Global malaria mortality between 1980 and 2010: a systematic analysis
Published in The Lancet, February 2012
Malaria caused over 1.2 million deaths worldwide in 2010, twice the number found in the most recent comprehensive study of the disease, according to researchers at IHME and the University of Queensland. This research, published in the study “Global malaria mortality between 1980 and 2010: a systematic analysis,” shows that while malaria is traditionally considered a childhood disease, there is a significant disease burden in adults.
During the past decade, attention to and funding for combatting malaria have greatly increased. Development assistance for malaria increased from $149 million in 2000 to almost $1.2 billion in 2008, which led to a rapid scale-up of malaria control in Africa.
Many efforts have been made to assess the burden of malaria and progress in fighting the disease, with different approaches leading to highly variable results. Global malaria death estimates since 2000 range from 800,000 to more than 1 million.
To aid assessment of progress toward development goals for malaria and to better focus future prevention efforts, an accurate assessment of the levels and time trends in malaria mortality by age, sex, and country is needed.
This work is part of IHME’s research for the Global Burden of Diseases, Injuries, and Risk Factors 2010 Study, which will produce new estimates measuring the impact of hundreds of diseases, injuries, and risk factors in 21 regions around the world over two decades. The study is being carried out in collaboration with more than 800 researchers and includes more than 220 conditions and injuries and more than 40 risk factors.
Researchers found that global malaria deaths increased from 995,000 in 1980 to a peak of 1.8 million in 2004. In 2010, there were 1.2 million malaria deaths, a 32% decrease since 2004. These results are largely driven by the pattern seen in sub-Saharan Africa, where deaths increased from 493,000 in 1980 to 1.6 million in 2004 and then decreased by about 30% to 1.1 million in 2010. Outside of Africa, the trend is much different, with deaths steadily decreasing from 502,000 in 1980 to 104,000 in 2010.
Although most malaria deaths are in children, the number of deaths in adults is high. In 2010, 20% of malaria deaths were in people aged 15 to 49 years, 9% were in people aged 50 to 69 years, and 6% were in those over 70 years of age. Compared to other assessments, this study estimates more deaths in individuals aged 5 years or older in 2010: 435,000 in Africa and 89,000 outside of Africa.
The risk of dying from malaria in 2010 is highest in western, eastern, and central sub-Saharan Africa. However, the risk in several countries in these regions that have scaled up malaria control efforts, including Zambia, Tanzania, Kenya, and Ethiopia, has decreased between 2000 and 2010.
In order to predict levels and trends over time in malaria mortality, the authors used an approach developed and applied for other causes of mortality, including breast and cervical cancer. For 105 countries with data on malaria transmission between 1980 and 2010, they identified all data for deaths due to malaria, correcting for known biases in the data, including misclassification of deaths to causes other than malaria. Data were obtained from countries’ vital registration systems and verbal autopsy studies. In a verbal autopsy, researchers interview the relatives of someone who has recently died to identify the cause of death.
Researchers studied key predictors of malaria mortality, such as prevalence of Plasmodium falciparum parasites, resistance to first-line antimalarial drugs, and vector control.
Many different models for analyzing the data were tested, including ensemble models, which are weighted averages of individual component models. To choose the final model, the researchers used out-of-sample predictive validity, which is tested by running a model with some of the data removed, and then checking the performance of the model at predicting the data that were removed.
The authors note there are differences between these results and those of other assessments, including the World Malaria Report 2011. The differences can be attributed to the larger number of malaria deaths included in this analysis, as well as child mortality estimates developed with an analysis that suggests fewer deaths from all causes than those used in the World Malaria Report. In contrast to other assessments, this study also took into account estimates of Plasmodium falciparum parasite rate from the Malaria Atlas Project, included the effect of interventions other than vector control, and developed models with rigorous out-of-sample predictive validity.
While substantial progress has been made in the fight against malaria over the past five years, these findings show substantially more deaths across all ages and regions than other assessments, especially in adults.
Traditionally, medical and public health schools teach that adults in countries with malaria develop immunity as children and are not likely to die from the disease. However, these results clearly show that a substantial percentage of malaria deaths occur in people aged 15 years and older, even in areas such as sub-Saharan Africa where adults are likely to be exposed to malaria as children.
