Written nutrition guidelines, client choice distribution, and adequate refrigerator storage are positively associated with increased offerings of Feeding America’s Detailed Foods to Encourage (F2E) in a large sample of Arkansas food pantries

Abstract

Background

Millions of food insecure households in the US obtain food from food pantries each year. These foods are often of insufficient nutritional quality.

Objective

Describe the frequency with which Arkansas food pantries offer foods included in Feeding America’s Detailed Foods to Encourage (F2E) framework, and examine food pantry characteristics associated with increased frequency of offering F2E and other foods.

Design

A 27-item cross-sectional online survey assessed food pantries’ characteristics (e.g., storage capacities, number of clients served) and typical food offerings.

Participants/setting

Partnering with five of Arkansas’s six food banks, 764 email invitations were sent to food agency managers across the state. A final sample of 357 food pantries were included in the analyses.

Intervention

N/A

Main outcome measures

The primary outcomes of interest were the frequencies of offering specific F2E and F2E in general. The F2E framework was developed by Feeding America to more accurately assess food banks’ inventories, and its categories (Fruits and Vegetables; Protein; Dairy; and Grains) are generally consistent with MyPlate.

Statistical analyses performed

Descriptive statistics were computed for all food pantry characteristics and frequency of foods offered. Associations between food pantry characteristics and the frequency of offering F2E were examined via multiple linear regression and path analysis.

Results

Only 18.5% of food pantries had written nutrition guidelines, and only 19.3% offered client choice distribution. The F2E most commonly offered were meat/poultry/seafood without breading and not fried (59.6%) and peanut butter (58.2%). The least commonly offered F2E were nuts/seeds with nothing added (3.8%) and low fat/1%/skim cheese (8.2%). Written nutrition guidelines (P<0.001), client choice distribution (P=0.003), and adequate refrigerator storage (P=0.010) were associated with more frequently offering F2E.

Conclusions

This study fills a gap in knowledge by documenting food pantry characteristics that are associated with the frequencies of offering specific types of healthy foods.

INTRODUCTION

In 2017, 11.8% of US households were food insecure.¹ Specifically, 15 million US households reported difficulty providing enough food for the household because they lacked adequate resources.¹ Food insecurity is both a cause and consequence of household economic difficulties,² and it is associated with chronic diseases including hypertension, heart disease, diabetes, and cancer.³

In the US, food insecure people utilize food pantries to access free, unprepared food for offsite consumption.⁴ In 2017, almost 6 million US households obtained food from a food pantry.⁵ However, a systematic review of nine studies showed typical food distributions from food pantries are of insufficient nutritional quality to support a healthy lifestyle.⁶ Additionally, the typical diet of food pantry clients lacks an adequate amount of fruits and vegetables, dairy, and several vitamins and minerals.⁷ This lack of dietary quality, coupled with associations between food insecurity and chronic disease, suggests that improving the nutritional quality of food pantries’ foods could yield health benefits for food pantry clients.

Food pantries differ widely with respect to characteristics that may influence the nutritional quality of foods offered. However, research associating nutritional quality with broad food pantry characteristics (i.e., not tied to any specific intervention) is not widespread. One characteristic that differentiates food pantries is the presence or absence of formal nutrition guidelines. A survey of 1,539 US food pantry directors found that only 20.9% of responding pantries had a written nutrition policy related to food distribution or donations, but published results did not address the scope or effectiveness of these policies.⁸ Food pantries also differ with respect to the extent to which they allow clients to select the foods they take home (versus providing clients with uniform preselected food bags). Client choice of foods has been associated with preserving client dignity,^9–11 promoting self-efficacy,^12,13 and being preferred by clients,^14,15 and it has been linked with positive nutrition-related outcomes in trials of specific interventions (e.g., FreshPlace¹² and SuperShelf¹⁶). However, the extent to which the presence of client choice is associated with the nutritional quality of foods offered is not well specified. Given food pantries’ heterogeneity in operations, policies, and budgets,¹⁷ there are many other food pantry characteristics (e.g., number of clients served, the presence/absence of adequate refrigerator and freezer storage) that may be related to the nutritional quality of foods offered.

