Six RCTs (2,552 patients, range 89 to 689) were included in the review. Three trials described the randomisation method and all conducted allocation concealment and blinded outcome assessment. Loss to follow-up ranged from 2% to 31%. Four trials were industry sponsored.
HbA1c was significantly reduced as a result of self-monitoring interventions at six months (adjusted estimate -2.7 mmol/mol, 95% CI -3.9 to -1.6; six trials), which represented a 0.25% decrease.
Significant reductions were also reported at three months (-2.0 mmol/mol, 95% CI -3.2 to -0.9; five trials) and 12 months (-2.5 mmol/mol -4.1 to -0.9; three trials), which represented decreases of 18% (three months) and 23% (12 months).
There were no statistically significant differences for blood pressure (three trials) and serum cholesterol (three trials) outcomes.
Subgroup analysis did not generally show any significant differences, but trials that analysed the impact of patients new to testing were too small to ascertain this with confidence. Sensitivity analysis did not materially alter the main analysis. The authors stated that there was no evidence of heterogeneity in outcomes.