[Development and evaluation of a modified version of the dietary variety score for community-dwelling older adults]

Objectives　The dietary variety score (DVS) was developed as an index to assess the variety of food intake among older Japanese adults and has been widely used in epidemiological studies and public health settings. However, this index has not been reviewed since its development in the 1990s and may not adequately reflect the current dietary habits of the older Japanese population. In this study, we developed a modified version of the DVS (MDVS) and examined its validity.Methods　We conducted a cross-sectional study of 357 community-dwelling adults aged ≥65 years (mean age: 76.2±4.6, men: 61.1%) who participated in the 2016 survey of the Hatoyama cohort study. The DVS and MDVS were assessed by the number of food groups eaten almost every day based on the weekly frequency of consumption. The DVS was calculated based on ten components: meat, fish/shellfish, eggs, milk, soybean products, green/yellow vegetables, potatoes, fruits, seaweed, and fats/oils. The components of the MDVS were determined by calculating the contribution of each food group to the total and nutrient intakes from the food groups comprising the main dish, side dish, and soup, based on the data from older adults aged ≥65 years who participated in the 2017 National Health and Nutrition Survey in Japan. Based on the contribution of each food group, we added "other vegetables" and "dairy products" in the MDVS. Dietary intake was assessed using a validated, self-administered diet history questionnaire. We calculated the probability of adequacy for each of the fourteen nutrients selected as per the estimated average requirement in the Dietary Reference Intake for Japanese 2020 and the mean probability of adequacy for all fourteen nutrients. In addition, we assessed the correlation between the DVS, MDVS, and each indicator and the significance of the difference between correlation coefficients.Results　The MDVS was significantly positively correlated with the energy ratios of protein and fat, dietary fiber, potassium, and the modified diet score based on the Japanese Food Guide Spinning Top (r=0.21-0.45) and negatively correlated with the carbohydrate energy ratio (r=-0.32). The MDVS was also correlated with the mean probability of nutrient adequacy (r=0.41). The correlation coefficients between the MDVS and each indicator were not significantly different from those of the DVS.Conclusions　Validity in terms of nutrient intake and diet quality was comparable between the DVS and MDVS. To revise the DVS, it is necessary to conduct nationwide studies based on highly accurate dietary surveys.