A commonality across food pantries is they often purchase food from food banks, which are umbrella organizations that acquire and distribute food to food pantries.⁴ Beginning in December 2017, five Arkansas food banks partnered with researchers from the University of Arkansas for Medical Sciences to identify sustainable ways to help food pantries offer healthier foods to their food insecure clients. Several indicators motivated this partnership. Specifically, the 2015–2017 average prevalence of household food insecurity in Arkansas ranked second highest in the US, at 17.4%.¹ Further, 23.2% of children in Arkansas were food insecure in 2016, ranking third highest in the US.¹⁸ In addition, Arkansas had an extremely high prevalence of health conditions associated with food insecurity, including overweight/obesity (70.4%), hypertension (41.4%), and diabetes (12.2%).¹⁹ The present study represents the first step for the partnership between the food banks and researchers: implementing an online survey of Arkansas food pantries to characterize their food offerings and capacities to offer healthier foods.

As a guiding framework to identify healthier foods, the Arkansas food banks and researchers agreed to use Feeding America’s Detailed Foods to Encourage (F2E) framework.²⁰ In 2012, Feeding America — the US’s largest hunger relief organization — convened a panel of nutrition experts and food bank staff to develop a national framework of nutritious foods recommended to be included in food bank inventories.²⁰ The panel selected these foods in part because they meet the US Department of Agriculture (USDA) 2010 Dietary Guidelines for Americans, including fruits, vegetables, protein, dairy, and whole grains, and because they were generally consistent with USDA’s MyPlate framework.^20–22

Use of the Detailed F2E framework was appealing to the Arkansas food banks because it uses categories that were already familiar in the hunger relief system and focuses on encouraging distribution of healthier foods, rather than banning specific foods. To help their food pantries offer healthier foods to their clients, the Arkansas food banks were interested in identifying food pantry characteristics (e.g., storage capacities, number of clients served per month, proportion of food purchased from a food bank) associated with offering healthier foods. However, previous literature associating specific food pantry characteristics with the nutritional quality of foods offered is limited.^8,23–26

The present study was designed to identify those associations. The study’s first objective was to describe the frequency with which food pantries offered food items included in the Detailed F2E framework. The second objective was to identify food pantry characteristics associated with offering Detailed F2E in general, specific categories of F2E, and other foods not included in the Detailed F2E framework (hereafter referred to as “Other Foods”). Given the limited literature associating food pantry characteristics with nutritional quality of food offered to clients, this study focused on identifying associations, rather than confirming specific predictions or hypotheses.

METHODS

Respondents and Recruitment

The survey’s sole inclusion criterion stipulated that respondents must be food agency staff responding on behalf of a food agency served by one of the five participating Arkansas food banks. All respondents were recruited via emails sent directly from the five food banks. These Feeding America-affiliated food banks serve 67 of the state’s 75 counties and 81% of Arkansas’s food bank-affiliated food agencies. The food banks agreed to participate in recruitment in exchange for the researchers’ agreement to: 1) manage the survey; 2) allow each food bank to supplement the survey with region-specific questions to be sent only to the agencies it serves; and 3) share customized results with each food bank, including statewide summaries and results specific to that food bank’s agencies.

The researchers provided the participating food banks with a customizable template used to invite their food agencies to participate in the online survey. The invitation template allowed the food bank to insert their name, logo, and any additional messaging they wanted to include. Each food bank’s invitation template included a link to a region-specific version of the online survey. Food bank personnel sent the email invitation to their primary points of contact at each food agency they served. The invitation provided a brief overview of the survey and encouraged an “agency leader” at each site to respond to the survey. The food banks sent two reminder invitations to their food agencies: two weeks and four weeks after sending the first invitation. Respondents were not compensated for completing the survey. Data collection occurred from May-June 2018. See Figure 1 for a diagram depicting survey recruitment and respondent flow.

Figure 1.

