In this issue of JAMA Internal Medicine, Haring et al1 provide what appears to be the first detailed examination of a Mediterranean diet index and 3 other dietary quality indexes in association with the risk of hip and total fractures. They report that the 4 commonly used indexes predict a lower risk of hip fractures.
These a priori dietary indexes are one form of dietary pattern analyses, with the other being empirical dietary patterns based on statistical methods that take into account correlations among consumption of different foods. The use of dietary patterns in epidemiologic studies and intervention trials to complement studies of specific nutrients and foods has increased because effects of diet are likely to be strongest and clearest when contributions from multiple aspects of diet are combined. In addition, because isolating the effect of a specific nutrient or food from other highly correlated components of diet can be difficult, we can sometimes have greater confidence that an association with an overall dietary pattern is causal than we can for associations with specific components of that diet. One of the early uses of an a priori dietary index was the Healthy Eating Index (HEI), which was created by the US Department of Agriculture to describe adherence to the 1995 US Dietary Guidelines. Because of concerns that the focus of the 1995 guidelines—reduction of total fat and a broad increase in carbohydrates—was not supported by good evidence, we used the HEI to score the diets of participants in the Nurses’ Health Study and Health Professionals Follow-up Study using dietary data that had been collected every 4 years since 1986. After adjusting for smoking, physical activity, and other health-related behaviors, HEI scores were not associated with a composite outcome of cardiovascular disease, cancer, and total mortality. Thus, we created the Alternative Healthy Eating Index, which accounted for type of fat, form of carbohydrate, and source of protein; when applied to the same dietary data, this score strongly predicted a lower risk of this composite of major chronic disease outcomes in both men and women.2 Since that time, the US Dietary Guidelines and corresponding modifications of the HEI have moved closer to the diet described by the Alternative Healthy Eating Index, and both dietary indexes predict better health outcomes.3 More recently, the Alternative Healthy Eating Index has been used to track US trends in diet quality since 2000, documenting a steady improvement that would account for major health benefits.4 The Mediterranean Diet Index was developed to describe adherence to the traditional diet of Greece; this score and a modification for countries in which olive oil is not traditional (the alternative Mediterranean Diet Index) have been strongly associated with better health outcomes in Greece and elsewhere.5 The diet score used in the randomized Dietary Approaches to Stop Hypertension (DASH) trial was developed to describe the dietary pattern documented to reduce blood pressure. Although these dietary indexes differ in some ways, they generally emphasize intake of fruits, vegetables, whole grains, and plant sources of protein and deemphasize refined starch, sugar, and red meat.