: Gastrofiberscopy(GFS) is a very important method to detect esophagogastroduodenal lesions which has high sensitivity. Biopsy and other interventions can be done by GFS. Despite of the merits, GFS is quite an uncomfortable procedure for patients. We studied to find discomfort of patients during GFS and to identify predictors for it.
Mothods : We recruited patients taken GFS at outpatient clinic, SNUH, Seoul, from June to July 1994. The patients recorded their discomfort during GFS by visual analogue scale score for 4 items respectively, pain, nausea, dyspnea, and air-fullness. The left end means no pain or discomfort, and the right end means extreme pain or idscomfort. Duration of examination. premedication. insertin method and the results were recorded. The X²test, simple linear regression, and multiple reogression were used. p value ≤ 0.05 was regarded as significant.

Results : 176 patients(mean age 45yr : 53% male) were recruited. The mean(SD) scores were : nausea 5.9(2.4), pain 3.9(2.5), dyspnea 3.7(2.7), air-fullness 3.6(2.8). The younger patients reported more nausea, dyspnea, air-fullness(P<0.05). Female patients reported more nausea than male (p<0.01). Insertion under direct visualization caused more nausea than blind insertion. Decompression of air-filling did not reduce patients' discomfort. Previous GFS exprience, biopsy operation, waiting time at endoscopy room, duration of examinationm education during procedure, bowel preparation, and premedication(IM buscopan, lidocaine gargling) and other characteristics of investigation pool were not associated with the degree of discomfort. The shorter time for insertion(from incisor teeth to upper esophagus), the more nausea was reported. But the association was not significant at p value ≤ 0.05.

Conclusion : The younger patient reported higher degree of discomfort during GFS. Female reported more nausea than male. Insertion under direct visualization caused more nausea than blind insertion. The other procedural and general factors did not show significant association in our finding.

Background
: Non-compliance with long-term pharmacologic treatment is an important cause of inadequately controlled high blood pressure. A self-reported questionnaire on medication compliance has already been tested in two specialized outpatient clinics in USA and validated.
However, a method of identifying noncompliant patients that is easy to apply has not been in practical use in our country. The purpose of this study was to test and validate a Korean version of the questionnaire in a family pracice setting.

Methods : We recreited a sample of diagnosed hypertensive outpatients receiving care in a family medicine unit July and september 1994(N=108). After measuring the blood pressure and administering the questionnaire, we tested the questionnaire against blood pressure control for which JNC V criteria was used.

Results : Fifty-five subjects give an affirmative answer to the first question, about forgetting medication, while a much smaller number of subjects to the 2nd, 3rd, and 4th questions. Fortyfour patients answered no to all four questions. About sixty percent answered yes to one or more questions. The questionnaire's reliability was tested with Cronbach's alpha (0.18) Concurrent validity of the questionnaire is followed: sensitivity=76%. The sensitivity for the first question is 63% and the specificity is 67%.

Conclusion : The self-reported quesionnaire of korean version has been shown to be useful in monitoring compliance to hypertension drug treatment. In the clinical setting, asking patients only the first question instead of all four questions will produce similar results. But further studies will be needed because the questionnaire's reliability was too low.