Flow diagram of the recruitment and participation of 357 Arkansas food pantries in an online survey

Note: Responses to the survey were examined to ensure only one response per agency was included in the analysis. When multiple responses were identified, the research team prioritized the response that had the least missing data, or, if multiple responses were complete, was most recent. Response rate to the survey was 51.4%, ranging from 35.7% to 65.9% across the five participating food banks. Each respondent’s zip code and city name was evaluated to ensure they were within the participating food banks’ service areas.

The institutional review board at the University of Arkansas for Medical Sciences determined this study not to be human subjects research (IRB#228011).

Survey

A 27-item survey was developed to assess the food pantries’ organizational and food distribution characteristics. The researchers collaborated with the participating food banks to develop the items, incorporating three rounds of refinement before a final survey draft was approved via consensus. Food pantry characteristics to be included in the survey were chosen according to the food banks’ expertise; the researchers’ previous work in food pantries in Arkansas,^9,27 Maryland,²⁸ and Minnesota;^16,29,30 and other assessments of food pantries’ characteristics.^31–33

Item development focused on face-valid categorical and ordinal response options with which agency personnel would be readily familiar and could quickly complete. Some items were adapted from existing instruments, including Roadrunner Food Bank of New Mexico’s³⁴ Partner Peer Groups Self-Evaluation and SuperShelf’s¹⁶ Manager Fidelity Tool (both unpublished). Because respondents were not compensated for their time, brief, simple survey items were prioritized in order to facilitate participation of smaller agencies with few to no paid staff.

During each round of refinement, agency relations staff from each partner food bank piloted the draft survey with internal staff and/or partner agency staff. Each round, each food bank returned aggregate feedback related to content and face validity and usability to the researchers and food banks. By the end of pilot testing, most respondents reported completing the survey in fewer than ten minutes.

The survey was administered online via Research Electronic Data Capture (REDCap).³⁵ Like most other communications between the partner food banks and their agencies, the survey was only made available in English. The final version of the survey used yes/no items, Likert-type scales, multiple-response items, and open-ended items to capture responses. Analyses of open-ended items in the present study are limited to the items assessing the number of clients served per month and the proportion of food pantries’ food that is purchased from a food bank. The survey instrument is included as a supplementary file.

Frequencies of Food Items Offered

Food pantries were asked to report the frequency with which they offered each of 24 selected food items, 20 of which were selected from the Detailed F2E framework, according to four categories: Fruits and Vegetables; Grains; Proteins; and Dairy. To provide contrast to the F2E items, four Other Foods were included: breaded or fried meat/poultry/seafood; salty snacks (chips, pretzels, crackers, etc.); prepared desserts (snack cakes, donuts, pastries, etc.); and candy. The Other Foods included were selected because they were regularly offered in food pantries that participated in a previous research project.²⁷ Many of these items (e.g., desserts, bakery, snacks, prepared foods) are outlined in the Detailed F2E framework category “Other Food”, as they do not meet F2E nutrient standards.²⁰

An aggregate F2E frequency score was calculated for each food pantry by summing the responses for the 20 F2E items. Scores ranged from 0 to 80, with higher scores indicating more frequently offering F2E. Cronbach’s alpha for the aggregate F2E frequency score was 0.88, which shows very good internal consistency.³⁶ An aggregate Other Foods frequency score was calculated for each food pantry by summing the responses for the four Other Foods items. Scores ranged from 0 to 16, with higher scores indicating more frequently offering Other Foods. Cronbach’s alpha for the aggregate Other Foods frequency score was 0.62.

Analytic Strategy

Raw data for all survey responses were exported from REDCap directly to SPSS 25,³⁷ and then converted to a STATA³⁸ file for analyses.

Descriptive Method

Descriptive statistics were computed for the food pantry characteristics variables described above. Frequency distributions for all 24 food items offered were also calculated.

Regression Method

Multiple linear regression analyses and generalized structural equation modeling (GSEM) were conducted with STATA 15.³⁸ Statistical significance was determined at an alpha level of 0.05.

