Research Paper of the Year 2017

Dr Eleanor Barry, Winner Research Paper of the Year 2018

GPs can see from their everyday clinics that type 2 diabetes is rising and that some people are at high risk of developing it. We know that both lifestyle change, for example, eating less and exercising more, and some drugs, notably metformin, can delay or prevent type 2 diabetes in some people.  

These general principles raise more specific questions. How exactly should we identify people whose blood glucose levels put them at high risk of developing type 2 diabetes, a state increasingly referred to as ‘pre-diabetes’? Which blood test for example, fasting glucose (FPG), HbA1c or oral glucose tolerance test (OGTT), and which diagnostic criteria should we use? The World Health Organisation and the American Diabetes Association (ADA), for example, disagree on cut-off values.  In addition, what interventions should we offer those diagnosed with pre-diabetes – and what is the chance that it will prevent diabetes?  ‘Lifestyle’ interventions cover numerous approaches at varying degrees of intensity. 

To answer these questions, we conducted two different systematic reviews and meta-analyses of previously published studies, using statistical techniques to draw together the findings from different primary studies. The first meta-analysis assessed the diagnostic accuracy of HbA1c and FPG in identifying pre-diabetes as defined by the gold standard (but expensive and cumbersome) OGTT. The second meta-analysis assessed the effectiveness of intervention trials in reducing participants’ relative risk of developing type 2 diabetes. One key sub-question was which individuals (and what proportion of the population) would count as ‘pre-diabetic’ with each of the three tests.  

We found that neither the HbA1c nor FPG was particularly good at picking up OGTT defined pre-diabetes. Compared to the gold-standard, the sensitivity of these two tests was 49 percent and 25 percent respectively, meaning that 49 percent and 75 percent of all people with abnormal OGTT’s would be missed. Furthermore, each test and different diagnostic thresholds identified a different sub-population of people at risk of type 2 diabetes, with large variations in the size of the at-risk population depending on the test and threshold used. Using the ADA definition of pre-diabetes, 50 percent of the adult population would be identified as ‘pre-diabetic’. 

Our meta-analysis of lifestyle intervention trials showed that people identified as pre-diabetic (using the OGTT) could lower their risk of developing type 2 diabetes. For interventions lasting between six months and two years, for example, the relative risk reduced by 31 percent (95% CI 15-44%, NNT 33); longer and more intensive interventions, some of which were prohibitively expensive had higher relative risk reductions. However, follow-up studies suggested that risk reductions achieved attenuated over time. Only half the trials reviewed used type 2 diabetes development as the end point; the rest used surrogate markers such as weight loss, which although positive, does not correlate closely with diabetes development. 
Our review has highlighted a maxim first proposed by Geoffrey Rose: that in all prevention programmes, there is a trade-off between looking for ‘sick individuals’ (with a view to providing individual preventive interventions) and addressing the entire ‘sick population’ (using population measures to tackle upstream determinants of ill health such as the obesogenic environment)  

The National Diabetes Prevention Programme offers an individual ‘screen and treat’ approach to identifying and managing pre-diabetes available to 70 percent of the UK population. With the rising epidemic of type 2 diabetes, the individual approach must be supplemented by population measures, especially in areas of high diabetes prevalence (notably socio-economically deprived and ethnically diverse settings).2

The take-home messages for GPs are these: 

  • The patient whose HbA1c or fasting glucose test comes back in the normal range should not be reassured that they are not at risk of diabetes, especially if risk factors exist within the individual. Whilst the oral glucose tolerance test is too cumbersome to be used in population-wide screening, it remains the gold standard for exploring the glycaemic grey zone between ‘normal’ and ‘diabetic’ – and we should interpret all other tests with caution. 
  • Lifestyle intervention programmes to prevent type 2 diabetes may help individuals reduce their risk of diabetes development– with the caveats that patients will need to fully engage in such programmes. 
  • Clinical Commissioning Groups and Public Health departments need to collaborate to address the community level influences on health to supplement individualist prevention strategies.

Barry E, Roberts S, Oke J, Vijayaraghavan S, Normansell R, Greenhalgh T. Efficacy and effectiveness of screen and treat policies in prevention of type 2 diabetes: systematic review and meta-analysis of screening tests and interventions. BMJ. 2017;356:i6538.

Funding: This study was funded by grants from the Newham Clinical Commissioning Group and University College London Partners, a National Institute for Health Research fellowship for EB, National Institute for Health Research senior investigator award for TG, and by internal funding for staff time from the Nuffield Department of Primary Care Health Sciences, University of Oxford.


1. Rose G. Sick individuals and sick populations. International Journal of Epidemiology. 2001;30(3):427-32

2. Swinburn BA, Sacks G, Hall KD, McPherson K, Finegood DT, Moodie ML, et al. The global obesity pandemic: shaped by global drivers and local environments. The Lancet. 2011;378(9793):804-14.

Find courses & events

The item has been added to your basket.

Continue shopping

Go to basket

This item is out of stock.

Continue shopping

The item is out of stock.

Yes Continue shopping

An error occurred adding your item to the basket:

Continue shopping