Screening for and prevention of type 2 diabetes

Abstract

Intervention should be sooner rather than later, even though exact costs and benefits are uncertain

Cost effectiveness models are widely used to help policy makers and clinicians make decisions about alternative interventions. Models have limitations that are well understood—a model is only as reliable, or generalisable, as the assumptions and data that inform it. Moreover, a cost effectiveness model will not deal with questions about feasibility, acceptability, or affordability that may be crucial to decisions about implementation. However, decision analysis models often help to clarify the key factors likely to influence cost effectiveness, so that decision makers can better understand the importance of remaining uncertainties and make more logical decisions in the face of uncertainty.^{1 2}

In the accompanying paper, Gillies and colleagues report a cost effectiveness analysis for screening and prevention of type 2 diabetes.³ They compare four strategies—screening for diabetes; screening for diabetes and impaired glucose tolerance, followed by either lifestyle interventions or drugs; and no screening. Their findings are consistent with the intuitively logical view that, given the strong evidence base for prevention of diabetes in people with impaired glucose tolerance, it is more cost effective to intervene early, rather than to screen but then ignore impaired glucose regulation until it is severe enough for a diagnosis of diabetes.

This evidence for early preventive interventions comes from the results of trials of interventions for diabetes prevention that have proved highly effective⁴ despite data that show how difficult it is to change behaviour,⁵ particularly in an obesogenic environment.⁶ Prevention is cost effective in part because we do not yet have effective interventions to prevent complications in screen detected or clinically diagnosed diabetes. This is highlighted by models that make more optimistic assumptions about progression and complication rates in screen detected diabetes, which then make preventive interventions seem less cost effective.⁷

A new Department of Health policy initiative for England is the promotion of systematic assessment of cardiovascular risk from the age of 40 to 75, including testing for diabetes and impaired glucose tolerance in high risk groups.⁸ Given this new policy and the substantial evidence base, why are clinicians not more enthusiastic about primary care screening programmes for impaired glucose tolerance and diabetes?

The reasons include remaining uncertainties about the real costs and benefits of screening and practical considerations about feasibility and affordability. The potential to do more harm than good through screening is still real, even when the screening test is relatively risk free. Most important is the potential risk of false reassurance for people with risk factors for diabetes but a negative test result, most of whom still have a high risk of diabetes and cardiovascular disease and would benefit from lifestyle changes.

Uncertainty also exists about key model parameters. Although we have effective drugs for coexisting cardiovascular risk factors (particularly hypertension and hyperlipidaemia), the effect of earlier treatment of hyperglycaemia on long term outcomes is still unclear. Similarly, the biggest uncertainty in relation to drug treatment for impaired glucose tolerance is whether the reduction in blood glucose will translate into benefits that matter to patients. This is still a powerful argument for remaining sceptical about the benefits of drug treatment for impaired glucose tolerance.⁹

There are also at least two major practical barriers to screening. Firstly, access to oral glucose tolerance tests is limited. To achieve the high sensitivity values used in Gillies and colleagues’ model, a high proportion of people at risk will need an oral glucose tolerance test. In many countries, including the United Kingdom, and even in well resourced screening pilots, access to these tests is often limited.¹⁰ Organising appointments for tests, performing tests, and getting valid results from fasting blood tests and oral glucose tolerance tests may require additional training for staff and improvements in arrangements for collection and analysis of samples.

Secondly, effective and affordable lifestyle interventions are still lacking. Most of the intensive interventions used in diabetes prevention trials are unavailable or unaffordable in everyday practice. Several major research programmes to develop and evaluate feasible and affordable interventions in different settings are under way, but most of these are still to report, and the long term effects will take years to evaluate.

So what should we do in the meantime? Given current trends in the prevalence of diabetes, inaction is not an option.¹¹ Instead, we can try to ensure that all patients with risk factors for diabetes receive advice and encouragement to reduce their risk through dietary change and increased physical activity. If oral glucose tolerance tests are used as a screening or diagnostic test for diabetes, we should be proactive when impaired glucose tolerance is detected. Advice and support should be offered, and the patient should be referred for supported lifestyle change, if available. Such a policy will probably be more effective and more cost effective than waiting to intervene further down the line.