This study evaluated the effect of diabetic group teaching programs in one university hoapital in Seoul to predict when re-education would be needed. This study examined the patients' knowledge at four points (before, directly after, three months and six months after the teaching program) and self-care performance related to diabetes twice (three months and six months after the teaching program). The subjects of the study were 34 admitted diabetic patients who participated in four-day teaching programs. Data were gathered from January to October, 1992 by means of an instrument developed from two diabetic knowledge tests which were equivalent in item differentiation and item discrimination coefficients. Collected data were analyzed by paired t-test, Pearson correlation, t and F tests. The results of study were as follows. 1. The analysis of the effect of the diabetic group teaching program and the duration of the effect of teaching. 1) The first hypothesis, that the diabetes knowledge score directly after the teaching program would be higher than before, was rejected (t=-1.40 ;p=.172). 2) The second hypothesis, that the diabetes knowledge score directly after teaching would be higher than three months later, was rejected(t=-4.27; p=.000). 3) The third hypothesis, that the diabetes knowledge score three months after teaching would be higher than six months later, was supported(t=2.43; p=.020). 2. The relation of knowledge and self-care performance 1) The forth hypothesis, that the level of self-care performance related to diabetes three months later would be higher than six months later, was rejected(t=1.49; p=.146). 2) The fifth hypothesis, that the higher the diabetes knowledge, the higher the level of self-care performance, was rejected(r=.2086: P=.118). 3. The relation of diabetes knowledge and self-care performance according to demographic variables and structural variables of diabetes. 1) Diabetic knowledge scores varied according to the educational levels of the clients directly after the teaching. Three months after the teaching program higher educational levels and higher economic status were related to higher diabetic knowledge scores and men had higher knowledge scores than women. 2) Self care performance scores of men were higher than those of women at three months and six months after the teaching program. 3) Before the diabetes teaching, the diabetic knowledge scores of subjects who had a dia-betic patient in the family were higher than those who did not have patient in their family. Six months after the teaching, the diabetic knowledge scores of subjects who read the distributed books about diabetes were higher than those who did not read them. 4) No significance differences were found between self-care performance and structural variables of dialetes. The results of this study indicated that the levels of diabetes knowlege and self-care performance incease of three months after the teaching program but decrease of six months. Reeducation would be needed between three and six months. The investigator thinks that a study of the content and teaching methodology is needed to increase the education effect. The subjects want to hear patient histories of diabetic management. Group discussion would be helped after the teaching sessions.