Other interventions that appear cost-effective in developed countries according to economic studies include lifestyle improvement and possibly metformin to prevent type 2 diabetes, smoking cessation programmes, annual eye examinations to detect retinal disease, ACE-inhibitor use in all persons with diabetes, and influenza vaccines in elderly persons living with type 2 diabetes . The cost-effectiveness of low-dose aspirin to prevent cardiovascular disease in all persons with diabetes has not been formally assessed. However, aspirin is cost-effective in persons with pre-existing cardiovascular disease   8  and risk-beneficial in others when, as in diabetes, cardiovascular risk is elevated   9   10 . Given the very low price of aspirin, its use is probably cost-effective or cost-saving for everyone with diabetes   11 .

The cost-effectiveness of screening for type 2 diabetes is less clear. The CDC group calculated that, in persons without risk factors for diabetes, screening for diabetes, at least in the USA, may not be cost-effective   12 . A European modelling study recently concluded that the cost-effectiveness of diabetes screening depends on the presumed benefit of treatment  20 . A simulation study in Taiwan concluded that mass screening for type 2 diabetes would be cost-effective there   13 .

Cost-effectiveness in developing countries
Developing countries cannot justify paying for all the treatments that high-income countries use. Still, many low-cost treatments should be cost-effective or cost-saving everywhere, as the IDF Task Force on Diabetes Health Economics argued in its publication, Cost-effective approaches to diabetes care and prevention   14 .

In 2006, the World Bank systematically assessed the cost-effectiveness and feasibility of diabetes interventions in developing countries . Their table of findings by region is reproduced here as Table 1. The World Bank classified cost-effective interventions into three levels of implementation priority, based on cost per quality-adjusted life year (QALY, its measure of cost-effectiveness), technical feasibility, and cultural feasibility.

Table 1 | Cost-effectiveness of interventions for preventing and treating diabetes and its complications in developing regions

Level 1 interventions were predicted to be cost-saving and highly feasible in all regions. Three interventions achieved level 1: glycaemic control when HbA1c is higher than 9.0%, hypertension control when blood pressure exceeds 160/95 mmHg, and foot care for persons at high risk of ulcers. (As explained above, low-dose aspirin also might be added to this list.) Level 2 interventions (N=6) are usually highly cost-effective but face some barriers to implementation. Level 3 (N=5) priorities all cost more than USD1,500 per QALY saved and were judged challenging to implement.

Although the World Bank assessment was not based on actual effectiveness trials in developing countries — none have been conducted — their work leaves little doubt that many diabetes interventions could and should be pursued aggressively throughout the world. A recent high-level modelling analysis suggests that large gains for the entire economy, as well as for individuals with diabetes, would result from recommendations like the World Bank’s, whether carried out in Bangladesh or Denmark   15 .

Researchers are now debating and testing the use of inexpensive and convenient ‘polypills’ that combine in a single, once-a-day treatment, several compounds that already have been judged cost-effective or risk-effective   16 . This approach could increase the feasibility of diabetes treatment in developing countries, and make it even more cost-saving.

Finally, it is worth noting that although ‘Level 2’ activities like smoking cessation, increased physical activity, and more prudent eating face administrative costs and implementation barriers when done as social programmes; but individuals in developing countries who decide on their own to stop smoking or increase their walking can do so very cheaply.

Efficiency and equity
Analyses of cost-effectiveness usually assess the impact of new programmes and policies under the assumption that they are implemented as designed, with little waste or corruption. Implicitly, analysts also assume that access will be based on the ability to benefit, with no discrimination by class, race, or tribe. These assumptions do not always hold in practice. Indeed, it has been estimated that if only 50% of persons with diabetes are diagnosed, and 50% of these receive care, and 50% of these achieve treatment targets, and if treatment prevents complications 50% of the time, only 6% of persons with diabetes will have a successful outcome   15 . Increasing the efficiency of diabetes care can yield large improvements. If the reach of intervention were increased to 75% in the calculation just described, the percentage of persons with successful outcomes would be five times larger.

Many existing policies are ill-designed to operate efficiently and equitably. In the Northwest Frontier Province of Pakistan, for instance, the government pays for insulin only when citizens are hospitalized. This causes multiple costly hospitalizations — patients become desperately sick when they lack insulin — and leads to severe complications and premature loss of life. In Pakistan, it would probably be cost-effective for government to buy insulin for all persons with diabetes   17 .

Finally, it is important to recognize that, just as diabetes harms economies, economies create diabetes. Policies towards agriculture, transportation, housing and land drive patterns of eating and exercise. Perhaps even more importantly, countries with less egalitarian distributions of income, wealth and power also have higher rates of insulin resistance, diabetes and cardiovascular disease. Much research now indicates that this relationship is causal, and that it operates in high- as well as low-income countries, independently of average levels of resources. Feelings of shame, vulnerability, powerlessness and uncertainty trigger many of the same physiologic cascades that transmit the ill-effects of genes, obesity and lack of exercise   18   19 .