Attaining and maintaining a healthy diet is considerably difficult for the millions of people in the USA who experience food insecurity, defined by the US Department of Agriculture as an economic and social condition of limited or uncertain access to adequate food^{(Reference Coleman-Jensen, Rabbitt and Gregory1)}. The likelihood of experiencing food insecurity is unevenly distributed across the population. As a product of historical and modern policies and practices, African American, Latino and Native American individuals experience food insecurity at higher rates than non-Hispanic white adults^{(Reference Coleman-Jensen, Rabbitt and Gregory1,Reference Nikolaus, Johnson and Benally2)} . The US government has several efforts, such as the Supplemental Nutrition Assistance Program (SNAP)⁽³⁾, which attempt to address food insecurity. However, some households experiencing food insecurity do not meet eligibility criteria for federal programs, or for those who do, the benefits are at times insufficient to address the households’ food needs^{(Reference Waxman, Joo and Gonzalez4)}. In response to these gaps in the social safety net, non-governmental private organisations – broadly termed the charitable food system – have arisen to distribute foods to individuals with unmet needs.

The charitable food system in the USA is composed of organisations with storage warehouses (i.e. food banks) that often accept all types and sizes of food donations, purchase large quantities of food and receive food from federal programs. Food banks then re-distribute the food through consumer-facing sites (e.g. food pantries, shelves, soup kitchens, programs, etc.). Consumer-facing sites then make groceries or hot meals available to individuals (commonly referred to as neighbours, guests or clients). This system serves individuals and households in immediate need of food, often with minimal or no screening requirements, and is increasingly providing additional services (e.g. SNAP application assistance, referral to community health services) aimed at addressing systemic and longer term needs^{(Reference Martin5)}.

Developed largely in the 1980s in the USA, this system was originally referred to as the ‘emergency food system’ to reflect the short-term supply of food distributed to households^{(Reference Poppendieck6)}. However, there is increasing evidence of long-term use among food pantry clientele^{(Reference Kaiser and Cafer7,Reference Kicinski8)} . Today, the largest non-profit network of food banks in the USA, Feeding America, includes 200 food banks and 60 000 food pantries and meal programs across the country to support the needs of individuals and families⁽⁹⁾. An estimated 20 % of households in a nationally representative sample were served by the charitable food system in 2020^{(Reference Waxman, Gupta and Gonzalez10)}. Given that the system serves individuals at high risk of food insecurity^{(Reference Kaiser and Cafer7,Reference Liu, Zhang and Remley11,Reference Gundersen, Engelhard and Hake12)} and provides a non-negligible portion of users’ diets^{(Reference Liu, Desmond and Wright13–Reference Mousa and Freeland-Graves15)}, the charitable food system is an opportune setting for nutrition-related programming, as well as efforts to meet individuals’ cultural food needs and preferences.

There are notable critiques of the charitable food system’s historical and, in some cases, current efforts to provide services that promote food security and nutrition. Assessments of food pantry inventories have indicated shortfalls of key nutrients and food groups^{(Reference Simmet, Depa and Tinnemann16)}. However, it is important to note this low quality of foods may be a reflection of the broader US food system and supply priorities, and a by-product of the charitable food system’s reliance on donated foods. Though actors in the charitable food system increasingly recognise health promotion, equity and nutrition as important – if not vital – parts of a food bank’s role, published estimates suggest a minority of food banks have formal nutrition policies^{(Reference Wetherill, White and Seligman17)}. Nutrition policies are one tool for food banks to develop decision markers for food donations, purchasing or data collection.

Food banks and food pantries face the challenge of persistent variability in the quantity, quality and type of foods received through donations. This variability has consequences for the staff and volunteers needed to handle and process donations^{(Reference Millar, Parizeau and Fraser18)}. Stakeholders have noted issues of safety and appeal of perishable foods received for distribution, which may be considered inedible and necessitate disposal by food banks, pantries or clients^{(Reference Dave, Thompson and Svendsen-Sanchez19)}. Cultural relevance and usability of items distributed are also of importance. Surveys of clients have indicated that some food pantries do not have food their families like, or they do not know how to prepare the available foods^{(Reference Dave, Thompson and Svendsen-Sanchez19–Reference Greer, Cross-Denny and McCabe21)}.