Multiple Linear Regression

Multiple linear regression was utilized to identify associations between food pantries’ aggregate F2E frequency score and specific food pantry characteristics (written nutrition guidelines, client choice distribution, adequate refrigerator storage, adequate freezer storage, service frequency, number of clients served per month, and proportion of foods purchased from a food bank). A second regression examined associations between food pantries’ aggregate Other Foods frequency score and these predictors.

Path Analysis

To test an exploratory conceptual model in which food pantry characteristics are associated with frequency of offering specific categories of food, a path analysis was conducted. The model included five latent variables; each latent variable represented an unmeasured construct that captured the propensity of food pantries to offer one of the four categories of F2E (Fruits and Vegetables; Grains; Proteins; and Dairy) or Other Foods. No unidirectional causal relationships were assumed to exist between any latent variables; each latent variable was allowed to covary with every other latent variable.

Each of the five latent variables was measured by at least four manifest (measured) variables. Cronbach’s alpha for the manifest Fruits and Vegetables variables fresh fruits; fresh vegetables; canned vegetables, low sodium (not including beans); canned fruits, low sugar/light syrup; and frozen fruits and/or vegetables offered was 0.74. Cronbach’s alpha for the manifest Grains variables 100% whole grain bread; whole grain cereal; oatmeal/rolled oats (not instant); wild/brown rice; and whole grain pasta offered was 0.79. Cronbach’s alpha for the manifest Proteins variables peanut butter; nuts/seeds with nothing added; canned beans, low sodium; dry beans/lentils; meat/poultry/seafood without breading and not fried; and eggs offered was 0.64. Cronbach’s alpha for the manifest Dairy variables low fat/1 %/skim yogurt; low fat/1 %/skim milk; unsweetened milk substitutes (soy or almond milks); and low fat/1%/skim cheese offered was 0.87. Cronbach’s alpha for the manifest Other Foods variables breaded or fried meat/poultry/seafood; salty snacks (chips, pretzels, crackers, etc.); prepared desserts (snack cakes, donuts, pastries, etc.); and candy offered was 0.62. Because the manifest variables were measured on an ordinal scale, and thus have a non-normal and non-continuous distribution, path analysis was conducted using GSEM — family ordinal and link logit. GSEM is an extension of structural equation modeling and, as a more advanced statistical method, covers models where the response variables are ordinal.³⁹

Goodness of fit was used to evaluate written nutrition guidelines, client choice distribution, adequate refrigerator storage, adequate freezer storage, service frequency, number of clients served per month, and proportion of foods purchased from a food bank as potential exogenous predictors of the five latent variables in the conceptual model.

Missing Data

Proportions of missing data on included variables varied between 4% and 11%. The multiple linear regression analyses were conducted on cases with complete data for the F2E (n=254) and Other Foods (n=285) frequency scores. However, GSEM does not apply a listwise-deletion rule, so observations containing missing values were included in the GSEM analyses (n=299).

Return of Results

In August 2018, the researchers returned a region-specific descriptive overview of the survey’s results to each of the five partner food banks via email. These overviews contained a summary of key findings, graphics, and tables showing statewide results alongside results specific to the food bank’s agencies. Food bank directors reported sharing the results with their staff and communicating the results with funders and other key stakeholders.

RESULTS

Food Pantry Characteristics

Table 1 presents descriptive statistics for the food pantry characteristics variables. Among survey respondents, 18.5% reported having written nutrition guidelines, and 19.3% reported allowing clients to choose their food, rather than receive prepackaged bags or boxes. A quarter of food pantries (25.8%) opened two to four times a week and 13.8% opened five or more days a week. The median number of unduplicated clients served each month per food pantry was 130. The median proportion of total food offered by each food pantry that was purchased from a food bank was 75%. (To address skewness, tertiles were calculated based on valid responses to each of these two items. See Figure 2.)

Table 1.

