A multi-stage sampling strategy selected 1387 on-reserve First Nations adults in Ontario. Foods from a 24-hour dietary recall were assigned to the 100 most common food groups for men and women. Nutrients from market foods (MF) and traditional foods (TF) harvested from the wild as well as MF costs were assigned based on the proportions of total grams consumed. Linear programming was performed imposing various constraints to determine whether it was possible to develop diets that included the most popular foods while meeting Institute of Medicine guidelines. Final models were obtained for both sexes with the top 100 food groups consumed while limiting the nutrient-poor foods to no more than the actual observed intake. These models met all nutrient constraints for men but those for dietary fibre, linoleic acid, phosphorus, and potassium were removed for women. MF costs were obtained from community retailers and online resources. A grocery list was then developed and MF were costed for a family of 4. The grocery list underestimated the actual weekly food cost because TF was not included. Contemporary observed diets deviated from healthier historic First Nations diets. A culturally appropriate diet would include more traditional First Nations foods and fewer MF. Novelty: Linear programming is a mathematical approach to evaluating the diets of First Nations. The grocery list is representative of food patterns within Ontario First Nations and can be used as an alternative to the nutritious food basket used for public health food costing.
Keywords: Autochtones; First Nations; Indigenous; Premières Nations; diet modelling; diet optimization; linear programming; modélisation de l’alimentation; nutrition transition; optimisation de l’alimentation; programmation linéaire; transition nutritionnelle.