Examining Dietary Patterns in Relation to Chronic Disease

In this issue of Circulation , Shikany and colleagues1 add to the growing body of literature on dietary patterns in relation to health outcomes. The purpose of their article was to test the hypothesis that dietary patterns extracted through factor analysis are associated with the risk of acute coronary heart disease (CHD) in the Reasons for Geographic and Racial Differences in Stroke Study. The authors note that studies examining associations between dietary patterns and CHD in samples with sociodemographic and regional diversity are lacking. Understanding the drivers of health disparities is critically important, and the topic of dietary patterns is of increasing interest, as noted by the most recent Dietary Guidelines Advisory Committee Report, “…because the totality of diet…may have synergistic and cumulative effects on health and disease.”2Article see p 804Shikany and colleagues state that the use of rigorous factor analysis was a major strength of their study because it is a data-driven method through which patterns are empirically derived. They conclude that “a dietary pattern characteristic of the southern US” was associated with greater hazard of CHD, whereas a “plant-based pattern” was not. We believe these assertions deserve further consideration.Factor analysis is 1 of several ways of defining dietary patterns, each of which addresses a particular question (Table). Methods for studying patterns can be grouped into 3 broad categories: data-driven, outcome-independent; data-driven, outcome-dependent; and investigator-defined. Note that the labels are simply a shorthand way to refer to how the patterns are derived. Data-driven does not mean that a method is more evidence based, and investigator-defined does not mean a method includes more subjectivity. Each method is built on some type of evidence and includes some degree of subjectivity.3View this table:Table. Methods Used to Identify Dietary Patterns, the Corresponding Questions They Address, and Other Considerations for …