Despite the barriers, there are many ongoing efforts to improve the charitable food system’s policies, systems and environments (PSEs) with the aim of supporting individual and community nutrition and health. Much of the published work in this area has focused on initiatives in food pantries, as the consumer-facing side of the system. These initiatives have included efforts such as distribution of suggested donation lists^{(Reference Nikolaus, Laurent and Loehmer22,Reference Nikolaus, Kownacki and Darvesh23)} , coordination with food policy councils^{(Reference Remley, Kaiser and Osso24,Reference Stluka, Moore and Eicher-Miller25)} , developing nutrition-focused policies, transitioning distribution from traditional pre-packaged boxes to client-choice models^{(Reference Nikolaus, Laurent and Loehmer22–Reference Remley, Kaiser and Osso24)}, incorporating wrap-around social services^{(Reference Remley, Kaiser and Osso24,Reference Martin, Wu and Wolff26)} , partnering with the health care system^{(Reference Reinoso, Haut and Claffey27)} and reorganising how items appear to ‘nudge’ clients towards healthier items^{(Reference Nikolaus, Laurent and Loehmer22–Reference Remley, Kaiser and Osso24,Reference Wilson28)} , among others^{(Reference Nikolaus, Laurent and Loehmer22,Reference Nikolaus, Kownacki and Darvesh23,Reference Martin, Wu and Wolff26)} . These initiatives – many of which began within the last decade – focus on making healthier choices easy and accessible^{(Reference Fleischhacker, Parks and Yaroch29)} and are coupled with established nutrition education programming^{(Reference Guthrie30)}.

Though not as widely documented in the literature, changes to practices and policies in food pantries have coincided with upstream initiatives undertaken by food banks^{(Reference Campbell, Webb and Ross31)}. In efforts to track and improve the nutritional quality of food banks’ inventories, a variety of nutrition ranking systems and guidelines were developed^{(Reference Battista Hesse, Peachey and Wang32–Reference Martin, Wolff and Callahan34)}. In 2019, a national panel of experts convened by Healthy Eating Research (HER; a national program of the Robert Wood Johnson Foundation) developed the HER Nutrition Guidelines for the Charitable Food System to promote consistent and approachable nutrition classifications for diverse charitable food settings across the USA.^{(Reference Schwartz, Levi and Lott35)} In addition to these national efforts to rank and shift the nutritional composition of food banks’ inventories, case studies have detailed how individual food banks have implemented incremental changes to policies and practices to focus on nutrition and well-being of their community members^{(Reference Swords36,Reference Rowland, Mayes and Faitak37)} . A food bank in New England modified the ordering system their food pantry agency members interacted with and found that the introduction of nutrition ranking information to this system resulted in significant increases in the nutritional quality of monthly orders^{(Reference Martin, Xu and Schwartz38)}. Another food bank in New York state assessed the optimal combinations of days and volunteers to increase the amount of fresh produce they could glean from local farmers to supplement their inventory^{(Reference Sönmez, Lee and Gómez39)}. As of 2021, 20 % of the food banks in the Feeding America network reported using the HER Nutrition Guidelines for the Charitable Food System (J. Hager, personal communication, May 4, 2022). Though the HER nutrition ranking system provides a method for food banks to assess the nutrition quality of their food inventory, there have been no published methods to capture the range of supplementary practices and policies that food banks have been using to promote nutrition and health.

The current study was undertaken to address the need for a method to assess, measure and track the different strategies used by food banks to promote health and nutrition through their practices and policies. This work was started by staff affiliated with the University of Illinois Extension who are responsible for implementing SNAP-Ed in the state of Illinois and who collaborate with food banks interested in updating policies and practices to better promote nutrition and health. The objective of the current study was to develop and evaluate a novel assessment tool that could be used by food bank staff and partnering community-based professionals, such as Extension staff, in diverse settings to evaluate efforts to improve the promotion of health and nutrition in food bank settings. The study was undertaken with the expectation that the final tool could be used by SNAP-Ed implementing agencies, such as Extension and other community-based health professionals, as well as food bank staff and volunteers.

Results

Initial development, iterative review and revisions

The FB-HANA development team produced a list of potential strategies and related objectives for the tool. This list was shared during meetings and via e-mail with seven stakeholders with experience working with the charitable foods system to provide feedback. The purpose of the tool was clarified during this phase, with the team deciding the goal of the assessment was to ‘provide perspective on ways food banks promote health and nutrition through policy, system and environmental approaches.’ The HER Nutrition Guidelines for the Charitable Food System were published during this phase^{(Reference Schwartz, Levi and Lott35)} and strategies focused on equity were added after the draft of the tool was compared against the guidelines. The preliminary list of strategies was organised into eight objectives (Table 1). Each scored strategy was structured as a statement which the assessor would answer yes or no to for ease of completion. To support programmatic planning, additional information about the food bank, such as participation in a formal network, special programs and reach, sources of food and funding and external conditions were collected (but not scored) on the FB-HANA.