Characteristics of 357 Arkansas food pantries, as reported by staff member responses to an online survey

	n (%)
Does the food pantry have written nutrition guidelines?a
Yes	60 (18.5)
No	264 (81.5)
Is the food pantry a client choice pantry (i.e., do clients select their foods rather than receive prepackaged bags or boxes with minimal choice)?b
Yes	65 (19.3)
No	272 (80.7)
Does the food pantry have adequate refrigerator storage?
Yes	221 (65.0)
No	119 (35.0)
Does the food pantry have adequate freezer storage?
Yes	257 (76.5)
No	79 (23.5)
Service frequency of the food pantry
No regular schedule, clients call as needed	15 (4.4)
Open once a month	68 (19.9)
Open twice a month	37 (10.9)
Open once a week	86 (25.2)
Open two to four days a week	88 (25.8)
Open five or more days a week	47 (13.8)

Note: One staff member per food pantry responded to the online survey. The denominators for percentages are based on the number of valid responses to each item; thus, number of respondents for any item may not total 357 due to missing data. Percentages may not total 100 due to rounding.

^aResponse options were: “Yes, written guidelines”; “Yes, unwritten guidelines”; and “No.” This variable was dichotomized as 1=“Written guidelines” and 0=“No written guidelines.” The decision to dichotomize this variable was made to emphasize formal written policy and was made before data were analyzed.

^bResponse options were: “Clients select foods from shelves, like in a store”; “Clients make selections via order form”; and “Clients receive prepackaged bags or boxes with minimal choice.” This variable was dichotomized as 1=“Client choice,” which combined selection from shelves and order forms, and 0=“No/minimal choice.” The decision to dichotomize this variable was made before data were analyzed.

Figure 2.

Path analysis from generalized structural equation modeling: exploratory analysis of predictors of the frequency with which Arkansas food pantries offer Foods to Encourage and Other Foods, as reported by staff member responses to an online survey, n=299

Note: *P<0.05; **P<0.01; ***P<0.001. Only statistically significant paths (P) are depicted. Adequate freezer storage had no statistically significant paths to F2E categories or Other Foods, so is not depicted. ξ, represent errors terms. Food pantries were asked to report the number of unduplicated clients served each month and the proportion of foods purchased from a food bank (i.e., by typing a number in a blank). Tertiles were calculated based on valid responses to each of the items. Number of unduplicated clients served per month: lowest tertile=≤72 clients; middle tertile=73–212 clients; highest tertile=≥225 clients. Proportion of Foods Purchased from a Food Bank: lowest tertile=≤50%; middle tertile=55–85%; highest tertile=≥87%.

Frequencies of F2E and Other Foods Offered in Food Pantries

For the aggregate F2E frequency score, the food pantries’ mean score was 34.3 (SD=13.5), with a possible and actual range of 0 to 80. For fruits and vegetables specifically, 56.6% of food pantries reported often or always offering canned vegetables, low sodium (not including beans), and 42.3% reported often or always offering canned fruits, low sugar/light syrup. Fresh vegetables were often or always offered by 38.9% of food pantries, and 30.9% reported often or always offering fresh fruits. Only 16.3% of food pantries reported often or always offering frozen fruits and/or vegetables (Table 2).

Table 2.

Frequency distribution of Feeding America’s Foods to Encourage and Other Foods offered in 357 Arkansas food pantries, as reported by staff member responses to an online survey

