Exploring culturally acceptable, nutritious, affordable, and low climatic impact diet for Japanese diets: Proof of concept of applying a new modelling approach using Data Envelopment Analysis

Sugimoto, Minami; Temme, Elisabeth H.M.; Biesbroek, Sander; Kanellopoulos, Argyris; Okubo, Hitomi; Fujiwara, Aya; Asakura, Keiko; Masayasu, Shizuko; Sasaki, Satoshi; ’t Veer, Pieter van


A future sustainable dietary pattern for Japanese is yet undefined. This study aimed to explore more sustainable Japanese diets, that are nutritious, affordable, and with low greenhouse gas emissions (GHGE) and particular emphasis on cultural acceptability. A newly developed Data Envelopment Analysis diet model was applied to 4-d dietary record data among 184 healthy Japanese men and 185 women volunteers aged 21–69 y. Alternative diets were calculated as the linear combinations of observed diets. Firstly, for each individual, four modelled diets were calculated that maximised cultural acceptability (i.e. minimize dietary change from observed diet), maximised nutritional quality assessed by the Nutrient-Rich Food Index (NRF), minimized monetary diet costs, or minimized diet-related GHGE. The final modelled diet combined all four indicators. In the first four models, the largest improvement was obtained for each targeted indicator separately, while relatively small improvements or unwanted changes were observed for other indicator. When all indicators were aimed to optimize, the NRF score and diet-related GHGE was improved by 8–13% with the lower monetary cost than observed diets, although the percentage improvement was a bit smaller than the separate models. The final modelled diets demanded increased intakes for whole grains, fruits, milk/cream/yoghurt, legumes/nuts, and decreased intakes for red and processed meat, sugar/confectionaries, alcoholic and sweetened beverages, and seasonings in both sexes. In conclusion, more sustainable dietary patterns considering several indicators are possible for Japanese while total improvement is moderate due to trade-offs between indicators and methodological limitation of DEA diet model.