To the best of our knowledge, this is the first study to examine food expenditure patterns in two territories in the Canadian Arctic (NU and the NWT). About 34% of total food expenditure by participants was spent on NNDF. Participants also spent more money on non-traditional meats and traditional meats, compared with fruit and vegetables, grains and potatoes, and dairy products.
The findings of the present study raise significant questions concerning the disproportionate amount of money being spent on NNDF, which do not provide adequate nutrition, but provide relatively high levels of energy. NNDF includes a relatively large group of items that were usually available in community food stores, in contrast to fresh produce, which was not as readily available on a consistent basis. This suggests that, on a daily basis, most of the NNDF items were more available than fresh produce. For individuals who are food shopping for the entire family, the purchase of NNDF may be the easiest way of obtaining the most energy while ensuring that all members of the family were satisfied. In addition, a taste preference may exist for high fat foods . Drewnowski and Specter  reported that the relationship between obesity and poverty can be explained by the low cost of foods that are energy dense and the taste and acceptability of these foods that are usually high in fat and sugar. Affordability, quality of store-bought perishable foods, exposure to fresh produce, and culinary knowledge are other factors that could be linked to lower expenditure on nutrient-dense food groups, particularly fruit and vegetables. Increased cost of fuel and hunting equipment, as well as reduced availability of land, sea and sky resources, such as caribou, Arctic char, and geese, possibly as a result of climate changes , are potential reasons that traditional meats expenditure is lower than non-traditional meat expenditure among this population.
Food group expenditure in this study varied among age, gender, and education subgroups. This finding is supported by results from an Expenditure Survey conducted by the U.S. Bureau of Labor Statistics . The reality in traditional and remote communities is that men generally spend more time outside the household for work and have less food-preparation skills than women; this could explain the gender difference of non-traditional meat food expenditure in this study. Physiological need difference between genders for protein and energy expenditure could also partly explain the variety of meat intake and expenditure between men and women in the study population . The finding that women consume diets with less meat reported by other studies [24, 25] support our result.
Our analysis showed a statistically significant, increasing trend for traditional meats expenditure and a decreasing trend for NNDF expenditure with an increase of age. Similarly, in the study by Fan et al. older households were less likely to be in the fast-food-dominant food expenditure category. Fast food tends to be more energy dense, higher in saturated fat and salt, and lower in micronutrients relative to other foods  and thus, could be considered equivalent to NNDF in this study. It is known that dietary habits tend to be relatively stable over time , suggesting that older people may have a higher tendency towards traditional meats in these communities.
The finding that participants with higher levels of formal education spent significantly more on fruit and vegetables is comparable with the results from the East Anglia cohort of the European Prospective Investigation of Cancer analysis . This finding could be linked to having greater nutrition knowledge regarding fruit and vegetables, being more familiar with method of preparing dishes using fruit and vegetables, and more taste preference for this food group. A high level of expenditure on traditional meats by people classified with a lower level of formal education is compatible with the similar expenditure pattern and amount among older participants (both groups spent 21% of the food expenditure on traditional meats). This indicates a preference towards traditional meats is likely related to interest in and commitment to customs as well as an awareness of nutritional values of traditional meats.
In 2007–2008, approximately 7.7% of Canadian households and 21% of off-reserve Aboriginal households were food insecure . In one recent survey, nearly 70% of Inuit families in 16 NU communities were food insecure . Previous studies among this population indicate that this may be an issue among Inuit and Inuvialuit . In other populations, individuals who are food insecure and have a low income often rely heavily on low-cost foods that are high in energy (sugar and fat) but lack essential nutrients [20, 30]. In this study, however, no difference in patterns of food expenditure was found between different levels of socioeconomic status (MSL scale), suggesting factors such as taste preference or availability may have more of an influence on food expenditure patterns, at least among people with adequate income. In addition, this finding indicates that any future intervention program could be broadly applied to all members of the community.
Nutrition related diseases such as obesity and diet related anemia can be caused by an insufficient intake of food, insufficient intake of certain nutrients or by an over consumption of certain foods . Addressing dietary inadequacies through appropriate and relevant intervention programming has the potential to positively impact health status and to reduce the cost to health systems in the Canadian Arctic.
Municipal, regional, territorial, and federal governments and organizations should consider the findings of this study in the development, design, and implementation of any programming or subsidies related to the food environment in the Canadian Arctic. Particular focus should be given to costs associated with hunting and gathering and supportive programs for younger generations of Inuit/Inuvialuit to learn hunting and gathering skills.
All future programming should consider challenges to obtaining a healthy diet in the Canadian Arctic, whether they relate to store-bought or traditional meats. High costs associated with household essentials, such as diapers and toilet paper, can take away from the disposable household income available for the purchase of food. As such, given the importance of traditional meats to the diet of Inuit/Inuvialuit, subsidies should be provided to reduce costs associated with hunting, such as snowmobile parts, ammunition, and fishing supplies. Any programs or subsidies should include rigorous community consultation to ensure programs are relevant and appropriate for the current Northern environment. Communities that are to receive the program and/or subsidy should endorse all promoted items.
Considering the findings of this study in relation to food choices and costs, programs should be developed to promote food items that have the largest impact on diet and be the most amenable to Inuit/Inuvialuit communities, such as eggs and yogurt. In addition, the current study shows that younger people tended to pay more for NNDF. Thus, particular emphasis should be paid to engaging young people in the development and implementation of household and individual budgeting support, in addition to healthy food choices, as a key element of any intervention program.
Further research into all factors impacting overall household income in the Canadian Arctic is required to ensure any program and policy decisions are appropriate, relevant, and timely. If a large percentage of household income is spent on NNDFs, then the available disposable income will decrease, considering a subsidy of only healthy food choices. Without community-led programs that support healthy food choices, simply subsidizing the cost of certain foods will continue to increase the likelihood of food insecurity in the Canadian Arctic.
Food prices were sampled over a 12-month period and therefore, variations in food costs due to seasonality and for different brands were captured. In addition, there was very limited diversity for food item brands between communities due to remoteness. Women contributed more than men in the study because the study preference was to recruit a member of each household who was the main person shopping for and preparing food for the family. This might limit the ability to generalize the findings to people living outside of these communities. However, it was necessary to collect detailed and accurate dietary information from participants to address the HFN project’s objectives. This need could be addressed by recruiting main food shoppers from each household as those people were more likely to have food knowledge and manage food expenditure for the household. Accordingly, collecting more accurate dietary data was more advantageous to the study compared with the disadvantage associated with potential biases. Majority (54%) of participants were between 30 and 49 years. This matches with the age pyramid in these communities where 46% of the population in the six communities was also between 30 and 49 years. A chi-squared test also did not show any significant difference between communities and the study participants for the proportion of people in the three age groups (P = 0.19). However, women usually attach greater importance to healthy eating (e.g. more fibre and less fat intake) . Only formal levels of education were recorded during data collection because the valuable traditional culture knowledge and education was not easily measurable.