Why the Atlas Cites IHME’s Tobacco Use EstimatesEconomic and Health Policy Research, American Cancer Society
The 5th Edition of The Tobacco Atlas aims to be a transparent clearinghouse of the best available data and analysis in global tobacco control. To this end, we have tried to be as open as possible in showing our methods and offering the data that underpin our maps and figures for easy download on our website. We use as much statistically equivalent data from as few sources as possible to cover as much of the world as possible in any one measure.
We have chosen to cite the tobacco use prevalence estimates produced by the Institute of Health Metrics and Evaluation (IHME) in the Atlas because their work is global, inclusive, and constantly striving to improve–just like The Tobacco Atlas.
The work of the epidemiologists at IHME has risen to the attention of the world through their groundbreaking Global Burden of Disease (GBD) study, which sought to determine the risk factors and diseases that caused all the death and disability in every person around the world. The GBD found that in 2010, tobacco use was third largest cause of death globally (after poor diet and high blood pressure). While some scientists dispute the precise numerical findings of the GBD, there is something incredibly useful about trying to make better and collective sense of the many individual studies that measure the impact of disease on the health of the public. The GBD aims to measure the impact of thousands of diseases caused by hundreds of risk factors and proximate causes in a manner that tries to apportion the harms of complex interaction with the environment.
In order to more accurately calculate this burden of tobacco use, IHME endeavored to produce longitudinal estimates of tobacco use in every country, age group, and sex, using a standardized methodology (described here).
IHME has produced tobacco prevalence estimates for 188 countries for a time span of 1980 to 2013 (although estimates for just 187 of those countries from 1980 to 2012 are available for download at the IHME website), and the organization plans on updating its models and the resulting estimates as more surveillance data flow in over time. They have maintained a scientific and professional posture as they have proven to be open to criticism of their approach, and will gladly add “missing” nationally representative survey data to their models if they were unaware such data existed.
A sign of vibrancy in the field of public health (and particularly in tobacco control) is that multiple surveys of tobacco use are often conducted simultaneously in a single country. The surveys may vary in scope, sample size, and the questions asked, but every bit of tobacco use surveillance data holds within it a glimpse at the true tobacco use rate within a country or an age group. Consider this example: IHME gathered data from 3 surveys of tobacco use in Turkish adults that were conducted in 2010. The Chronic Renal Disease in Turkey Study (CRIDS), the Turkey Health Interview Survey (THIS), and the BREATHE study were all conducted at the same time, and because they found widely varying tobacco use prevalence estimates for 40 to 44 year old males, from 37% to 61% of males in the age group, creating a process to reconcile these differing findings was necessary.
IHME has created that necessary process, by which the findings of CRIDS, THIS, and BREATHE are reconciled, along with other historical data sources to produce a smoking prevalence estimate of Turkish males aged 40 to 44 years in 2010 of 44% (95% CI, 41-55%). Admittedly, this is a large confidence interval for a prevalence estimate, but it reflects the best-known compilation of all the available epidemiological data.
By synthesizing all available data into a single prevalence estimate, IHME does a great service to us all by not forcing policymakers, researchers, and advocates to pick and choose which individual survey estimate represents reality.
The Tobacco Atlas team deeply respects IHME’s work, and perhaps just as much, their efforts at transparency and openness, and willingness to update their own thinking, and looks forward to future cooperative efforts with their organization. The Tobacco Atlas team believes that the IHME prevalence estimates are an honest attempt to derive the best possible answers from all the available data, a kindred mission to our own. For these reasons and more we will continue to use IHME tobacco use prevalence estimates in the current paper and web versions of The Tobacco Atlas.