Craig Moscetti cmoscetti@globalhealth.org
February 3, 2012
WHO urges the malaria global health community to exercise caution when interpreting the malaria burden estimates published in the Lancet on 3 February 2012 by Christopher Murray and colleagues. The data published in the study have very high uncertainty intervals and coverage in the news media has led to some misleading conclusions, including that this study will "overturn decades of assumptions" about malaria deaths. WHO feels it is important to underline that while it appreciates the debate that the study will generate in the scientific world, some key findings do not seem to be based on strong evidence. WHO therefore does not believe that any immediate changes are required to the policies currently being applied by malaria endemic countries with the support of implementing partners and donors.
"Global malaria mortality between 1980 and 2010: a systematic analysis" is part of the "Global Burden of Diseases, Injuries and Risk Factors 2010" study that has been underway at the Institute for Health Metrics and Evaluation (IHME) for the last four years. WHO appreciates all the efforts by partners and the scientific community to improve estimation methods for the global malaria burden and welcomes this study as well. Mortality is challenging to measure under the best of circumstances and there is substantial uncertainty around any estimate of cause-specific deaths in countries which lack adequate vital registration systems - which is true of most countries with high malaria burdens.
While most media stories will be limited to highlighting the key differences between the IHME and WHO estimates, the malaria policy community will be aware that, in fact, there are many similarities between our findings. Both approaches show that an unacceptably high number of people still die from malaria, and that trends in the disease burden are steadily downward by both methods, thanks to investments in scaling up life-saving tools. In most settings and ages, the uncertainty intervals between WHO and IHME estimates overlap, meaning they are not statistically different. The similarities between the estimates produced by IHME and WHO include:
the total number of deaths has fallen consistently since 2004, following the rapid scale-up of life-saving malaria control interventions;
the percentage of all malaria deaths occurring in Africa is the same (91%);
the number of deaths estimated to occur in children are similar both in and outside of Africa (the uncertainty intervals of these estimates overlap);
the number of deaths estimated to occur outside of Africa in individuals over the age of five are similar (the uncertainty intervals of these estimates overlap);
The most important difference between WHO and IHME estimates is the number of deaths occurring in those greater than 5 years of age in Africa. WHO estimates 55,000 deaths and IHME 435,000 (nearly two-thirds of the total estimated deaths in children under 5 years of age in Africa). Such a result is unexpected in stably endemic areas, since partial immunity to malaria generally develops at an early age, protecting against severe disease and death. The IHME study uses verbal autopsy to identify cause of death for persons of all ages and IHME has previously conducted validation exercises in six study sites to support this work. However, five of those study sites had little or no malaria, and in the one site with a significant amount of malaria (Dar es Salaam), the quality of microscopic diagnosis, used to validate cause of deaths, has been challenged in other peer-reviewed articles (Kahama-Maro J et al 2011). Moreover, other studies have found considerable over-diagnosis of severe malaria in health facilities (e.g. Makani et al 2003). Hence, there is doubt about the classification of these deaths as malaria.
WHO constantly strives to improve its disease burden estimates and works with other partners including the RBM Monitoring and Evaluation Reference Group, the Child Health Epidemiology Reference Group, and governments of endemic countries to achieve consensus on methods for estimating the burden of disease using all available data. The fact that we must rely on estimates of malaria deaths (rather than primary collected data) highlights the need for improved malaria diagnostic testing, surveillance and vital registration. WHO and partners need to increase efforts to support endemic countries in this undertaking and stay on course to fight malaria with our existing tools and interventions.
Robert D. Newman, MD, MPH: Director, Global Malaria Programme: World Health Organization
E-mail: newmanr@who.int
February 3, 2012
WHO urges the malaria global health community to exercise caution when interpreting the malaria burden estimates published in the Lancet on 3 February 2012 by Christopher Murray and colleagues. The data published in the study have very high uncertainty intervals and coverage in the news media has led to some misleading conclusions, including that this study will "overturn decades of assumptions" about malaria deaths. WHO feels it is important to underline that while it appreciates the debate that the study will generate in the scientific world, some key findings do not seem to be based on strong evidence. WHO therefore does not believe that any immediate changes are required to the policies currently being applied by malaria endemic countries with the support of implementing partners and donors.
"Global malaria mortality between 1980 and 2010: a systematic analysis" is part of the "Global Burden of Diseases, Injuries and Risk Factors 2010" study that has been underway at the Institute for Health Metrics and Evaluation (IHME) for the last four years. WHO appreciates all the efforts by partners and the scientific community to improve estimation methods for the global malaria burden and welcomes this study as well. Mortality is challenging to measure under the best of circumstances and there is substantial uncertainty around any estimate of cause-specific deaths in countries which lack adequate vital registration systems - which is true of most countries with high malaria burdens.
While most media stories will be limited to highlighting the key differences between the IHME and WHO estimates, the malaria policy community will be aware that, in fact, there are many similarities between our findings. Both approaches show that an unacceptably high number of people still die from malaria, and that trends in the disease burden are steadily downward by both methods, thanks to investments in scaling up life-saving tools. In most settings and ages, the uncertainty intervals between WHO and IHME estimates overlap, meaning they are not statistically different. The similarities between the estimates produced by IHME and WHO include:
the total number of deaths has fallen consistently since 2004, following the rapid scale-up of life-saving malaria control interventions;
the percentage of all malaria deaths occurring in Africa is the same (91%);
the number of deaths estimated to occur in children are similar both in and outside of Africa (the uncertainty intervals of these estimates overlap);
the number of deaths estimated to occur outside of Africa in individuals over the age of five are similar (the uncertainty intervals of these estimates overlap);
The most important difference between WHO and IHME estimates is the number of deaths occurring in those greater than 5 years of age in Africa. WHO estimates 55,000 deaths and IHME 435,000 (nearly two-thirds of the total estimated deaths in children under 5 years of age in Africa). Such a result is unexpected in stably endemic areas, since partial immunity to malaria generally develops at an early age, protecting against severe disease and death. The IHME study uses verbal autopsy to identify cause of death for persons of all ages and IHME has previously conducted validation exercises in six study sites to support this work. However, five of those study sites had little or no malaria, and in the one site with a significant amount of malaria (Dar es Salaam), the quality of microscopic diagnosis, used to validate cause of deaths, has been challenged in other peer-reviewed articles (Kahama-Maro J et al 2011). Moreover, other studies have found considerable over-diagnosis of severe malaria in health facilities (e.g. Makani et al 2003). Hence, there is doubt about the classification of these deaths as malaria.
WHO constantly strives to improve its disease burden estimates and works with other partners including the RBM Monitoring and Evaluation Reference Group, the Child Health Epidemiology Reference Group, and governments of endemic countries to achieve consensus on methods for estimating the burden of disease using all available data. The fact that we must rely on estimates of malaria deaths (rather than primary collected data) highlights the need for improved malaria diagnostic testing, surveillance and vital registration. WHO and partners need to increase efforts to support endemic countries in this undertaking and stay on course to fight malaria with our existing tools and interventions.
Robert D. Newman, MD, MPH: Director, Global Malaria Programme: World Health Organization
E-mail: newmanr@who.int
