Study aim
The aim of this audit protocol is to describe the method to assess the dietary risk of donated food, including the type, amount, nutrition quality, and safety of donated foods at Australian food banks. The protocol could also be applied to food banks in other countries or other organisations within the Australian CFS.
Setting
The setting was a metropolitan branch of an Australian food banking organisation responsible for state-wide distribution of food (referred to as “the Food Bank” hereafter). The Food Bank operates out of a 6000sqm warehouse with six branches located throughout the state. This Food Bank selection was both convenient and purposeful as it is one of the largest charitable food relief providers in any Australian state. The nutrition quality of donated food has not been assessed previously in Australia using an audit approach that is objective, transparent, and rigorous, yet practical, acceptable and replicable is required.
Study design
An audit of all food donated to and procured by the Food Bank was conducted over five consecutive days (May 2022). As the Food Bank is only opens on weekdays, this represents a week’s worth of food. The Standard Protocol Items: Recommendations for Intervention Trials (SPIRIT) checklist was used to guide this protocol.
Audit process
Consultation and collaboration
Consultation with the Food Bank’s operational staff to organise the audit was undertaken from August 2021 to May 2022, including observing warehouse operations. A Memorandum of Understanding between the University and the Food Bank outlined the aims of research, data collection methods, and reporting activities.
Audit approaches
Different audit approaches were considered in an iterative process that spanned several months. A subset of Bowen et al [60] areas of focus relating to ‘acceptability’ and ‘practicality’ was used to gauge the feasibility of different approaches. Acceptability relates to how those involved are predicted to react to the intervention [60]. Practicality is the extent, likelihood and manner in which an intervention can be delivered when resources, time, commitment or some combination thereof are constrained in some way [60]. The feasibility of interrogating delivery receipts, contacting donors for packing slips and auditing existing warehouse inventory, similar to the one-time inventory method described by Caspi [61] were considered. However, receipts lacked specific detail about the types of food donated and contacting donors was considered an unacceptable risk to donor/Food Bank relationships. A one-time stock inventory was considered impractical because of the requirement to use machinery (e.g., boom lift) to access warehouse shelves, and the time needed to be spent in cool storage and freezer areas. An audit of all incoming deliveries to the food bank warehouse over a set period was deemed the most feasible in terms of Food Bank staff burden (e.g., additional duties (acceptability)) and the researchers could work within the flow of the warehouse without disrupting usual operations (practicality). Although the number and nature of deliveries were unpredictable, the delivery window for each day was fixed between 0700 and 1500 when the warehouse was open for deliveries.
Food donations
All donated and procured food is sorted and distributed at the central Food Bank warehouse. The Food Bank receives food donations from growers, manufacturers, major supermarket chains, food service providers (e.g., institutional caterers), meal delivery companies and third-party logistics firms. Food donations are made for many reasons including: surplus stock resulting from over-ordering or under-selling, cancelled orders, changes in weather conditions or buyer preferences; production errors; damaged packaging; or because food is too close to the ‘Best Before’ date (BBD) [62]. The Food Bank permits donated foods that are close to their ‘Use By’ date (UBD) and up to six months past the BBD but food must not have any obvious signs of damage (e.g. broken packaging) or spoilage (e.g. mould). There are no nutrition guidelines or policies determining the types of food donated to the CFS in Australia. Similar to the UK, food banks in Australia are required to collect the donated food at a time and place convenient to donors, often with very little notice [63]. Community members also donate food directly to food banks.
Defining the food donation process
In consultation with the Food Bank’s operations staff, the process was defined in six stages, see Fig. 1. Each stage (steps 1–6) of the food donation/delivery process was defined to develop the audit procedures. Donations arrived as either single (same product) or mixed loads (unsorted, miscellaneous products).
