Balanced diet and nutrient intake is an important component of a healthy lifestyle. Dietary guidelines have been mainly based on observational evidence. However, observational designs are prone to high risk of bias, among which confounding is the most common and difficult to deal with. Dietary and nutritional factors are often highly correlated with socioeconomic, lifestyle and other factors, which are often inaccurately measured. Multivariable adjustment, matching, stratification and other approaches are commonly used to reduce confounding bias, but only known and measured confounders can be addressed. Residual confounding can be resulted from unknown or unmeasured confounders, as well as inaccurate measurement of confounders. Limiting residual confounding has been pivotal to causal inference in nutritional epidemiology.
Randomized controlled trial is the gold standard design for assessing the effect of an intervention on health outcomes, whereby randomization process removes all known and unknown confounding and achieves comparability between groups.
Mendelian randomization is a study design of using genetic variants as instrumental variables to examine exposure-outcome associations. It is similar to randomized controlled trials in the way that genotypes are assorted randomly at meiosis and fixed at conception, thus removing residual confounding.
Mediation analysis provides another approach for causal inference, by seeking to identify potential mechanisms underlying an observed exposure-outcome association. It examines both direct and via-mediator indirect effects of exposure on outcome.
The goal of this Research Topic is to present epidemiological studies that address the association between dietary/nutritional factors and health outcomes using confounding-reduction methods, as well as methodological studies that advance how to further overcome residual confounding and/or other bias in nutritional epidemiological research.
We welcome Original Research articles, reviews, systematic reviews and perspectives with a focus on enhancing causal inference in nutritional epidemiology, including but not limited to the following topics:
• Randomized controlled trials assessing the effect of dietary or nutritional intervention on health outcomes;
• Mendelian randomization studies assessing the associations between dietary or nutritional factors with health outcomes;
• Nutritional epidemiological studies involving mediation analysis to examine possible mechanisms for observed associations;
• Studies that assess the effect of residual confounding on observational associations;
• Studies that discuss/propose novel approaches to deal with residual confounding or other bias in nutritional epidemiology.
Balanced diet and nutrient intake is an important component of a healthy lifestyle. Dietary guidelines have been mainly based on observational evidence. However, observational designs are prone to high risk of bias, among which confounding is the most common and difficult to deal with. Dietary and nutritional factors are often highly correlated with socioeconomic, lifestyle and other factors, which are often inaccurately measured. Multivariable adjustment, matching, stratification and other approaches are commonly used to reduce confounding bias, but only known and measured confounders can be addressed. Residual confounding can be resulted from unknown or unmeasured confounders, as well as inaccurate measurement of confounders. Limiting residual confounding has been pivotal to causal inference in nutritional epidemiology.
Randomized controlled trial is the gold standard design for assessing the effect of an intervention on health outcomes, whereby randomization process removes all known and unknown confounding and achieves comparability between groups.
Mendelian randomization is a study design of using genetic variants as instrumental variables to examine exposure-outcome associations. It is similar to randomized controlled trials in the way that genotypes are assorted randomly at meiosis and fixed at conception, thus removing residual confounding.
Mediation analysis provides another approach for causal inference, by seeking to identify potential mechanisms underlying an observed exposure-outcome association. It examines both direct and via-mediator indirect effects of exposure on outcome.
The goal of this Research Topic is to present epidemiological studies that address the association between dietary/nutritional factors and health outcomes using confounding-reduction methods, as well as methodological studies that advance how to further overcome residual confounding and/or other bias in nutritional epidemiological research.
We welcome Original Research articles, reviews, systematic reviews and perspectives with a focus on enhancing causal inference in nutritional epidemiology, including but not limited to the following topics:
• Randomized controlled trials assessing the effect of dietary or nutritional intervention on health outcomes;
• Mendelian randomization studies assessing the associations between dietary or nutritional factors with health outcomes;
• Nutritional epidemiological studies involving mediation analysis to examine possible mechanisms for observed associations;
• Studies that assess the effect of residual confounding on observational associations;
• Studies that discuss/propose novel approaches to deal with residual confounding or other bias in nutritional epidemiology.