Table 1 Objectives and strategies in the food bank health and nutrition assessment (FB-HANA)

Note. CI = confidence interval; ICC = intra-class correlation coefficient.

ICC estimates for inter-rater reliability were produced from one-way random-effects models based on 3–5 raters at 11 food banks.

ICC estimates for test-retest reliability were produced from mixed effects models based on two assessments completed by 3–5 raters at 11 food banks.

Pilot testing

All Extension (n 2) and food bank (n 2) staff who participated in pilot testing provided feedback via interviews or e-mail correspondence. Overall, the FB-HANA was considered useful with good flow between sections and feasible to complete. Staff estimated the FB-HANA took approximately 70 min to complete but shared that almost half of this time was dedicated to finding the documents needed to provide accurate information on food and funding sources for unscored components of the FB-HANA. To improve efficiency of time used, suggestions to clarify what the assessment entails, who in the food bank should complete it and the documents which may be needed were incorporated into the cover page of the FB-HANA. Feedback led to additional clarification of terms, reasons for some data collection and an explicit acknowledgement on the FB-HANA cover page that not all strategies would be relevant for each food bank. At the end of the pilot test, the food bank staff indicated interest and eagerness to use the results to begin planning their next steps; given this, the FB-HANA was placed in a web-based surveying platform (Qualtrics), formatted for ease of completion on mobile or portable computing devices (i.e. tablets or laptops) and structured to auto-generate a report of results for users immediately upon completion.

Content validity assessment

The eight content validity reviewers included four food bank staff, one Extension staff, and three academic researchers. All reviewers rated the FB- HANA overall and objectives 1–3 and 5–8 as relevant (3 or above on 4-point scale), far exceeding the 80 % threshold required. Objective 4 was rated as relevant by most reviewers (88 %), still meeting the necessary threshold for content validity. Qualitative comments from reviewers resulted in re-ordering of objectives and strategies as well as revisions to strategy and objective wording. As one example, a strategy focused on provision of microwave recipes was updated to more general phrasing that recipes provided ‘require minimal cooking equipment.’ Additionally, four new strategies were added to increase the number of relevant practices and policies assessed.

The final version of the FB-HANA was produced after incorporating content validity feedback. All sixty-five strategies were answered with binary response of options of ‘Yes, this strategy is present at the time of the assessment’ or ‘No, this strategy is not present at the time of the assessment.’ These response options are not sensitive to strategies which are in development or generally present, but not on the day of assessment. Yet, the restrictive binary option of yes or no was selected to maximise ease of use and reliability. The full FB-HANA is freely available online on a web platform which incorporates display logic⁽⁴²⁾, and a list of the individual strategies is shown in a Supplemental Table. Of note, some strategies are organised as ‘sub-strategies’ and are displayed only if a preceding question was answered to indicate the ‘sub-strategies’ were relevant. For example, if the food bank indicates they do not have a written nutrition policy, the following twelve strategies under objective 3 are not displayed (and subsequently unaffirmed). Finally, the FB-HANA development team decided to incorporate a definition of nutritious into the instructions to create a more uniform understanding among assessors. If the food bank had a nutrition ranking system in place, this was to be used to inform the interpretation of the term nutritious used throughout the FB-HANA; if the food bank did not have a ranking system in place, items categorised as green by the HER guidelines^{(Reference Schwartz, Levi and Lott35)} were used as the reference for what was to be considered nutritious.

Inter-rater and test-retest reliability assessment

All eleven food banks invited to participate agreed to contribute to the reliability assessment. Descriptive characteristics of these food banks based on food bank staffs’ selections for non-scored elements of the FB-HANA during the first assessment are shown in Table 2. A total of forty individuals participated as raters. There were twenty-six Extension staff serving as raters with a mean of 2·4 (0·5 sd) Extension staff completing ratings per food bank. There were fourteen food bank staff completing assessments, with a mean of 1·3 (0·5 sd) staff per food bank. Each food bank staff only assessed their own respective food bank, whereas Extension staff rated a range of 1–7 unique food banks (mean = 2·2 and sd = 2·1).