Identifying audit attributes
This audit measured and documented the total weight (kilograms) of all deliveries received over the 5-day period using the total weight of each individual delivery, the total weight of each product within a delivery of mixed load, and the total weight by donor organisation. As the weight of food is the accepted performance measure for food banks, recording the weight of food audited, if repeated, directly contribute to reports of annual turnover.
Donor details and the following product information were collected including for packaged goods: brand and product name, product description and variety (if applicable); and, for fresh or unpackaged food: product description. Shelf-life category (frozen, chilled, or ambient) was also recorded.
The inherent quality of foods was categorized as either, ‘satisfactory’ or ‘damaged’ based on visual observations [27]. Food packaging protects products against contamination and quality loss [64]. Product packaging was categorised as damaged if the: packaging seal had been broken with/without the product exposed; packaging itself has been damaged (e.g., ripped or dented); or there was visible mould or the product was swollen/‘blown’ (an indication that the physical and microbial integrity has been compromised) [64]. Date marks were also recorded and provide a guide as to how long food can be kept before it begins to deteriorate, but still safe to eat (BBD) or becomes unsafe to eat (UBD) [65].
Summary of information audited
Table 1
details the audit attributes collected during the audit at the Food Bank warehouse:
Data collected
|
Description
|
Digital images
|
Of each load and product
|
Date food was received
|
Date (day/month/year)
|
Donation or procurement details
|
Name of the donor or purchase program
|
Product information
|
Brand and product names
Product description and variety
If a whole/fresh/unpackaged food, a description of the product
|
Shelf-life category
|
Frozen, chilled or ambient
|
Weight
|
Total donated per load (kg)
Number of products and pack sizes(g)
Total weight of each product for mixed loads (kg)
|
Product quality
|
Damaged or satisfactory
If damaged, a description
|
Date marking
|
Best Before and/or Use By date
|
Table 1
Summary of information audited
For each audit attribute, corresponding research questions (Table 2) were formulated to inform both the audit process and data analysis stage.
Attribute
|
Research questions
|
Data required
|
Food received
|
• What is the total weight (kg) of food received (procured or donated)?
• What proportion of food received was procured?
• What proportion of food received was donated?
|
Weight of each delivery received
Analysis of weight data
|
Donated food
|
• Who were the donor organisations?
• What proportion of the total weight of donations is attributable to each donor?
• What many types of products were donated?
• How much variety existed across and within donations?
|
Name of donors
Analysis using donor information and weight data
Products assigned to 1 of 22 product categories and 1 of 163 product category groups
|
Nutritional quality
|
• How is donated food categorised within each group of AGTHE?
• How is donated food categorised within each category of the NOVA system?
|
Brand and product names or whole food description, weight of each donation, total weight of donations received
|
Food safety
|
• What proportion of donated food was deemed safe to consume?
• What proportion of donated food was deemed unsafe to consume?
• What were the key reasons for unsafe foods?
• Which donors are responsible for the donation of unsafe food?
|
Weight of each donation, total weight of donations received
Product quality data
Data marking data
|
Table 2. Audit attributes with corresponding to research questions and data collection requirements
Data collection process
A hand-held mobile device was used to take digital photographic images of all incoming deliveries. This method enabled quick data collection [66] within a busy warehouse setting. Food product images have been previously used to document nutritional quality of foods in the CFS [34, 38] and to monitor food environments more broadly [67]. Photographic images of each delivery were manually annotated electronically when taken [name of the donor; weight (kilograms); date marking (best before or use by date)]. Although manually annotating images on a mobile device is time consuming [68], within the busy warehouse setting, the single handheld device was an operationally effective way to capture deliveries. Other details [daily delivery number, donor, number of loads within delivery] were documented on paper to help keep track of all deliveries.
An audit framework (Table 3) was constructed using Microsoft Excel (Version 2019, Redmond, Washington, USA) profiling all donation and product attributes, enabling the methodical entry of each product photograph of each delivery once the audit was completed.
