Global Burden of Disease 2010 Study: A personal reflection

[first paragraph of article] The Global Burden of Disease (GBD) 2010 project is a massive effort to systematically describe the epidemiology of a wide array of major diseases, injuries and risk factors. It collates millions of observations and has involved collaboration of nearly 500 scientists from 50 countries. Just to give some idea of the size of the endeavour, the analysis has included 800 million deaths. It has taken 5 years to come to the point where The Lancet was able to publish the headline results, but the database is so incredibly rich that there are likely to be hundreds of subsequent papers using this data source. One can anticipate papers on each risk factor (or groups of risk factors – such as dietary risk factors); each disease or specialty; each region; each country, and then local papers on disease outcomes within each country or region.

increasingly more likely after 55 years of age, but the rate of changes in life expectancy with age are intriguing. Figure 4 shows the decline in deaths by age.
The age-specific decline in deaths is reported by Wang et al. (Figure 4) 1 . It is a little confusing since a higher point on the graph actually reflects better outcomes. Mortality has decreased less among young adults (especially young males) than in other age groups. This might be because this age group have a naturally low rate of death, but it might also reflect the fact that young adult men are primarily those who are at increased risk of death as combatants in war; this age group is also the most likely to engage in risky behaviours -and to die as a result of traffic accidents. 2 Even so, every age group now shows a higher life expectancy than the 1970 estimates.
It is encouraging to see there have been huge strides in reducing infant mortality: deaths among the 1-4 year age group have dropped by 70%. In 1970, deaths in those younger than 5 years accounted for 38% of deaths worldwide; by 2010 this was down to 13%. Whether or not it is causal, it seems that the Millennium Development Goals regarding infant mortality are well on the way to being successfully met. The job is not complete since deaths in the first 4 years of life are still greater than any other age group except over 80s ( Figure 3) but now might be a time to begin planning for the next focus of priorities.
Over the past 20 years global population life expectancy has increased in almost all countries. GBD 2010 also presents the changes in deaths by geographic region. 1 Much more detailed analysis can be found at www.healthmetricsandevaluation.org/gbd/visualizations. All regions of the world have seen an increase in life expectancy since 1990 as demonstrated in Figure 5 where all points are below the line. Life expectancy is longest in high-income countries, and shortest in Africa. But Africa has made massive strides. In Western Europe life expectancy has increased by 8 years (from about 68 years to 76 years), but in North Africa and the Middle East, for example, life expectancy has more than doubled from 22 years in 1970 to 52 years in 2010. An extra 30 years of life. Life expectancy in this region is, however, still relatively short.
What is striking is that with the exception of HIV/AIDS, the ranking of deaths due to infections have decreased substantially. By contrast non-communicable diseases have become more important causes of death. The lines between the ranking tell the story of relative importance, but one needs to look beyond the lines to get a better idea of change. Although malaria has declined in relative    importance from 9 th to 11 th , this masks a 21% increase in absolute numbers of deaths as shown in the right hand column. Overall, however, the picture of change generally reflects absolute changes. Ischaemic heart disease and stroke remain the two biggest causes of death. Of non-communicable diseases, only stomach cancer, rheumatic heart disease and congenital anomalies have declined in importance. Perhaps emphasis on research and policy should now start to focus more on non-communicable diseases. The GBD database is a massive resource and the papers published in December 2012 are headline summaries. It can be expected that there will be many future papers looking at specific diseases and regions. One can expect at least one detailed paper on cardiovascular disease, but the headline data provide some interesting insights, such as the years living with disability due to heart failure around the world. 5 These data might be of interest to physicians and health policy-makers in specific regions and can be useful to prioritise policy in particular regions. It can be seen in Figure  deficiency is largely confined to the North Africa and Middle East. Perhaps this might stimulate more investment in iodised salt in North Africa and Middle East countries. It is also apparent that endocarditis is more prevalent in Oceania than elsewhere, and this might inform funding for endocarditis research in that region. The headline analysis also provides insights into subspecialty causes of deaths. 2 A global picture of causes of cardiovascular deaths is in Table 1. The number of cardiovascular deaths has increased 31% from 12 million to 16 million deaths between 1990 and 2010, 2 but of course this needs to be considered in the context of an increase in the global population. In that perspective, the 25% reduction in the number of deaths due to rheumatic heart disease is notable. The absolute numbers of deaths due to most other cardiovascular conditions have increased, but so has the mean age of the population. Diseases such as atrial fibrillation which are more prevalent in the elderly, are therefore expected to be more common as the population ages. For most of the cardiovascular diseases in Table 1 age-standardised mortality rate has decreased, but atrial fibrillation shows an 89% age-standardised increase. The other condition with an increased age-standardised mortality is peripheral vascular disease. This probably reflects the increase in diabetes associated with changing life styles and increased age.
