Predicting the impact of primary prevention programmes on the incidence and prevalence of type 2 diabetes in future years has even more difficulties than predicting the future of diabetes in the ‘do nothing’ scenario. The extent to which the potential for success will be realised depends upon a number of factors:

• The extent to which the spectacular results of the explanatory randomized controlled trials (RCTs) can be replicated in pragmatic RCTs and in the ‘real world’ of clinical public health practice.
• The extent to which transition to type 2 diabetes is prevented as opposed to being delayed.
• The extent to which the micro- and, particularly, the macrovascular complications of diabetes can be modified by this prevention or delay.
• The extent to which other factors, as yet unknown, contribute to a change in the epidemiology of the condition.

In the context of the explanatory RCT, transition to type 2 diabetes can be reduced by as much as 50% over three years. The annual rate of progression from IGT to type 2 diabetes in observational studies varies from 3%-13%   1   2  and, in the control arms of the DPP and Finnish Diabetes Prevention Study (DPS) were, respectively, 12% and 6% at the end of the first years, increasing to 28% and 20% at three years. Public health interventions (as distinct from RCT interventions) could not be expected to decrease these rates by as much as 50% but, even if they decreased this conversion rate by 10% or 25% the future maps of type 2 diabetes would be very different.

A more likely scenario is that public health interventions will delay the transition from IGT to type 2 diabetes rather than prevent it entirely. A ‘first pass’ in predicting the extent to which such preventive programmes will change future experience of diabetes and its complications in one country (the United Kingdom) can be provided by modelling a number of ‘what if’ scenarios. The DiabetesForecaster model  3  investigates the impact of delaying the onset of type 2 diabetes on the rate of progression to micro- and macrovascular complications and death, and estimates the impact on direct financial healthcare costs and quality-adjusted life years (QALYs) of delayed onset.

Running the model over a 40-year time horizon for 1,000 people in the absence of any preventive intervention predicts just under 500 coronary heart disease (CHD) events, approximately 450 stroke events and 350 cardiovascular-related deaths. Running the model once more, this time with a delay of 11 years in the transition to type 2 diabetes (as predicted in DPP simulation studies), decreased the number of predicted CHD, stroke and fatal cardiovascular events by 22%, 25% and 20% respectively. Figure 1 shows the effects on total cardiovascular disease mortality of delaying the onset of type 2 diabetes by one, three, five or 10 years compared with no delay.

Figure 1 | Kaplan-Meier survival plots for total cardiovascular mortality*

*Assumes no delay in the onset of diabetes and a one, three, five and 10-year delay in onset.
Source: McEwan et al  3 

The cost savings and QALY improvements relating to macrovascular complications when running the model over this time horizon resulted in a mean, total discounted cost decrease of GBP1,200 per subject from a mean of GBP11,500 per subject assuming no intervention. Discounted QALYS per subject were 12 assuming no intervention, increasing by approximately one for lifestyle intervention (that is, an important increase in expected length and/or quality of life).

Clearly these are modelled estimates and take no account of the cost of introducing these interventions on a community scale. However, they do suggest that the gains in adverse outcomes averted, healthcare resources available for other uses and improvements in the quality of life will be real and worthwhile striving for.