From the American College of EpidemiologyWhat matters most: quantifying an epidemiology of consequence
Section snippets
Our charge as epidemiologists and the limits of risk factor epidemiology
Individuals who are overweight might have longer life spans [1], [2] (or not [3]), blueberry and strawberry consumption may decrease risk for heart attacks [4] (or not [5]), moderate alcohol consumption is good for cardiovascular health [6] (or not [7]), green tea consumption might prevent stomach cancer [8] (or not [9]), and on and on. The list of diet, lifestyle, environmental, and genetic factors that purportedly cause or prevent a wide range of chronic diseases is voluminous, and often in
A re-emphasis on what matters most
The emphasis on identifying risk factors within a paradigm that hunts for precise causal effects obscures what we argue is the broader goal of the field—an attempt to identify “what matters most.” This approach urges us to identify what we can do about those factors that do indeed matter most for the health of populations, which necessarily involves both theory-driven approaches and a pragmatic assessment of what is likely to make a difference. Instrumentally, this would mean a focus on the
Mathematically understanding the effect of prevalent versus rare causes
The magnitude of the risk ratios and risk differences we obtain in our studies for the effect of an exposure on an outcome is dependent on the prevalence of those causes that interact with the exposure of interest [47], [39]. Thus, the idea that we can identify “the” causal effect of an exposure on an outcome is not only inefficient but also at odds with the very math of risk ratio and difference estimation when the exposure of interest is not sufficient to produce the outcome. We can clearly
From simulation to the community: shifting exposure prevalence across geographic space and time
One could potentially write off our previously mentioned example as a convenient mathematical exercise, but there is substantial empirical literature to indicate that such variations in the magnitudes of our effect estimates occur frequently in the empirical literature. These variations are sometimes explained by random chance [54], [55], faulty study design, or other methodologic bias, exposing our innate preference for exposures to have one true causal effect in the population.
Of course, the
Implications and conclusions for conducting an epidemiology of consequence
To conduct an epidemiology of consequence, we need to identify what matters most for population health so that we can guide public health stakeholders toward strategies that reduce the burden of these factors. It is hard to argue that we should not be thinking about what matters most as we endeavor to build our research questions and design studies to answer these questions, and our mathematical simulation demonstrates the importance of such an approach. The next question is, then, how do we
Acknowledgments
Funding was provided to Katherine Keyes by the National Institute on Alcohol Abuse and Alcoholism (K01AA021511).
References (82)
Epidemiology and the web of causation: has anyone seen the spider?
Soc Sci Med
(1994)- et al.
Mental illness and reduction of gun violence and suicide: bringing epidemiologic research to policy
Ann Epidemiol
(2015) - et al.
Should the mission of epidemiology include the eradication of poverty?
Lancet
(1998) - et al.
Adolescents with and without gestational cocaine exposure: longitudinal analysis of inhibitory control, memory and receptive language
Neurotoxicol Teratol
(2011) - et al.
A personal view on systems medicine and the emergence of proactive P4 medicine: predictive, preventive, personalized and participatory
N Biotechnol
(2012) - et al.
Mortality, health outcomes, and body mass index in the overweight range: a science advisory from the American Heart Association
Circulation
(2009) - et al.
Overweight, mortality and survival
Obesity
(2013) - et al.
Body-mass index and mortality among 1.46 million white adults
N Engl J Med
(2010) - et al.
High anthocyanin intake is associated with a reduced risk of myocardial infarction in young and middle-aged women
Circulation
(2013) - et al.
Strawberry intake, lipids, C-reactive protein, and the risk of cardiovascular disease in women
J Am Coll Nutr
(2007)