Forty patients were randomly assigned to receive either the individualized learning activity packages or classroom instruction. Effectiveness was evaluated on the basis of fasting plasma glucose, glycosylated hemoglobin, percent of ideal body weight, and knowledge and behavior scores. No significant differences were found between the groups at entry or at the 2-month follow-up. At the 5-month follow-up, the group that received the learning activity packages scored significantly higher on knowledge assessment, significantly increased their behavior score, and decreased their percent of ideal body weight. Patients who received classroom instruction increased their behavior score and exhibited significantly decreased glycosylated hemoglobin levels. Behavior and blood glucose levels were significantly correlated. Although the learning activity packages proved effective in increasing knowledge, no significant improvement was observed in blood glucose levels