“Thinking back across the days you were open during the past 6 months, how often does your agency offer each of the following?”	Never (Approximately 0% of the time)	Rarely (Approximately 25% of the time)	Sometimes (Approximately 50% of the time)	Often (Approximately 75% of the time)	Always (Approximately 100% of the time)
FOODS TO ENCOURAGE
Fruits and Vegetables
Fresh fruits	17.1	33.9	18.0	16.2	14.7
Fresh vegetables	11.5	29.3	20.2	19.9	19.0
Canned vegetables, low sodium (not including beans)	4.8	19.0	19.6	22.9	33.7
Canned fruits, low sugar/light syrup	5.7	26.4	25.5	20.1	22.2
Frozen fruits and/or vegetables	32.0	32.0	19.6	10.0	6.3
Grains
100% whole grain bread	19.9	20.6	23.3	19.9	16.3
Whole grain cereal	8.3	27.2	28.8	20.5	15.3
Oatmeal/rolled oats (not instant)	24.7	32.9	23.8	10.7	7.9
Wild/brown rice	16.3	31.4	23.0	16.3	13.0
Whole grain pasta	16.5	31.7	25.3	17.7	8.8
Proteins
Peanut butter	2.7	12.4	26.7	23.6	34.6
Nuts/seeds with nothing added	43.3	39.2	13.8	1.6	2.2
Canned beans, low sodium	12.7	27.5	29.0	18.1	12.7
Dry beans/lentils	9.8	16.9	20.9	23.3	29.1
Meat/poultry/seafood without breading and not fried	13.7	9.7	17.0	23.7	35.9
Eggs	47.4	27.1	10.6	8.2	6.7
Dairy
Low fat/1%/skim yogurt	46.3	27.4	11.3	10.1	4.9
Low fat/1%/skim milk	35.8	27.2	11.3	15.3	10.4
Unsweetened milk substitutes (soy or almond milks)	54.2	26.7	9.7	5.5	3.9
Low fat/1%/skim cheese	53.5	28.6	9.7	4.9	3.3
OTHER FOODS
Breaded or fried meat/poultry/seafood	22.5	23.1	21.9	18.2	14.3
Salty snacks (chips, pretzels, crackers, etc.)	7.6	29.0	29.0	20.7	13.7
Prepared desserts (snack cakes, donuts, pastries, etc.)	8.4	21.1	23.8	25.6	21.1
Candy	26.4	42.3	21.0	7.6	2.7

Note: One staff member per food pantry responded to the online survey. Foods to Encourage refers to food categories included in Feeding America’s Detailed Foods to Encourage framework.²⁰ Each number in the table refers to a row percentage (i.e., the number of food pantries represented in that table cell divided by the number of total food pantries represented in the row of the table, converted to a percentage). The denominators for percentages are based on the number of valid responses for each row. Percentages may not total 100 due to rounding.

With respect to grains, 36.2% of food pantries reported often or always offering 100% whole grain bread; 35.8% reported often or always offering whole grain cereal; 29.3% reported often or always offering wild/brown rice; 26.5% reported often or always offering whole grain pasta; and 18.6% reported often or always offering oatmeal/rolled oats (not instant).

With respect to proteins, 59.6% of food pantries reported often or always offering meat/poultry/seafood without breading and not fried; 58.2% reported often or always offering peanut butter; 52.4% reported often or always offering dry beans/lentils; and 30.8% reported often or always offering canned beans, low sodium. Only 14.9% of food pantries reported often or always offering eggs, and only 3.8% reported often or always offering nuts/seeds with nothing added.

With respect to dairy, 25.7% reported often or always offering low fat/1%/skim milk, and 15.0% reported often or always offering low fat/1%/skim yogurt. Only 9.4% of food pantries reported often or always offering unsweetened milk substitutes (soy or almond milks), and 8.2% reported often or always offering low fat/1%/skim cheese.

For the aggregate Other Foods frequency score, the food pantries’ mean score was 7.3 (SD=3.3), with a possible and actual range of 0 to 16. Specifically, 46.7% of food pantries reported often or always offering prepared desserts; 34.4% reported often or always offering salty snacks; 32.5% reported often or always offering breaded or fried meat/poultry/seafood; and 10.3% reported often or always offering candy.

Predictors of Overall F2E and Other Foods

For the model with the F2E aggregate frequency score (range: 0 to 80) as the dependent variable, food pantries with written nutrition guidelines — as compared to those without — reported, on average, offering more F2E (P<0.001 ). Similarly, client choice food pantries reported, on average, offering more F2E (P=0.003). Reporting adequate refrigerator storage was also associated with reported offering of more F2E (P=0.010). No associations were found between reported offering of F2E and reported number of clients served per month, proportion of food purchased from a food bank, or having adequate freezer storage (Table 3 Supplement).

