Two new studies suggest that newly diagnosed people with type 2 diabetes are unlikely to gain any benefits from monitoring their blood glucose themselves. Self-monitoring of blood glucose (SMBG) may also result in a lower quality of life. In people with established type 2 diabetes, SMBG approximately doubles the net health-care costs with no benefits and a decrease in quality of life. SMBG in newly diagnosed patients and those with established type 2 diabetes should now be reserved for certain people treated with insulin and, conceivably, in some very specific circumstances (e.g. patients who are at risk of hypoglycaemia during intercurrent illness or fasting). Instead, we should direct our attention to interventions likely to make a difference to patients’ symptoms and cardiovascular risk, and consider using the resources currently allocated to SMBG to help this. These include support and advice around nutrition, exercise, smoking cessation, foot care, etc.
Patients, health professionals and commissioners of health care should look carefully at the use of SMBG. These two studies cast further doubt on its usefulness in newly diagnosed patients and those with established type 2 diabetes, other than for certain people treated with insulin and, conceivably, in some other very specific circumstances.
More benefit may well come from directing attention to interventions likely to make a difference to patients’ symptoms and cardiovascular risk. It may be worthwhile for commissioners to consider using the resources allocated to SMBG to fund increased focus on support and advice around nutrition, exercise, smoking cessation, foot care, etc. In 2004, it was estimated that around 1.5 million people had type 2 diabetes. If we make a conservative estimate, that SMBG could be discontinued in two-thirds of these, the DiGEM analysis suggests that around £90 million could be found each year to fund such interventions.
What is the background to this?
SMBG is discussed in detail on the type 2 diabetes section of NPC. Although different studies have produced slightly different results, it seems clear that in most people with established type 2 diabetes who are trying to exercise strict glycaemic control with diet or tablets, the long-term effect of SMBG on HbA1c and major clinical outcomes is, at best, modest, may be non-existent or is even harmful.
It has been suggested that, in the recently diagnosed, SMBG may help patients understand more about the implications of food choices, promote adherence to medication and improve satisfaction with treatment. However, this does not seem to have been assessed in a clinical trial until now. In addition, although it is known that the NHS spends more on the prescription costs of SMBG materials than it does on oral hypoglycaemic drugs, there has not been a UK-based cost-effectiveness assessment of SMBG.
What did these studies find?
In people newly diagnosed with type 2 diabetes, O’Kane et al found that adding SMBG (with advice and guidance on how to respond to high or low readings) to a comprehensive, structured, education programme did not produce greater reductions in HbA1c compared with the education programme alone. HbA1c decreased from 8.8% to 6.9% in the SMBG group, and from 8.6% to 6.9% in the control group; the mean differences were not significant at baseline (SMBG vs control -0.33, 95% Confidence Interval (95%CI) –0.77 to 0.51), at 12 months (0.07, 95%CI –0.25 to 0.38) or at 3, 6 or 9 months follow up. There were no differences between the groups in the incidence of reported hypoglycaemia, use of oral hypoglycaemic drugs or Body Mass Index (BMI). The study did not measure differences in clinical outcomes such as rates of cardiovascular events.
Assessments of the patients well-being found that patients in the SMBG group were significantly more depressed than in the control group. There were no statistically significant differences in any other well-being, treatment satisfaction or diabetes attitude measures.
In non-insulin-treated people with type 2 diabetes in the DiGEM study, Simon et al found that SMBG increased annual net costs of care (which include health professional time, prescription costs, hospital care, etc). Average net annual costs were £89 in the no-monitoring group, £181 in the less intensive SMBG group (an extra cost of £92, 95%CI £80 to £103) and £173 in the more intensive SMBG group (an extra cost of £84, 95%CI £73 to £96). SMBG was also associated with a reduced quality of life.
As we have blogged previously, focusing on controlling blood glucose in type 2 diabetes rather than a range of interventions to reduce cardiovascular risk is not likely to be helpful. Indeed, aiming for very low blood glucose levels can be harmful. Set in context of the rest of the evidence base (see the type 2 diabetes section of NPC), it is increasingly difficult to justify the use of SMBG other than in very specific circumstances. These specific circumstances include patients with type 2 diabetes treated with insulin who adjust their dose on the basis of SMBG results. Pragmatically, it may also be useful to monitor blood glucose during episodes of intercurrent illness, and self-monitoring might also be useful in people who fast for religious or other reasons (eg during Ramadan) if they are at risk of hypoglycaemia (but note not all people are).
Simon J et al. Cost-effectiveness of self-monitoring of blood glucose in patients with non-insulin treated type 2 diabetes: economic evaluation of data from the DiGEM trial. BMJ April 2008 (online first) doi: 10.1136/bmj.39526.674873.BE
O’Kane et al:
Patients:184 newly diagnosed people aged under 70 years (mean around 59 years) referred to participating hospitals in Northern Ireland. It is not possible to assess whether allocation was concealed, but unconcealed allocation would tend to overestimate benefit so this potential bias is probably less important in this study.
Intervention and comparison: Patients were randomised to SMBG (four fasting and four post-prandial capillary blood glucose tests each week) with education on monitoring and response to high or low readings, or no monitoring. Both groups underwent an identical structured, multidisciplinary core education programme.
Outcome: HbA1c decreased from 8.8% to 6.9% in the SMBG group, and from 8.6% to 6.9% in the control group, mean differences not significant at baseline (mean difference SMBG vs control -0.33, 95%CI -0.77 to 0.51), at 12 months (mean difference 0.07, 95%CI -0.25 to 0.38) or at 3, 6 or 9 months follow up. There were no differences between the groups in the incidence of reported hypoglycaemia, use of oral hypoglycaemic drugs or Body Mass Index (BMI). In an analysis of co-variance, participants in the SMBG group scored 6 points higher (that is 6%) on the depression subscale of the well-being questionnaire at 12 months (P=0.011). There were no statistically significant differences in any other well-being, treatment satisfaction or diabetes attitude measures.
Sponsorship: Northern Ireland research and development office. Blood glucose meters provided free of charge by Johnson and Johnson.
Simon et al
Incremental cost-utility analysis of the DiGEM study. SMBG increased annual net costs of care (which include health professional time, prescription costs, hospital care, etc). Average net annual costs were £89 in the no-monitoring group, £181 in the less intensive SMBG group (an extra cost of £92, 95%CI £80 to £103) and £173 in the more intensive SMBG group (an extra cost of £84, 95%CI £73 to £96). SMBG was also associated with a reduced health-related quality of life at 12 months, as assessed by the EuroQoL EQ-5D score, compared with the no-monitoring group. This was statistically significant for the intensive SMBG group only (difference –0.072, 95%CI –0.127 to –0.017). There were indications that the worsening of quality of life with self monitoring was due to increased levels of depression and anxiety.
Sponsorship: UK NHS and NHS health technology assessment programme.