Using sophisticated modelling techniques the results of this study illustrated how the obesity epidemic will unfold across Latin America. Over the next twenty years overweight and obesity was projected to increase. Reflecting these trends the incidence of each disease is also set to increase. Interventions that are effective in reducing BMI will be important in reducing rates of cardiovascular disease and diabetes.
In general, overweight and obesity levels were projected to increase in all countries with the highest rates seen in Cuba. Rates of change are likely to be a result, at least in some part, to economic changes. The positive effect of the economic transition has been to help eradicate undernutrition, but, unregulated, it also promotes unhealthy lifestyles which favour obesity. Traditional diets in the region are often meat-based, in particular red meats are popular staples and food preparation involving frying is also popular. When food security and incomes increase, it is not surprising that the per-capita consumption of these traditional foods increase, together with other Western influences on diets. Chile had the best available data of any Latin American country allowing for more accurate predictions than other countries. In the 1990s Chile doubled its per capita income. An important amount of this increase has been spent on modern living such as television sets, cars and unhealthy high fat, sugar and salted processed foods fuelling energy imbalance and subsequent increases in obesity 
. Similar transition has occurred across Latin America.
Cuba is an interesting case due to the macroeconomic changes and their consequences on health behaviour and obesity. Since the 1950s the economy grew at a rate higher than the rest of Latin America despite the US embargo. The collapse of the Soviet Union in the 1990s cut Cuban trade by 80% and GDP decreased by 33% 
resulting in food shortages, increased physical activity and reduced BMI 
. However, the budget for education and health increased during this period. Since then, availability of high fat foods has increased, whereas labour-intensive activity has reduced, which may be one contributing factor of the plethora of causes explaining the obesity epidemic and the high levels of obesity and cardiovascular diseases observed in the present study.
Cumulatively, from 2010 to 2030 the incidence rate of CHD & Stroke in Cuba would rise to a staggering 15022 per 100,000 of the population, diabetes would rise to 11624, and cancer to 2523 per 100,000 of the population. Cuba has a good quality health system relative to the rest of Latin America. That the current estimates of Cuba's CHD rates are higher than elsewhere in Latin America might be a result of higher survival rates due to better quality health care.
As well as economic changes, socio-economic and demographic differences between and within countries are likely to impact the rate of obesity. Obesity is shifting to be a disease of the poor, as it is in most high-income countries 
which underscores obesity as a social phenomenon highlighting the need to take action on sociocultural and economic factors. This social shift in obesity usually happens first in urban women perhaps because of differences in working patterns between men and women with men being in more traditionally manual labour roles. A study by PAHO/WHO 
on obesity in Latin America found that a higher prevalence of obesity is correlated with per capita income especially in urban areas. Despite this, under nutrition is still a major problem and it is increasingly apparent that Latin America is experiencing a double burden of disease where both underweight and obesity coexist.
Interestingly, more urban countries show a higher rate of disease than less urban countries. For example, in highly urbanised Chile the cumulative incidence rate of CHD and stroke by 2030 is around 8100 per 100,000 of the population in 2010 
. In Nicaragua, where urbanisation has reached 58%, the projected cumulative incidence (for women) is 5400 by 2050 per 100,000 of the population in 2010. There is also a huge ethnic diversity within and between countries in Latin America. Interestingly, nations with the highest white or European ethnicities and that are highly urbanised (e.g. Uruguay, Chile) have the highest projected prevalence rate of cardiovascular diseases compared with countries such as Bolivia, Colombia and Nicaragua where the ethnic makeup is mostly Amerindian and/or Mestizo. However, in Bolivia and Nicaragua we were only exploring women who have a lower risk of CVD than men especially in the younger age groups and these populations are still quite young. Sampling both men and women will allow for more accurate comparisons to be made.
