The purpose of this study was to evaluate, through a meta-analysis study, whether the use of computer-based systems reported in the literature improves the metabolic control of diabetic patients. On the retrieved papers, a set of meta-analysis studies were performed: first the difference of HbA1c between cases and controls at follow-up was evaluated (sign test); then the difference between cases and controls in the total variation of HbA1c from the beginning to the end of the trial was considered (method of effect sizes). The latter methodology was reapplied also on three more homogeneous article subgroups. The sign test was performed on 16 papers: in two of them, the HbA1c level was higher in the intervention group than in the control group at follow-up: it is unlikely that this is a random occurrence (p < 0.01). The method of effect sizes was first applied to 13 papers, as in the others some needed data were missing: the results obtained showed a statistically significant amelioration of metabolic control in the intervention group in comparison to the control group (p < 0.01). A progressive reinforcement of this outcome was obtained on the trial subgroups. Our study supports the hypothesis that the use of computer-based systems can be an effective means of improving metabolic control. The differential benefit obtained in the amelioration of HbA1c does not justify, by itself, the applicability of such systems into clinical practice; additional investigations should be carried out to evaluate the enhancement of other clinical and organizational indicators.