Table 3
Date (DD/MM/YY)
|
Delivery #
|
Donor/
Purchase
|
Type of load
|
Product category
|
Product category group
|
Brand name
|
Product name
|
Product variety
|
Whole foods
|
Total amount donated (kg)
|
Pack size (g)
|
No. products/ pallets
|
Product quality
|
Product quality notes
|
Shelf-life category
|
Date marking
|
(Free text)
|
Donor/ Purchase
(Select)
|
Single/
Mixed (Select)
|
e.g., “Frozen Food”
(Select)
|
e.g., “Frozen fruit” (Select)
|
As written on packaging
(Free text)
|
Describe (Free text)
|
(Free text)
|
(Free text)
|
(Free text)
|
Satisfactory/ Unsatisfactory
(Select)
|
Describe if unsatisfactory
(Free text)
|
Ambient/
Chilled/
Frozen
(Select)
|
UBD/
BBD (Select)
|
Specify date (Free text)
|
Pilot test
The food donation audit process (Fig. 1) was reviewed with the Food Bank’s operations staff. Feedback led to the incorporation of the type and weight of the various delivery vessels (e.g., wooden pallets, cardboard produce bins, plastic ‘mega’ bins, metal cages) to the documentation. Researchers piloted the data collection tool on several incoming deliveries and the methodology was deemed appropriate for data collection because it was the most operationally effective way to capture delivery details. The process of data extraction was trialled using a sample of the pilot photos and the utility of the audit framework established.
Data collection
Two researchers (SM and MC) and two research assistants recorded deliveries on site in May 2022 for five consecutive days between 0700 to 1500. Data collection commenced at Step 3 (Fig. 1) and each delivery was sequentially weighed after being unloaded. Pre-determined weights of the delivery vessel were subtracted from the total weight. If donations arrived in a vessel where the weight was unknown, the delivery would be unpacked, then the vessel weighed separately.
One researcher took the images of the load and the other verbally confirmed the weight. Each image was then annotated [name of donor, total (gross) weight and date marking information]. Next, additional photos were taken of the product [brand and product names, variety, and pack/product size] and the total number of packs/products donated added. This was used to determine the total (net) weight of donations during the data analysis stage. Single product loads (deliveries of the same product) were typically transported to storage (Fig. 1 – Step 4). Mixed loads of unsorted, miscellaneous products were manually sorted so that each product could be counted (Fig. 1 – Step 5). Each unique product within each load was photographed to capture key product details.
Frozen mixed loads were unable to be sorted as the time required to sort, record and repack each product presented a food safety risk. Only details of the donor and total weight were recorded. Damaged products data collection required additional photographs to document the damage.
Warehouse staff were notified to dispose of items that were deemed unsafe for consumption, e.g., where packaging was ‘blown’, bloated or spoiled, leaking, crushed or obviously contaminated.
For quality control, researchers reviewed the photographs and notes at the end of each day to ensure every delivery had been accounted for. All photographs were date and time tagged by the device used. This enabled an assessment of efficiency of the audit methodology.
Data extraction and analysis
A computer filing system was established with folders for each day of data collection and sub-folders for each delivery, organised chronologically by number and identified by donor name and product photographs were filed accordingly. The data on the annotated photographic images was extracted for each product and for each delivery according to categories in Table 3. Pre-coded responses were used to specify the type of load [single or mixed], product quality, shelf-life category and date marking type. Pre-coded product categories were assigned, [category (e.g., ‘Beverages’) and sub-category (e.g., ‘Carbonated Drinks’)].
Donor name, brand name, product name, product variety, whole food description, and product quality description and date marking all used free text. Quantities of each delivery was determined through the annotated photographic images, with the total (net) weight for single and mixed product loads calculated. The paper-based notes were used to confirm all delivery details.
Missing product information was obtained from the relevant food manufacturer website, visits to a local supermarket or contacting the food manufacturer directly, as sometimes important data was obscured on an image. For quality control, SM completed data entry, MC checked 10% of entries, and any discrepancies were discussed then reviewed by RG.
