Clinical evaluation of decision support system for insulin-dose adjustment in IDDM

We developed a wallet-sized learning memory decision support system that helps patients with insulin-dependent diabetes mellitus adjust their insulin dosages.

To determine the efficacy of the support system, we designed a randomized clinical trial with patients participating in a program in a diabetes education center. Patients were assigned to two groups of 21 patients each. All patients performed self-monitoring of blood glucose (SMBG) and were treated with multiple daily injections of insulin. Each of the patients was examined over a 32-day period. The basic educational program, i.e., practical advice in SMBG, diet, and exercise under homelike conditions, was identical in both groups. The only difference was that the first group used the computer for adjusting the insulin dose, whereas the second group received recommendations from the education team.

The baseline HbA1 levels (9.8 +/- 1.6 vs. 9.9 +/- 1.6%) of both groups did not differ significantly. Mean blood glucose over the last 2 wk of the study was higher (P less than 0.01) in the second group (8.4 +/- 1.4 vs. 9.2 +/- 2.0 mM); the frequencies of hypoglycemic episodes were not different (1.7 vs. 2.3%).

Metabolic control and safety were comparable in both groups. Thus, patients may benefit from such a system at home where no support by diabetes educators is available.