Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Glucose self-monitoring and glycaemic control

Type 1 diabetics
Type 2 diabetics
Diabetics with type 1 diabetes using insulin are recommended to test their blood glucose levels several times a day. In the USA type 2 diabetics using insulin or oral hypoglycaemic agents are recommended to test their blood glucose at least daily. There are no recommendations about monitoring for type 2 diabetics using diet and exercise to control their glucose levels.

Glucose self-monitoring is not without cost, and the annual cost of strips for someone monitoring their glucose level three times a day is about $850 in the USA and £330 in the UK. The efficacy of glucose self-monitoring controversial. There are concerns over its effectiveness even for type 1 diabetics, let alone type 2 diabetics. A new study from the Kaiser Permanente diabetes registry in Northern California covering about three million people [1] may help.


The registry collects information about diabetics, including prescribed drugs and control strips, outpatient and hospital visits, and is estimated to have information on 96% of all diabetics in the programme. The study set out to examine the average daily strip use by diabetics throughout 1996, and relate that to the first haemoglobin A1c measurement in 1997 as a measure of glycaemic control. At the time of the study there were no Kaiser Permanente practice guidelines for self-monitoring.

Information from patients about themselves and their diabetes was collected by questionnaire or telephone interview, including information about treatment and use of diet and exercise. Automated pharmacy records were used to calculate glucose strip use, and linked records to evaluate emergency room and hospital admissions.

The American Diabetes Association 1997 guidelines were for self monitoring three times a day for type 1 diabetics, and daily for type 2 diabetics treated with insulin or oral hypoglycaemic agents. These rules were used to define adherence and non-adherence.Definitions of average daily reagent strip use from prescriptions are shown in Table 1.

Table 1: Definition of average daily reagent strip use from prescription analysis

Strips a day
Three times daily or more ≥2.5
At least daily <2.5 but ≥0.75
Occasional <0.75 but >0
No self monitoring 0

Haemoglobin A1c results were corrected for demographic differences between different categories of strip use. Information on excluded patients was also given and results were similar to those in the main study.


There were just under 49,000 diabetics with full membership of the health programme for the whole of 1996 and 1997, and 23,412 who additionally responded to the survey, in whom type of diabetes could be determined and in whom haemoglobin A1c was measured in 1997.

Type 1 diabetics

There were 1160 such patients, of whom 395 measured blood glucose on average three times a day. Their haemoglobin A1c was, on average 1.2% lower (7.6%) than those who monitored less frequently (8.8%). Adherent diabetics who monitored three times a day tended to be a few years older and more likely to be women than non-adherent diabetics. More monitoring was associated with lower levels of haemoglobin A1c (Figure 1).

Figure 1: Average 1997 haemoglobin A1c values according to diabetes diagnosis and average daily glucose strip use during 1996

Type 2 diabetics

Seven thousand five hundred of 23,000 patients measured their glucose at least daily. Haemoglobin A1c in diabetics treated with insulin or oral hypoglycaemic agents was 0.7% lower (8.1%) than in those who monitored less frequently (8.8%). Adherent diabetics who monitored once a day tended to be older and were more likely to be women than non-adherent diabetics. More monitoring was associated with lower levels of haemoglobin A1c (Figure 1).

Adherent type 2 diabetics were more likely than non-adherent diabetics to attend for eye examination and use diet and exercise as part of their therapy. They also had more frequent emergency room attendance (31% vs 26%) and hospital admissions (17% vs 13%), but we don't know if these were for hypoglycaemic episodes.


Evidence for the benefits of glucose self-monitoring, especially in type 2 diabetics, is lacking. The paper [1] has a good reference list for the evidence that exists. Some randomised studies have been done, and there is even a good meta-analysis [2], but a good meta-analysis of poor trials. The total number of patients for analysis in randomised trials was 558, dwarfed by this study, and with high dropout rates and other methodological problems. New randomised trials are unlikely, despite the apparent equipoise.

So observational studies assume increasing importance, especially when, like this, they are large and detailed. The result, that increased self monitoring up to recommended levels is associated with better glycaemic control, is the expected one. Unfortunately such trials do not allow us to conclude that increased self-monitoring results in better glycaemic control, though that might seem obvious.

For type 2 diabetics in particular there were interesting differences between those patients who measured their blood glucose daily and those who did not. This limits the applicability of the study, and the authors themselves conclude that this is only supportive evidence in favour of the benefits of self-management in type 2 diabetics.

This was a relatively short-term study. Bandolier 84 described a long-term study from Washington State where diabetic patients who achieved a 1% fall in haemoglobin A1c reduced their overall healthcare costs by almost $1,000 a year, and sustained it, but only after the second year.

These large, real-world database studies emerging from United States healthcare providers often have important academic affiliations. This study [1] had NIH and American Diabetes Association support. The motives of better health care and lower cost are often mutually supportive.


  1. AJ Karter et al. Self-monitoring of blood glucose levels and glycaemic control: the Northern California Kaiser Permanente diabetes registry. American Journal of Medicine 2001 111: 1-9.
  2. S Coster et al. Self-monitoring in type 2 diabetes mellitus. A meta-analysis. Diabetic Medicine 2000 17: 755-761.
previous or next story in this issue