Comparing patient-generated blood glucose diary records with meter memory in diabetes: a systematic review

Aims To synthesize evidence relating to comparisons between patient-generated blood glucose records and meter memory in diabetes and to identify any predictors of agreement. Methods A systematic literature search was performed to identify articles comparing meter and diary records in those unaware of this assessment. Results Eleven observational studies, covering patients with Type 1, Type 2 and gestational diabetes were included spanning 1984-2009. Failure to record blood glucose measurements in the diary was the most extensive 'error', but addition of values, which were not measured, was a greater cause for concern. When present to a high degree, 'errors' lead to decreased variability in diary records compared with meter records. Allowing for a minimal amount of disagreement, just over 50% of adult diaries can be considered as 'accurate/reliable'. Disagreements were most extensive in teenagers and young adults, but the pregnant populations were only slightly better. Agreement was not related to sex, number of insulin injections or duration of monitoring. Those who were younger were more likely to have 'errors', while those who monitored more frequently had more 'accurate' diaries. Conclusions The lack of meter-diary agreement suggests that the real reason for monitoring is not understood by many patients, raising issues about motivation, perceived need to impress healthcare providers and denial of poor control. Considering that diaries are used to inform decisions about therapy when HbA(1c) is raised or in pregnancy, when HbA(1c) is not suitable, there is significant cause for concern in relation to their clinical utility.