How different acceptability modelling approaches influence sustainable diet solutions: an overview of methods and a case study

Diet optimisation modelling is increasingly used to design sustainable diets that meet nutritional and environmental goals. However, adequately capturing (cultural) acceptability in these models remains a major challenge. While most studies rely on a few common methods to address acceptability – defined as distance to observed diets – their impact on model outcomes is not well understood. This study systematically evaluated the extent to which different acceptability modelling approaches affect sustainable diet solutions. First, we reviewed existing strategies to incorporate acceptability into diet optimisation models. Second, we applied a selection of the most frequently used approaches in two models (iOTA Model® and SHARP Model) under consistent data and parameter conditions. Sixteen combinations of acceptability approaches were tested, covering different model components (algorithm, decision variables, objective function and constraints). Results showed that the choice of acceptability approach substantially influenced both the quantitative and qualitative outcomes of the models, although some consistent patterns emerged. Strong agreement was observed for meat (decreased consumption from 20% greenhouse gas emissions reduction onwards) and directional agreement for cereals, fish and seafood, fruits, legumes, nuts and seeds, potatoes, and vegetables (maintained or increased consumption). Minimal agreement was found for dairy products or substitutes, eggs, fats and oils, meat substitutes, and sugar-rich or salty foods. These findings underscore that methodological choices regarding acceptability influence sustainable diet solutions, with important implications for dietary recommendations and food policy. To enhance robustness and relevance, future studies should explicitly justify acceptability modelling choices and test multiple approaches to assess the sensitivity of model outcomes.