Blood glucose discrimination training in patients with type II diabetes

This study was designed to determine whether patients with Type II diabetes could be taught to discriminate blood glucose after experiencing a variety of blood glucose levels and receiving feedback on the accuracy of their estimates. Thirty-six subjects (18 on oral agents and 18 on insulin) were randomly assigned to one of two feedback conditions: (1) current feedback, which received accurate information regarding their blood glucose levels, (2) noncurrent feedback, which received blood glucose levels from the preceding session. Subjects were exposed to a wide range of blood glucose values in six training sessions by ingesting drinks with three different caloric loads. In pre/post comparisons using several indices of accuracy, both groups showed significant improvement in estimating blood glucose levels. However, feedback on current blood glucose levels did not produce greater improvement than noncurrent. Accuracy was unrelated to the degree to which subjects reported associating internal sensations to their estimates. Failure to find differences between the two feedback conditions may have been due to the noncurrent feedback group's receiving fairly accurate information, to the difficulty of the discrimination task, and to the limited number of training trials.