The fact that malaria is a significant factor in adult mortality indicates that control strategies should shift to pay more attention to adults and underscores the dangers posed by the global economic crisis. One of the biggest forces in the decline in malaria deaths was the advent of the Global Fund to Fight AIDS, Tuberculosis and Malaria (GFATM), which provided 40% of development assistance for health targeted toward malaria from 2003 to 2008.
GFATM’s recent announcement that it would cancel its next round of funding threatens the gains made in preventing malaria deaths. If malaria elimination and eradication and broader health and development goals are to be achieved, donor support needs to be increased.
Citation: Murray CJL, Rosenfeld LC, Lim SS, Andrews KG, Foreman KJ, Haring D, Fullman N, Naghavi M, Lozano R, Lopez AD. Global malaria mortality between 1980 and 2010: a systematic analysis. The Lancet. 2012; 379:413-431.
Data and Methods
Data for download
Malaria deaths and cumulative probability of death (per 1,000) for all ages, by country, and year (Global), 1980-2010 (75KB xls)
Malaria deaths and cumulative probability of death (per 1,000) by age group, country, and year (Global), 1980-2010 (82KB xls)
For additional results data, visit our Global Health Data Exchange (GHDx). The GHDx includes data records with information on more than 200 countries.
Related Visualizations & other Tools
Related News & Events
In February 2012, IHME published in The Lancet global estimates for malaria mortality: Global malaria mortality between 1980 and 2010: a systematic analysis. The new estimates differed from multiple previous assessments of malaria mortality at the global level and in individual countries, the most recent being the World Malaria Report 2011 by the World Health Organization (WHO).
- We based our estimates on the broadest range of available data to date for 105 countries: a total of 1,150 site years of data from 1980 to 2010. This includes data from vital registration systems and from verbal autopsy studies.
- Both WHO and IHME use verbal autopsy studies for estimating mortality from all causes and for estimating mortality from malaria specifically. The World Malaria Report 2011 relied on verbal autopsy data for its child mortality estimates for some of the regions it studied. The report says, “Child malaria deaths were estimated using a verbal autopsy multi-cause model (VAMCM) developed by the WHO Child Health Epidemiology Reference Group (CHERG) to estimate causes of death for children aged 1-59 months in countries with less than 80% of vital registration coverage.” IHME used a wider range of verbal autopsy studies to generate its estimates and included adult deaths.
- WHO does not use verbal autopsy studies for estimates of adult deaths from malaria in sub-Saharan Africa, but it does use verbal autopsy studies for estimates of adult deaths in other areas. This is based on two assumptions. The first is the assumption – taught in medical and public health schools – that adults develop immunity from early exposure to malaria and do not die from the disease. The second is the assumption that verbal autopsy studies are not accurate enough to diagnose malaria, because the symptoms of malaria can be similar to other causes of death. IHME, using vital registration data and verbal autopsy studies for both children and adults, found that verbal autopsy studies yield conservative estimates of malaria mortality, meaning that the true number could be even higher.
- For African countries, WHO estimates are based on a model of malaria mortality that takes into account only population growth and the effects of vector control. IHME’s estimates include the effect of chloroquine resistance, the scale-up of artemisinin-combination treatment, environmental factors such as rainfall, and broader socioeconomic determinants.
- To overcome misclassification of malaria deaths attributed to other cases, IHME used an approach developed by IHME Assistant Professor Dr. Mohsen Naghavi and colleagues to account for changes in the International Classification of Diseases and Injuries. With this approach, IHME redistributed deaths that had been called “fever of other and unknown origin,” “disseminated intravascular coagulation,” “other and unspecified infectious diseases,” and “sequelae of other and unspecified infectious and parasitic diseases.” As a result, some of those deaths were reclassified as malaria deaths.
- IHME’s estimates confirm some findings from previous studies. For example, they show that malaria deaths increased by three times through the 1980s and 1990s to a peak in 2004. Previous studies also show an increase in malaria deaths in this period of two to three times. Both the IHME study and previous studies have noted an association with increasing chloroquine resistance.
For a detailed discussion of our methods, please see the Web Appendix with the journal article.