Background: Compared with traditional outpatient follow up, short message service (SMS) provides more convenience for diabetes self-management. However, the feasibility, effectiveness, and satisfaction of SMS intervention remain unclear. Methods: A systematic retrieval of databases, including PubMed, EMBASE, and Cochrane Library, was used to evaluate the effect of SMS on the diabetes glycemic control. Complete quantitative reports on the changes of glycosylated hemoglobin (HbA(1c)) before and after intervention were collected. The study type was limited to randomized controlled trials (RCTs). Results: 13 RCTs were identified as eligible for this subject. Overall result revealed a statistical decline of HbA(1c) by -0.62% (95% CI -0.82 to -0.41). The longest intervention duration was 12 months with the noteworthy improvement of HbA(1c) by -1.63% (-2.27 to -0.99). Compared with control groups, five trials quantitatively demonstrated a significant decline in fasting plasma glucose (FPG) by -0.45 (95% CI -0.7 to -0.19) and -1.05 (95% CI -1.35 to -0.75) in 2-h postprandial blood glucose (2hPBG). Conclusions: This meta-analysis demonstrated that text message intervention indeed leads to a decline in HbA(1c) and improvement of blood glucose control. No convincing evidence was found on quality of life improvement, satisfaction, body mass index (BMI), blood lipid levels, the frequency of hypoglycemia, self-monitoring of blood glucose or complications reduction.
