Self-monitoring in Type 2 diabetes mellitus: a meta-analysis.

AIMS Self-monitoring of blood or urine glucose is widely used by subjects with Type 2 diabetes mellitus. This study evaluated the effectiveness of the technique at improving blood glucose control through a systematic review and meta-analysis. METHODS Randomized controlled trials were identified that compared the effects of blood or urine glucose monitoring with no self-monitoring, or blood glucose self-monitoring with urine glucose self-monitoring, on glycated haemoglobin as primary outcome in Type 2 diabetes. RESULTS Eight reports were identified. These were rated for quality and data were abstracted. The mean (SD) quality score was 15.0 (1.69) on a scale ranging from 0 to 28. No study had sufficient power to detect differences in glycated haemoglobin (GHb) of less than 0.5%. One study was excluded because it was a cluster randomized trial of a complex intervention and one because fructosamine was used as the outcome measure. A meta-analysis was performed using data from four studies that compared blood or urine monitoring with no regular monitoring. The estimated reduction in GHb from monitoring was -0.25% (95% confidence interval -0.61 to 0.10%). Three studies that compared blood glucose monitoring with urine glucose monitoring were also combined. The estimated reduction in GHb from monitoring blood glucose rather than urine glucose was -0.03% (-0.52 to 0.47%). CONCLUSIONS The results do not provide evidence for clinical effectiveness of an item of care with appreciable costs. Further work is needed to evaluate self-monitoring so that resources for diabetes care can be used more efficiently.