Study Design
This cross-sectional study used data from the Family Food Skills Study, which is a family-based study designed to understand associations between parental food literacy and dietary intake among families with young children. Families were eligible if they had at least one child between 2-5 years of age (in 2017) and 2-8 years of age (in 2018) and resided in Guelph-Wellington area of Ontario. In addition, parents had to be comfortable speaking and reading in English as well as having no prior food or nutrition training (e.g. Registered Dietitian, Chef, or Culinary Student). Families were recruited through social media, events in the community and local childcare centres. Data collection took place over a 4-week period. Initial home visits were conducted prior to data collection to provide families with an overview of the study and to obtain written, informed consent. After the data collection was completed, final home visits were completed to collect study material and to provide the grocery gift card incentive.
Recruitment and data collection occurred in two waves. The first wave occurred in August-September 2017 and the second occurred in August-September 2018. Of 55 families enrolled in the study in the 2017 wave, 1 family withdrew, and 7 families were excluded due to missing data. A total of 50 families enrolled for the study in the 2018 wave, but 8 families dropped out and 4 families were excluded due to incomplete sets of data. Thus, our final analytic sample included 85 families. The primary reason for withdrawal from the study was scheduling conflicts which made it difficult to complete the data collection. In families with more than one child in the target age range, the oldest child was chosen to participate in 2017; whereas, in 2018, the child with the nearest birthdate was selected to participate in the study. While some families had two parents participate in the study, only data from parent 1 (the first parent to sign up) were included in these analyses. All details are illustrated in Figure 1.
Fig. 1 Flowchart of the Family Food Skills Study families in 2017 and 2018.
Variables
Diet Quality (Healthy Eating Index-2015)
Parents completed 3-day food records, which included details about all food and beverages consumed on two weekdays and one weekend day for themselves and for their participating child. Our study protocol did not specify that food recording had to be consecutive days. Detailed instructions on how to complete the food records were provided during the initial home-visit. Parents were provided supplementary documents to help complete the food records including a guide for estimating portion sizes and an example of a complete food record. Completed food records were entered into the Food Processor Nutrition Analysis Software version 11.6.441 (ESHA Research, Salem, OR, USA) by a trained research assistant. A second research assistant then checked the entered data to ensure accuracy.
Parent and child diet quality was assessed using the Healthy Eating Index-2015 (HEI-2015). Under the supervision of a Registered Dietitian (RD), the food records were manually examined by research assistants for HEI moderation and adequacy components, which are expressed relative to energy intake, i.e., as densities, and converted to HEI equivalents. Sodium and fatty acid intakes were obtained from ESHA Food Processor Nutrition Analysis Software and were also used as adequacy and moderation components. Added sugars were manually calculated through retrieving information from designated food labels. The research assistants then calculated HEI-2015 scores based on average intake over the three days. The HEI-2015 is a validated measure of diet quality for individuals aged 2 years and older [9]. Scores range from 0 to 100, with higher scores indicating better diet quality. The HEI has been adapted for use with Canadian populations [10]. Since the version of ESHA used data based on the USDA recommendations – the HEI-2015 was most appropriate for this analysis.
Daily Per Capita Food Waste
Waste audits were conducted over a four-week period in 2017 and a three-week period in 2018 to determine daily per capita food waste. The shorter audit period in 2018 was driven by logistical constraints (i.e., availability of auditors and a holiday long-weekend, which can change waste behaviors). Research assistants collected all three waste streams (garbage, recycling, and organic bins) on the day the family would normally have their waste collected by the municipality and delivered the material to a waste sorting facility each week for four (2017) and three (2018) consecutive weeks. Individual food items were categorized and weighed (in grams) separately and sorted into the following six categories: fruits and vegetables, milk, cheese and eggs, meat and fish, breads and cereals, fats and sugars, and other (e.g. primarily coffee grounds). The other category was not used in the statistical analysis for this study. These categories were taken from the Waste and Resources Action Programme (WRAP) Household Food Waste Collection Guide [11].
Subsequently, each item was categorized as avoidable, unavoidable or possibly avoidable food waste. Avoidable was defined as food being discarded that is edible (e.g. bread, apples) [12]. Food items in this category typically could have been eaten if managed better [13]. Unavoidable was defined as food that is not edible (e.g. banana peels, coffee grounds) [12]. Possibly avoidable were items that can be eaten or prepared in different ways depending on the individual (e.g. potato skin, apple skin). Some food items were categorized as unidentifiable (e.g. waste that could not be recognized as belonging to one of the six food categories) or unknown (e.g. waste that could not be identified as a particular food but could be identified as belonging to a food group category). Unidentifiable/unknown foods made up approximately 11% of the total food waste. The weights of these unknown/unidentifiable foods were assigned proportionally to the six food categories to develop estimates of total waste output. For example, if a household had “carrot” food scraps that constituted 15% of the household’s known vegetable weight, the “carrot” category would receive 15% of that household’s “unknown vegetable” category. Lastly, some items were identified as non-food organics (e.g. teabags, paper towel). There was no unavoidable waste in the breads and cereals, as well as, fats and sugars food category; therefore, these were not included in the analysis on daily per capita unavoidable food waste. The total mean weights of the 2017 and 2018 subsamples were compared using a Mann-Whitney test and the sub-groups were found to be appropriate for combined subsequent analysis.
For the purpose of this analysis, avoidable and unavoidable food waste were reported as a daily average by dividing the weekly amount of waste by 7 days. Subsequently, a per capita amount of food waste was generated by dividing the daily average by number of family members in the household.
Household Income
Household income was measured and used as a covariate in the analysis as it has been observed to be associated with household food waste [14], as well as diet quality [15]. A single item was used to assess household income: “What is the total annual income of your household before taxes? Your household income includes income from you and anyone who lives with you who depends on the same income. Be sure to include income from all sources, such as salary and wages, child support, interest, public assistance and pensions.” Response options included: “Less than $10,000, $10,000 to $19,999, $20,000 to $29,999, $30,000 to $39,999, $40,000 to $49,999, $50,000 to $59,999, $60,000 to $69,999, $70,000 to $79,999, $80,000 to $89,999, $90,000 to $99,999, $100,000 to $149,999, $150,000 or more, I don’t know, and I am not comfortable answering this question”. The mid-point for each quantitative category was calculated. Response options were coded as a continuous variable; response items “I am not comfortable answering this question” and “I don’t know” were coded as a non-response. The first three income categories (less has $10,000; $10,000-$19,999 and $20,000-$29,999) were combined and the midpoint of $20,000 was used due to low numbers in these categories.
Statistical Methods
The statistical analysis was performed with SAS (University Edition, Version Studio, SAS Institute, Cary, NC, USA). Linear regression was used to explore the association between HEI-2015 scores and both avoidable and unavoidable daily per capita food waste (total and for each food category). Household income was included in all models. A p-value less than 0.05 was used to establish statistical significance.