OBJECTIVES: Health promotion using web-based tools may provide some benefit to patients with diabetes mellitus. This meta-analysis assessed the effectiveness of web-based tools in improving blood glucose control. METHODS: Randomized, controlled trials (web-based tools vs. usual care) were identified from research databases using the following criteria: participants were patients with diabetes (type 1 or 2) who had suboptimal blood glucose control (glycated hemoglobin [A1C] >7%), and outcomes included A1C. Two reviewers independently screened and extracted data. Study quality was evaluated based on randomization; concealment of allocation; blinding; accounting for dropouts and losses to follow-up; and overall quality based on the Grading of Recommendations Assessment, Development and Evaluation scale. RESULTS: Nine studies were included. Pooled estimates showed a significant mean difference in A1C favouring webbased tools. The mean differences were -0.71% (95% CI -1.00, -0.43) after 3 months, -0.52% (95% CI -0.75, -0.29) after 6 months and -0.55% (95% CI -0.70, -0.39) after 12 months. There was heterogeneity among studies with 12 months of intervention (I2=78%). There was also a significant mean difference in low-density lipoprotein cholesterol (LDL-C) favouring web-based tools (-0.23 mmol/L, 95% CI -0.28, -0.19). CONCLUSION: This study showed that web-based tools were better than usual care in improving A1C and LDL-C. Future studies should assess the cost benefit of web-based tools and further improve their effectiveness.