GBD 2010 doesn't only report deaths. It also reports the burden of disease in terms of years lived with disabilities, 5 disability adjusted life years 6 and risk factors for disease and injury. 7 The years of life lost (YLL) analysis shows a slightly different pattern. Ischaemic heart disease and stroke are still near the top of the list, 2 but Figure 9 reflects the effect of disease early in life on total years lost. Serious diseases of infancy and childhood achieve a greater prominence when analysed by YLL. In this analysis   infectious diseases are more important causes of YLL but they have all (except HIV) also decreased in number -presumably as a result of the success of immunisation programmes. There remain too many people in the world who are hungry. It is encouraging that deaths due to malnutrition have fallen by 32%, but it still ranks as 21 st in the list. And there is also a decline in ranking by YLL but, because malnutrition in childhood has long term effects, it remains as the 15 th commonest     cause of years lost (Figure 9). The decrease in malnutrition probably reflects major advances in agriculture as well as medicine. Although less headline-grabbing than deaths, DALYs are particularly important to inform government health policies. The causes of DALYs by regions 6 are interesting as a reflection of events (both natural and man-made). It is well known that there has been an explosion of HIV in Southern subSaharan Africa and this is clearly demonstrated in Figure 10. The effect of the Haitian earthquake is also reflected in the war and disaster statistics in the Caribbean region. And in good news, the DALYs due to diarrhoea and respiratory infections have declined substantially in central and southern America and in south east and central Asia. The data presented here are regional estimates, but national estimates (available at www.healthmetricsandevaluation.org/gbd/country-profiles/) can help governments decide which diseases should be their priority for treatment and prevention. In terms of reducing incidence of disease, the risk factors data provides information that needs careful consideration that can have an impact on government decisions. Some of the risk factors for disease are not a surprise: smoking causes cancer; high glucose is associated with diabetes. The change in relative importance of risk factors between 1990 and 2010 can be found in the paper by Lim et al. 7 Risk factors are especially important for the predicting the future global burden of diseases. Identifying the most important risk factors to reduce DALYs can help decide on health policy strategies. For example, is money best spent on anti-smoking campaigns, on promoting exercise, or reducing pollution? Figure 11 shows the analysis for 2010 and this can be a starting point for policy changes that will influence future health outcomes.
A careful look at Figure 11 provides some interesting insights into how policies to address health issues might sometimes promote solutions which might not necessarily be those that will have the biggest effect on health. The analysis in Figure 11 provides an example of where perhaps these analyses might influence future priorities. It is no surprise that tobacco is a prominent cause of DALYs, and around the world Governments are working to reduce tobacco consumption. It might be less obvious that household air pollution is so important. These data might stimulate promotion of cleaner Ascending order in rank Descending order in rank household fuels. However, there are some anomalies. High blood pressure is identified as the major burden of DALYs. To counteract this there are moves to reduce salt intake, but this might not be the most effective way to promote health since a diet high in sodium is way down the list of risk factors. That discrepancy supports the fact that, despite association studies linking salt intake and hypertension, direct evidence that high salt intake increases blood pressure in non-hypertensive people is very weak. Perhaps focusing health campaigns on other factors would be more beneficial.
Much further up the list of important risk factors is a diet low in fruit. It is known that fruit is beneficial in avoiding major killers like heart disease and cancer, 8 but this gets little positive action from governments. At the very least, the risk factors identified in GBD 2010 should encourage a critical scientific re-evaluation of the hard evidence underpinning policies on risk factors. It might be fashionable (and easy) to blame the fast food industry for obesity or the soft drinks industry for diabetes, but it seems that increasing fruit intake would be a more important initiative for improving health. This paper indicates some of the ways in which the Global Burden of Disease 2010 study can be of interest and value, but it merely scratches the surface. The main analyses of GBD 2010 are  published in The Lancet 1 -7 and regional, country and specialty specific analysis are likely to follow. Many more GBD 2010 analyses are available at www.healthmetricsandevaluation.org/ gbd/visualizations. In conclusion, GBD 2010 can be important in framing health policies and objectives both globally and in specific countries. The first country specific analysis from the GDB team was an assessment of the UK. 9 The UK government was represented at the launch of the GBD 2010 presentation in London in 2012 and recognised the findings should inform policy. In March 2013 the Department of Health in the UK published a call to action. 10 It would be surprising if other governments did not follow this lead, and more country-specific analyses are expected. In depth analysis of the data will also help refine attitudes to risk factors and influence research priorities in specific regions and countries. As NCDs are the major cause of death and disability perhaps it is time to reconsider the priorities for research and policy initiatives away from the current HIV, tuberculosis and malaria programmes. And now might be the time to start considering whether more emphasis might be placed on death and disease in adults now infant mortality has been so substantially reduced. Analysis by medical specialty will also shape research priorities and research programmes in specific regions. Finally, the GBD 2010 has highlighted weaknesses in data collection in some geographic and specialty areas and moves have already begun to improve the databases to make future assessments more rigorous.

Conflicts of interest
SS is a full time editor at The Lancet but has no other relations with any aspect of the Global Burden of Disease Study, and no financial interests.