Table 2 Characteristics of food banks who participated in the inter-rater and test-retest reliability phase for the evaluation of the food bank health and nutrition assessment (FB-HANA)

Note. Estimates rely on data provided by the food bank staff at the initial assessment. In the case that >1 food bank staff member was present at the initial assessment and responses were not identical, the last response recorded was used.

The inter-rater reliability of the overall FB-HANA among all raters met the threshold for excellent reliability (Table 1), and the inter-rater reliability for individual objective scores indicated good to excellent reliability. When ICC estimates for inter-rater reliability of the overall FB-HANA score were produced from a one-way random effects model restricted to the two to three Extension staff at the initial assessments, the ICC increased to 0·96 with a narrower 95 % CI of 0·80, 0·98 (data not shown). The second assessment was completed at all food banks 3 to 5 weeks (mean = 3·9, sd = 0·6) after the initial assessment. The test-retest reliability of the overall FB-HANA when comparing first and second assessments from all raters met criteria for excellent reliability (Table 1). Test-retest reliability for individual objectives varied from moderate to excellent reliability, with the lowest ICC produced for Objective 1: Food Bank Identifies and Responds to Nutrition Needs of Communities Served and the highest produced for Objective 3: Food Bank Adopts, Implements and Shares Nutrition Policy.

A second mixed-effects model with a dummy variable indicating affiliation as Extension or food bank added was built to compare how scores differed by assessor type. Extension staff produced significantly lower FB-HANA scores (β = −1·28; 95 % CI (−2·45, −0·11); P = 0·032) than their food bank staff counterpoints. Additionally, the ICC estimate of test-retest reliability was marginally higher (ICC = 0·82) with a narrower 95 % CI of 0·65, 0·92.

Discussion

The objective of the current study was to develop and test a tool that could be used by food bank and community-based nutrition professionals to assess current practices to promote nutrition and health through PSE initiatives and monitor change over time. Findings indicate that the resulting tool, the FB-HANA has strong content validity as well as good to excellent inter-rater and test-retest reliability. Results indicate the FB-HANA can be employed directly by food banking professionals, as a self-assessment, or by external staff, such as Extension (or other SNAP-Ed implementing) staff, to guide the provision of technical assistance.

The FB-HANA can be used to initiate preparation and planning to promote nutrition and health as well as to supplement existing resources. Many of the food banks who contributed to reliability testing intended to use the results to inform organisational plans. These plans included the development and adoption of nutrition policies, distribution of surveys to staff or clients and initiation of a nutrition advisory board. Extension staff are using the assessments conducted for the reliability phase as baseline measures to evaluate the implementation of policies and practices. Results suggest these evaluations could be completed by Extension or food bank staff, given the strong test-retest reliability for the FB-HANA regardless of assessor type. However, it is worth considering that food bank staff systematically scored their food banks higher on the FB-HANA than Extension staff. Future research, particularly qualitative studies, might consider exploring what contributes to these scoring differences as well as the unique approaches that Extension and food bank staff take in planning PSE initiatives.

Planning for PSE initiatives, regardless of whether they are lead by Extension or food bank staff can be further supported by Feeding America’s Nutrition in Food Banking Toolkit⁽⁴³⁾ as well as the published case studies of other food banks’ efforts to implement changes in practices^{(Reference Swords36,Reference Rowland, Mayes and Faitak37)} . FB-HANA evaluations and subsequent planning might also support and supplement downstream efforts to assess and improve the consumer nutrition environment in food pantries^{(Reference Nikolaus, Laurent and Loehmer22,Reference Nikolaus, Kownacki and Darvesh23)} . In a Minnesota sample, food banks comprised an average of 50 % of food pantries’ inventories, but this varied from 3 to 99 % in the sample^{(Reference Nanney, Grannon and Cureton44)}, so complementary efforts in both food banks and food pantries are warranted.

A clear insight from the FB-HANA development process was the need for recognition and assessment of who is being served, and by whom, within the charitable food system. At the national level, young adults, women, non-Hispanic Black adults, Hispanic adults, low-income households, adults with a disability and single parents are among those most likely to need the supports of the charitable food system^{(Reference Waxman, Joo and Gonzalez4,Reference Waxman, Gupta and Gonzalez10)} . However, there is geographic and cultural variation in characteristics of individuals of families seeking support from the charitable food system, therefore food banks and pantries are increasingly striving to reflect the cultural composition and diversity of their respective communities among their staff and volunteers as well as with the food and services offered. The need for community assessment, tailoring to community needs and representation of clientele in decision making are recommended in the Nutrition in Food Banking Toolkit⁽⁴³⁾. Thus, these considerations were incorporated into numerous strategies in the final FB-HANA. Assessments completed by food bank staff at the initial reliability assessment indicated that over half of food banks identified the diet-related conditions of concern (55 %) or sought feedback about preferred foods (73 %) among individuals served by their member agencies, but few had advisory boards which included current or former recipients of charitable food assistance (18 %; data not shown).