The results of this study have important policy implications. Given the high social and economic cost of NCDs, further work into the health economics of obesity in Latin America is necessary so that future health policy can be planned for. In 2000, diabetes was estimated at US$65.2 billion across Latin America 
and a recent review reported that obesity accounted for 0.7–2.8% of a country's total health care costs and medical costs were 30% higher for obese than normal weights 
. Thus the problem of obesity poses an enormous challenge and institutionally Latin America needs to be equipped to deal with increasing numbers of chronic diseases. Some countries have responded to the obesity challenge by implementing interventions. Chile introduced nutrition and physical activity initiatives to reduce obesity in preschool children though this was not enough to shift the rising rate of obesity. Although, it was argued that the intervention's lack of success is perhaps because obesity rates have reached a plateau 
. To address the problem of sedentary lifestyles, Colombia has launched a free bike scheme and bicycle lanes in the capital Bogota which has since been named the worlds 3rd
most bike-friendly city. The Caribbean Public Health Agency (CARPHA) 
have recently set non-communicable diseases as a key public health priority (2011) and in early 2011 more than 40 Latin American organisations launched The Healthy Latin American Coalition (HLAC) to develop a declaration recognising the public health emergency of NCDs and the importance of government action. Clearly, if trends are set to continue rising more work is required.
Very little data were available for Latin American countries making analysis of obesity trends in this area limited and it difficult to draw affirmative conclusions. For Bolivia, Nicaragua and Peru data for females only was available. These data were from Demographic Health Surveys which only measure women who had had a child in the past five years, thus biasing the data. The direction of the bias is unclear, and might differ according to age, but higher rates of overweight might be expected since mothers often do not return to pre-pregnancy weight. For Costa Rica, Cuba and Panama only one data point was available and so 2008 estimates were used based on a recent analysis by Finucane and colleagues 
which used BMI means. This is disadvantageous since one cannot then reliably infer the proportion of obese to overweight. However the estimates are of use when looking at the proportion of normal weight to overweight and obese combined. This highlights the need for greater surveillance work across Latin America which samples both men and women in nationally representative samples. This is imperative if accurate estimates of trends are to be made and policies to be built around more precise data.
The projections can only be as good as the data that is input. Our extensive searches found very little data were available for Latin America and no set of complete age and sex-specific BMI and disease data for one Latin American country. We were also unable to include data on children due to lack of consistently measured data. Since projections are mere extrapolations from these data, inaccuracies in the output are likely. However, there was insufficient time to undertake time consuming error analysis. Furthermore, we have insufficient knowledge of BMI growth patterns following interventions and insufficient knowledge of the future.
Our model incorporates a sophisticated economic module using Morkov-type simulation estimation of long-term health benefits, health care costs and the cost-effectiveness of specified interventions. With access to country-specific cost data our model can be adapted to include cost burden and allow us to simulate costs of obesity-related diseases for application in public health policy. Our recent work projected future health and related medical costs based on available disease data in the UK and US 
allowing for more accurate projections to be forecast, however these data were not available for Latin America and so could not be included in the present study. Other diseases beyond those studied here have been related to obesity such as infertility 
, sleep apnoea 
, osteoarthritis 
, asthma 
. It was beyond the scope of this study to include them. The flexibility of our model means that there is scope to model these diseases given the right input data are made available.
The programme is limited in that it assumes that people do not reverse in the BMI categories. Unfortunately this mirrors reality, where body weight loss is often only temporary. Moreover, the analysis has not taken into account unforeseeable changes in circumstances, such as fluctuations in food prices and changes in medicine. It relies on our best estimate based on previous trends.
The 95% confidence intervals for the microsimulation were derived from simulation of the BMI distributions corresponding to the upper and lower limits of each of the obesity growth scenarios. Unfortunately there was insufficient time to undertake time consuming error analysis and this has been noted in the limitation of the paper. The results do not vary significantly when different simulations are run.
Despite some limitations, this study is timely and an important first step in quantifying the future burden of obesity-related diseases in Latin America. It highlights the need for urgent action to curb obesity levels and reduce the burden of disease. The challenge is to understand how best to initiate change and to quantify the cost of health consequences of obesity. If governments take action by implementing effective policies that reduce overweight and obesity, then a substantial number of new cases of cancer and cardiovascular diseases can be avoided in the coming decades.