Self monitoring of glucose by people with diabetes: evidence based practiceBMJ 1997; 314 doi: http://dx.doi.org/10.1136/bmj.314.7085.964 (Published 29 March 1997) Cite this as: BMJ 1997;314:964
- Marilyn Gallichan, diabetes specialist nursea
- Accepted 21 November 1996
The inappropriate use of self monitoring of glucose is wasteful of NHS resources and can cause psychological harm. Although a few patients find that self monitoring enables them to understand and take control of their diabetes, many people with diabetes are performing inaccurate or unnecessary tests. There is no convincing evidence that self monitoring improves glycaemic control, nor that blood testing is necessarily better than urine testing. It may be appropriate for some patients not to monitor their own glucose but to rely instead on regular laboratory estimations of glycaemic control. Glucose self monitoring should be performed only when it serves an identified purpose.
It is widely assumed that glucose self monitoring, preferably of blood glucose concentrations, is desirable or even essential for everyone with diabetes. It is common for patients who have previously tested their urine, or have done no glucose monitoring at home, to be taught to measure their blood glucose when they are admitted to hospital. In the community too, patients are often encouraged to monitor their blood glucose, and newly diagnosed patients of all ages are usually taught to measure their blood glucose concentrations. Self monitoring can sometimes be useful, but evidence is mounting that its indiscriminate use is of questionable value. In 1995, £42.6 million was spent on home monitoring of glucose in the United Kingdom (Intercontinental Medical Statistics, personal communication). Is this enormous cost justified? Is blood testing necessarily better than urine testing? Is glucose self monitoring always necessary, or is it sometimes a waste of time and money? Are recommendations for self monitoring based on sound evidence?
We now have conclusive evidence that improved control of glycaemia is associated with a significantly lower risk of the complications of diabetes1, but there is no convincing evidence that glycaemic control is consistently influenced by self monitoring of blood or urine. When it was first introduced, home monitoring of blood glucose was claimed to lead to a sustained improvement in glycaemic control in insulin dependent diabetes.2 3 However, the absence of control groups in these studies has made it impossible to separate the effects of increased education and medical attention from the effects of the blood testing itself. More recent studies have suggested that regular self monitoring of blood glucose may be a waste of time for many patients receiving insulin. A comparison of two groups of such patients aged over 40 showed that patients who tested their blood did not have better glycaemic control than those who tested their urine,4 and in a study of young people with insulin dependent diabetes there was no difference in glycaemic control between those who tested their blood frequently and those who did not.5
Even less evidence links self monitoring of glucose with improved glycaemic control in non-insulin dependent diabetes. Randomised comparative trials of self monitoring in non-insulin dependent diabetes have found no difference in glycaemic control between patients who tested their blood and those who tested their urine.6 7 A comparison of patients with non-insulin dependent diabetes who carried out self monitoring and a matched group who did not,8 a randomised comparative trial9, and a retrospective study10 all found no difference in the control of those who monitored their glucose and patients who did not test their blood or urine at all.
Guilt or empowerment?
Self monitoring of blood glucose enables some patients treated with insulin to take control of their diabetes, allowing them to adjust their insulin dosage or diet in the light of their results, especially in relation to exercise, illness, or dietary changes. For these people, the ability to take an instant measurement of blood glucose and act on the result is enormously helpful. It improves the quality of life and amply justifies both the inconvenience of carrying around the testing equipment and the discomfort of the test itself. However, only 10-15% of diabetic people in Britain have insulin dependent diabetes,11 and only a relatively small proportion of this group have the need or desire to make frequent adjustments to their insulin dosage.
Most patients do not alter their treatment on the basis of results of their self monitoring but merely collect results that may or may not be useful to their doctor or nurse. For many of these people, self monitoring proves to be counterproductive. Anxiety is often generated when values repeatedly fall outside the desired range, and patients may experience feelings of frustration, helplessness, or guilt.12 For some people, no matter how hard they try and how rigidly they adhere to their treatment regimen, their blood glucose values continue to fluctuate in an alarming way. Not surprisingly, they may become fearful of the complications of diabetes, and this fear may lead to despair and to the psychological state of learned helplessness, which describes a combination of a loss of motivation, emotional disturbance, and cognitive impairment, induced by repeated exposure to an uncontrollable and unpleasant situation.13 Some become obsessive in their self monitoring. For example, I recently met a man who became so anxious if he had not tested his blood within the last few hours that he would have to perform a blood test to discover whether his trembling and sweating were due to hypoglycaemia or to his anxiety that he had not tested.
Reliability of tests
Even when patients perform regular blood or urine tests and religiously record the results in their home monitoring diaries, can we rely on the accuracy of these measurements? Despite appropriate training, almost half of patients testing their blood may obtain inaccurate results through poor technique14 and, although portable blood glucose meters have become much simpler to use, they are not yet foolproof.
As well as technical inaccuracies, deliberate falsification of results is common across all age groups and social classes. By asking patients to use blood glucose meters with a hidden memory, researchers showed that the results recorded in home monitoring diaries were often lower than the actual readings. Patients frequently omitted to record high readings and made up extra results so that it appeared that they had tested more frequently than they had in reality.15 Colin Dexter, the writer who created Inspector Morse, admits to making a New Year's resolution for 1996 not to invent quite so many satisfactory blood sugar readings when he goes for his diabetic check ups.16
Individual self monitoring plans
Many patients abandon self monitoring tests if their purpose is not clear.17 Home glucose monitoring should be performed only if it serves an identified purpose that is clear to both the patient and the nurse or doctor (box).
