Accuracy of subjective blood glucose estimation by patients with insulin-dependent diabetes

A between-groups design using a baseline, treatment, follow-up procedure was used to investigate the accuracy of 20 patients with insulin-dependent diabetes when subjectively estimating their blood glucose levels. Patients were encouraged to attend to their mood for cues when making estimates of their blood glucose. Their capacity for reducing estimation errors when given immediate or delayed feedback of actual blood glucose was examined. The results showed that neither delayed nor immediate feedback produced a significant improvement in the mean estimation accuracy of these groups of patients or in their ability to predict whether their blood glucose was in the acceptable or unacceptable range. Patients were particularly inaccurate in detecting Low <4.0 mmol/L (<72.0 mg/dl)] and Very High >16.0 mmol/L (>288.0 mg/dl)] blood glucose levels. Examination of moodblood glucose relationships revealed consistent patterns for individual subjects and considerable differences between subjects.The authors wish to thank the consultant physician Dr. J. D. Ward, and Drs. G. Knight and A. J. M. Boulton from the Royal Hallamshire Hospital, Sheffield, for referring the patients who took part in the study. The research was funded by NIH Grant number AM28196 to Dr. C. Bradley.