Another important insight from the FB-HANA development was the common practice of food banks connecting to a variety of organisations in the food system and larger community. Though speculative, food banks which have numerous, diverse connections in the community may be able to provide more variety in their food and services as well as have greater capacity to restrict or reject donated goods from a single donor source. Diverse connections may position food banks to be more nimble and adaptable to surges in charitable food reliance, such as the spike in need in 2020 as a result of the COVID-19 pandemic^{(Reference Waxman, Gupta and Gonzalez10)}. Future climate or public health emergencies may cause similar rapid increases in charitable food system reliance. Food banks may be positioned to adapt to disruptions in the food system by adopting a nutrition policy, which can guide the use of funds to purchase foods. These hypotheses should be tested in future observational studies assessing how food bank characteristics relate to PSE initiatives.

Notably, there are factors upstream of the food bank which can impact the food bank’s ability to promote nutrition and health. For example, national and state-level policies that impact the ability of farmers to sell or donate produce can act as an incentive or deterrent to improving the charitable food system’s quality of foods^{(Reference Osland and Baker45)}. Gleaning un-harvested foods from farmers’ fields is not commonly reported by food banks^{(Reference Millar, Parizeau and Fraser18)}, but food banks expanding their presence and connection to farmers, gardeners and other community members can help optimise the ability of gleaning to contribute nutritious fresh produce^{(Reference Sönmez, Lee and Gómez39)}. Though challenges to transporting, storing and distributing perishable foods in the charitable food system will need to be taken into consideration^{(Reference Osland and Baker45–Reference Levi, Schwartz and Campbell47)}.

Though tools like the FB-HANA and the corresponding instrument for food pantry settings, the Nutrition Environment Food Pantry Assessment Tool^{(Reference Nikolaus, Laurent and Loehmer22)}, have a role in supporting the nutrition and health promotion in the charitable foods system, there are inherent issues related to justice and equity within the system. For many, there is stigma associated with eating donated food, though still healthy and safe, which otherwise may have been discarded^{(Reference Millar, Parizeau and Fraser18)}. Volunteers often far outnumber staff in organisations within the charitable food system^{(Reference Mousa and Freeland-Graves15)}, which can result in unstable or unpredictable labour. Further, food pantry clientele are often not represented among leaders in the charitable food system^{(Reference Cooksey Stowers, Marfo and Gurganus48)}. This may, in part, explain why results of a recent qualitative study found that stakeholders wrongly perceived the charitable food system as contributing only negligibly to clients’ diets overall^{(Reference Cooksey Stowers, Marfo and Gurganus48)}. In response to these structural inequities, some food banks have shifted to community-hub models where redistributing food is supplemented with additional activities, such as a Speaker’s Bureau which focuses on leadership development and advocacy among food pantry clientele^{(Reference Swords36)}, incorporation of a milk bank, policy change advocacy and benefits enrollment support^{(Reference Strickland and Whitman49)}. Though the research team incorporated many of these innovations into strategies in the FB-HANA, additional efforts outside of the charitable food system are likely needed to address food access inequities.

Limitations

The current research study is not without limitations. The sample of food banks who participated in reliability testing for the FB-HANA were limited to four midwestern states. These states may employ policies and practices which are systematically different than those used by food banks in other regions. Fortunately, content validity reviewers had experience with food banking practices in states outside this region, which may have resulted in a greater variety of strategies being added to the final FB-HANA prior to reliability testing. Nonetheless, future studies which draw data from food banks across the USA would be valuable to characterisze practices in different settings and could provide valuable data to assess floor or ceiling effects within the measure. Though the overall FB-HANA scores had evidence of excellent reliability, individual objectives’ reliability estimates were lower. This variation in reliability was likely related to the lack of formal training needed for the FB-HANA and may improve with additional training. One final limitation is that results do not provide insight into how policies and practices may relate to downstream nutrition or health effects among clients of the charitable food system. Additional research will be needed to investigate whether food banking policies and practices translate to outcomes among food pantry clientele.

Fig. 1 Five phases and timeline of food bank-health and nutrition assessment (FB-HANA) development and evaluation