Dietary risk assessment
Nutritional quality
The nutritional quality of audited products was assessed. Each product was categorised by two systems: 1) Food groups according to the Australian Guide to Healthy Eating (AGTHE) [69], and, the 2) level of food processing according to the NOVA classification [70]. The AGTHE, based on the Australian Dietary Guidelines (ADG) classifies food into the recommended five food groups (1) grain/cereal foods, 2) vegetables and legumes/beans, 3) fruit, 4) lean meat, poultry and alternatives, and 5) dairy and alternatives) and discretionary food (e.g. food and beverages high in fat, added sugar and/or salt) [69]. NOVA classification is according to the extent and purpose of processing (from unprocessed (Group 1), processed culinary ingredients (Group 2), processed (Group 3), to ultra-processed (Group 4)), with ultra-processed foods associated with dietary health risk [70]. The AGTHE is Australian Government’s food selection guide based on the recommendations of the ADGs. The AGTHE presents the serving size and proportions of food group foods recommended for daily consumption [69], and incorporates variety and nutrition adequacy recommendations [71]. Based on individual dietary recommendations, the types and proportions of food groups can be applied to food service, and in this case, assessment of ‘apparent consumption’ of the food supply. NOVA is a food-based classification system, informed by evidence of a food’s structure or composition, and associations with health outcomes [70]. Utilising both these systems considers the appropriateness of products in terms of both the role of food processing and dietary patterns on health outcomes [72, 73]. Nutrient-based classification systems were not used due to the reductionist focus on individual nutrients [71], however, the concept of nutrients ‘at-risk’ is incorporated into the development of the AGTHE. Importantly, it is the food supply, rather than single nutrients that are being assessed.
SM conducted the initial assessment, MC checked 10% of entries, with any identified discrepancies discussed, and reviewed by CEP.
Food safety assessment
Food safety is an important public health priority and the Australian Dietary Guideline #5 ‘Care for your food; prepare and store it safely’ [69] highlights the importance of food safety in terms of dietary risk. The proportion of damaged or unsafe products was determined as a proportion of the total number of entries rated as ‘unsatisfactory’ for product quality. Assessment was based on visual observations (the current practice in the CFS) [27]. Products deemed ‘unsatisfactory’ were categorized into 1) visibly damaged and rotting (e.g., food appears blown, swollen or has visible mould), 2) beyond date marking (past UBD or > 6 months BBD), 3) broken or damaged packaging with product exposed, 4) broken or damaged packaging/label with product unexposed; or 5) other food safety concern (e.g., product recall).
Data analysis
The audit comprised 1500 photographic images collected over 74 deliveries with only two deliveries procured over five days. Donors included supermarkets (n = 27), manufacturers (n = 11), the community (n = 9), growers (n = 8), logistics/distribution companies (n = 9), other food retailers (n = 3), meal delivery providers (n = 3), and inter-Food Bank branch donations (n = 2). The total weight of deliveries (procured and donated) was 108,509kg, with donations accounting for 79% (86,050 kilograms) of the total weight of food received. Data analysis will be guided by the research questions in Table 2. All data will be entered into Microsoft Excel (Version 2019, Redmond, Washington, USA) and descriptive statistics such as frequencies applied.
Assessing these attributes against the criterion developed to answer the research questions will form the development of a tool to assess the dietary risk of donated food at Australian food banks, similar to FODR [26]. The tool will transcend the current weight-based performance indicator to increase the performance focus on a client-centred outcome by incorporating measures to assess dietary risk. The assessment of the apparent dietary quality and food safety of the donated food supply will better indicate performance against the definition of food security, and be used to initiate a Food Bank QIAS with transparent policy measures. The audit protocol was developed with quality improvement as the main intent, and this will be incorporated into development of the risk assessment criterion in the tool.