Box 1–Purposes of self monitoring of glucose
To provide patients with information about their day to day glycaemic control, enabling them to make appropriate adjustments to their diet or diabetic medication, especially in relation to illness, strenuous exercise, or potentially dangerous activities such as driving
To provide the nurse or doctor with information about the patient's day to day glycaemic control, enabling them to give appropriate treatment advice–for example, after a raised measurement for glycosylated haemoglobin
To detect hypoglycaemia: home monitoring of blood glucose can confirm or rule out hypoglycaemia
Most people with diabetes feel guilty that they do not test often enough.18 This can be avoided if an individual home monitoring plan is agreed. This should include the method, timing, and frequency of tests and a review date. Regular reviews of the plan will prevent unnecessary testing after the need for tests has passed and will also lessen the guilt experienced by patients who fail to comply with the testing regimen recommended by their nurse or doctor. The patient should be able to perform the test accurately, according to the manufacturer's instructions, and must know what results to expect and what action to take if the results are outside the desired range.The method of monitoring should depend on the purpose of monitoring and the patient's manual dexterity, visual and cognitive ability, and personal preference.
Glucose monitoring methods
Urine testing remains a useful method of monitoring glycaemic control, especially among older patients, when the aim of treatment may not be strict normoglycaemia. Results should ideally show that the urine is free of glucose, indicating that the blood glucose concentration has not risen above the renal threshold.
Most patients use reagent test strips. Diastix (Bayer Diagnostics) require accurate timing of 30 seconds; Diabur-Test 2000 (Boehringer Mannheim) tests take longer, but the timing is less critical. Clinistix (Bayer Diagnostics) test strips are unsuitable for self monitoring as, being designed for screening for glycosuria, they do not show the range of results available from the other methods. Clinitest (Bayer Diagnostics) urine testing kits, with tablets, droppers and test tubes, are still available for patients who prefer this method, and the larger volume of colour in the sample is often helpful for those with poor vision.
Self monitoring of blood glucose is usually the method of choice for younger patients and for most patients treated with insulin. It is particularly useful during pregnancy and for women planning pregnancy and is the only method that can detect hypoglycaemia. BM 1-44 (Boehringer Mannheim), Glucostix (Bayer Diagnostics), and Hypoguard blood glucose test strips can all be read visually, and there is now a wide range of meters which measure blood glucose without the need for visual comparison.
The size of the blood drop, whether it is smeared or dropped, and the preciseness of the timing are some of the many factors that can significantly affect the reliability of blood testing. It is therefore essential to assess the accuracy of patients' tests, either by direct observation of technique or by asking them to complete a blood spot series for laboratory analysis, allowing comparison of the patients' own recorded results with the laboratory measurements on the same samples. Although ideally most blood tests should be within the normal range, the longer term risks of hyperglycaemia must be balanced against the immediate risks of hypoglycaemia, especially in elderly patients, and an appropriate individual target range should be agreed.
No self monitoring
Some patients may prefer not to monitor their glucose levels, choosing rather to rely on regular laboratory estimations of glycaemic control. Concentrations of glycosylated haemoglobin in venous blood samples are reliable measures of average glycaemia in the preceding 50-60 days in both insulin dependent and non-insulin dependent diabetes.19 In non-insulin dependent diabetes, the fasting plasma glucose value is a reliable indicator of prevailing glucose concentration.20 It is not only elderly patients who may opt for no self monitoring. In a recent letter to a diabetes journal, a psychologist with insulin dependent diabetes wrote, “I do not worry about my day to day control provided the HbA1 readings stay good. It is only during periods of illness (eg, a viral infection) that I feel the need for daily blood testing.” 21
Frequency of testing
There are wide variations between individuals and between localities in the method, timing, and frequency of self monitoring tests, and many unhelpful or unnecessary tests are performed.7 In addition to the unnecessary physical discomfort, inconvenience, and possible adverse psychological effects, every unnecessary urine test wastes 5p, and every unnecessary blood test wastes 28p plus the cost of lancing devices, lancets, blood glucose meters, and cotton wool. With several tests per day by hundreds of thousands of people, the potential waste of scarce NHS resources is phenomenal.
The patient and doctor or nurse should agree on the most appropriate timing and frequency of tests for any individual. Tests before breakfast, mid-morning, and two hours after the main meal are the most useful for assessing hyperglycaemia; hypoglycaemia is most likely before meals or at night, and tests before the main meals and at bedtime are useful in assessing the efficacy of the preceding insulin doses. It is often useful to perform several tests a day during illness or a change of treatment, but when diabetes is stable one of the regimens shown in the box may be appropriate.
Evidence based practice
Existing evidence suggests that self monitoring does not improve glycaemic control, that blood testing is not necessarily better than urine testing, and that it may be appropriate for some patients to perform no self monitoring at all. Research shows that some patients give up self monitoring if they cannot see its purpose, while others feel guilty if they do not comply with the recommended monitoring regimen. Others experience anxiety, frustration, and helplessness in the face of unsatisfactory results which they feel powerless to improve. It has been shown that self monitoring tests are often inaccurate and unreliable and that inappropriate and unhelpful testing is widespread.
If self monitoring is to serve a useful purpose, recommendations must be based on the available evidence. The patient must know why, when, and how to test and how to interpret the results. Avoiding inappropriate and unnecessary tests will result in enormous cost savings to the NHS, as well as increasing the psychological wellbeing of people with diabetes.
Box 2–Self monitoring regimens
Diabetes not treated with insulin:
Test before breakfast and two hours after the main meal on one or two days each week
Diabetes treated with insulin:
Test before meals and at bedtime on one or two days each week, or
Test once a day, varying the timing of the test–for example, before breakfast on Monday, before lunch on Tuesday, and so on
Occasional tests at 2 am or 3 am may also be useful
Conflict of interest: None.