Evaluating adherence to recommended diets in adults 1991–2015: revised China dietary guidelines index

BACKGROUND: The China Dietary Guidelines Index (CDGI) is a diet quality evaluation index that can present the overall diet quality and is comparable between individuals. The aim of this study was to revise CDGI for Chinese adults according to the Chinese Dietary Guidelines 2016 (CDG-2016), evaluate adherence to recommended diets between 1991 and 2015, and analyze the trend, variation, and determinants of diet quality. METHODS: Food, cooking oil, and condiment intakes were estimated based on twenty-four-hour dietary recalls over three consecutive days and the household weighing method. Based on the food and nutrients recommendations for people with different energy requirements in CDG-2016, CDGI was revised as China Dietary Guidelines Index (2019)-Adults (CDGI(2019)-A) by equal weight continuity scoring. Three-level random intercept-slope growth models were applied to analyze the trend, variation, and determinants at both the community and individual levels. RESULTS: CDGI(2019)-A, the sum of fourteen component scores with a range of 0–110 points, increased significantly from 38.2 in 1991 to 47.3 in 2015. Components with a score of less than half were milk (91.6%), fruits (72.0%), nuts (82.5%), other cereals and beans (82.6%), and seafood (77.7%). Between-individual accounted for 25.6% of the total score variation, of which 87.4% derived from the community level. CDGI(2019)-A score displayed a positive association with being female, having higher education, having higher income, living in an urban area, and knowing the CDG-2016 recommendations. The impact of income and awareness of CDG-2016 varied significantly across communities. CONCLUSIONS: Although quality of diet has been improving in China, overall quality remains poor, primarily because of inadequate intake of milk and dairy products, nuts, fruits, other cereals and miscellaneous beans, and seafood. Intervention at the community level may improve diet quality more efficiently than at the individual level, and the most effective intervention should be selected in different communities according to local conditions.