Population Approaches to Prevention of Type 2 Diabetes

Martin White argues that whole population interventions will be needed in addition to those targeted to people at high risk in order to respond to the global challenge of type 2 diabetes.

individual level, and their delivery in a societal context, population interventions are generally considered safe [9].
There are potential concerns also about the differential effectiveness of some population interventions. Not all population-level interventions are likely to have the same impact, and some may be less equitable than others. This is a particular concern for interventions that require a higher level of engagement by individuals, such as food labelling, which demands literacy and numeracy as well as an ability to process and apply the information presented in making healthy food choices, often from a bewildering array of available products [12]. Interventions requiring low levels of individual engagement, such as regulation of TV advertising of unhealthy foods, may be more equitable and offer greater overall health impacts. However, such policy interventions need to be formulated to ensure population benefit and be rigorously evaluated [13].
Simulation studies have modelled the potential impacts on T2DM incidence of small changes in risk factors for T2DM, such as physical activity at a population level, demonstrating significant potential benefits [14]. The challenge is to develop and deliver interventions that can bring about such small changes across the whole population cost-effectively. Is this achievable and, if so, how?
Population level interventions can be delivered via a range of modalities. Each is underpinned by a policy measure (e.g., a law or voluntary agreement) and involves a mechanism to achieve change in risk exposure. The mechanisms are diverse and include, for example, reformulation of foods (e.g., to reduce sugar content), provision of information (e.g., food labelling), fiscal measures (e.g., taxes on less healthy food products), or structural and environmental measures (e.g., new infrastructure for active commuting, such as cycle lanes). These can usefully be categorized as technology, education, or resource-based and sometimes involve combinations of these modalities (Table 1).
Many such interventions have been implemented with apparent success [15][16][17]. While the primary aim has usually not been T2DM prevention specifically, their contributions to this goal may be considerable. For example, an evaluation of a new guided busway in Cambridgeshire, United Kingdom, found that it resulted in an increase in active commuting; participants living 4 km from the busway were one-third more likely to have increased their cycle Subsidies for physical activity at local leisure centers commuting between 2009 and 2012 than those living 9 km away, reporting a mean increase of 87 minutes of cycling per week [18]. Similarly, early evaluation of the tax on sugar-sweetened beverages in Mexico suggests a decrease in purchases of taxed beverages and an increase in purchases of untaxed beverages [19]. Some evidence suggests that exposure to favorable diet and physical activity environments, such as healthy food provision and physical activity resources at neighborhood level, may have measureable effects on T2DM incidence [20]. In these intervention examples, it is possible that planned environmental changes will impact total activity level, or total sugar and energy consumption, with potentially important consequences for body mass, glucose regulation, and associated health outcomes at a population level.
Evaluating the impact of population-level preventive interventions on health outcomes is a challenging long-term goal that needs investment in routine data systems, data linkage, and new paradigms of experimental research that are able to capture impacts across a range of potential outcomes at a system level, taking into account their contextual complexity. The challenges of such evaluations have been helped considerably by methodological developments in recent years, which have been given prominence by guidance on evaluation of natural experiments from the UK Medical Research Council [21]. Nevertheless, evaluating population interventions is likely to continue to prove challenging, especially in countries with poorly developed routine heath information systems.
While rigorous science is needed to build the evidence base for population interventions, this alone will not be sufficient to ensure that the range of interventions needed to reshape population risk of T2DM is delivered by countries and regions. Such interventions are delivered in real world contexts by policy makers at local and national levels outside the control of researchers. Scientific evidence is just one of a number of inputs to the policy-making process, and political and economic considerations are usually of equal or greater importance [22,23]. Research thus needs to focus on these too, identifying, in particular, the economic case for such interventions, not just in terms of cost-savings to health systems, but also in terms of the wider benefits to society from reduced morbidity, including improved employment prospects and economic productivity.
The global challenges of obesity and T2DM are inextricably linked, and their scale and economic impact demand global action from political leaders. Interventions need to include both those targeted to people at high risk (to address the existing burden of prediabetes) and those aimed at whole populations (to reverse current upward trends). Solutions will need to be sensitive to context in low-income, middle-income, and high-income countries. Population interventions may prove the most viable options in resource-poor nations [7]. Bold policy decisions will need to be made to avert the economic and health crises that are the likely consequences of current trends in obesity and T2DM, since we cannot wait for perfect evidence [24].

Author Contributions
Wrote the first draft of the manuscript: MW. Contributed to the writing of the manuscript: MW. Agree with the manuscript's results and conclusions: MW. ICMJE criteria for authorship read and met: MW.