For the model with the Other Foods aggregate frequency score (range: 0 to 16) as the dependent variable, there was no association between reports of having written nutrition guidelines and reported offering of Other Foods. However, client choice food pantries reported, on average, offering more Other Foods (P=0.007). Food pantries that reported serving the most clients per month were more likely to report offering Other Foods than food pantries that reported serving the least clients per month (i.e., highest vs lowest tertile; P=0.001). No associations were found between reported offering of Other Foods and reported proportion of food purchased from a food bank, having adequate refrigerator storage, or having adequate freezer storage.

Predictors of Specific F2E Categories and Other Foods

The F2E model was fitted with and without service frequency as a predictor of latent variables. Goodness of fit indices — likelihood ratio test — indicated that the model without this predictor constituted a better-fitting model.

Results from GSEM are presented in Figure 2. All coefficients for the path linking the latent variables to the manifest variables were positive and statistically significant (P<0.05). Among the paths linking the latent variables to the exogenous predictors, the association between reporting written nutrition guidelines and reporting more frequently offering Fruits and Vegetables was statistically significant (P=0.045). Relative to food pantries that did not report client choice distribution, client choice food pantries were likely to report more frequently offering Fruits and Vegetables (P=0.001). Food pantries that reported serving more clients per month were likely to report more frequently offering Fruits and Vegetables than food pantries that reported serving the least clients per month (P<0.001 and P=0.001 for middle and highest tertiles vs lowest tertile, respectively). Food pantries that reported having adequate refrigerator storage were likely to report more frequently offering Fruits and Vegetables (P<0.001).

Food pantries reporting written nutrition guidelines were likely to report more frequently offering Grains (P<0.001). Food pantries that reported client choice distribution were likely to report more frequently offering Grains (P=0.042). Food pantries that reported serving the highest numbers of clients per month were likely to report less frequently offering Grains than were food pantries that reported serving the lowest number of clients per month (i.e., highest vs lowest tertile; P=0.018). Food pantries that reported an intermediate proportion of food purchased from a food bank were likely to report more frequently offering Grains than were those that reported sourcing the least proportion of their food from a food bank (i.e., middle vs lowest tertile; P=0.002).

Food pantries reporting written nutrition guidelines were likely to report more frequently offering Proteins (P=0.023), as were food pantries reporting client choice distribution (P=0.005). Food pantries reporting written nutrition guidelines were likely to report more frequently offering Dairy (P=0.002). Food pantries that reported serving the highest number of clients per month were likely to report more frequently offering Dairy than were food pantries that reported serving the lowest number of clients per month (i.e., highest vs lowest tertile; P=0.005). Food pantries in the middle tertile for the reported proportion of food sourced from a food bank were likely to report more frequently offering Dairy than were those that reported sourcing the least proportion of their food from a food bank (i.e., middle vs lowest tertile; P=0.005). Food pantries that reported having adequate refrigerator storage were likely to report more frequently offering Dairy (P<0.001).

Reporting written nutrition guidelines was not associated with reported offering of Other Foods (P=0.512). No association was found between reporting client choice distribution and reported offering of Other Foods (P=0.081). Food pantries that reported serving the most clients per month were likely to report more frequently offering Other Foods than were food pantries that reported serving the least clients per month (i.e., highest vs lowest tertile; P=0.001).

DISCUSSION

Written nutrition guidelines, client choice distribution, and adequate refrigeration were consistently associated with more frequent offering of F2E. Written nutrition guidelines were only present at 18.5% of respondent pantries, but they were associated with self-reports of more frequently offering F2E in general, as well as all four of the F2E categories (Fruits and Vegetables; Protein; Dairy; and Grains). The proportion of food pantries reporting written guidelines aligns closely with the finding of a 2019 survey that 20.5% of 1,539 US food pantries reported a written nutrition policy, as well as other previous research indicating most food pantries do not have formal nutrition guidelines in place.^8,40,41 Previous research has called for an investigation of how nutrition policies impact the nutritional quality of food in the food bank/food pantry system.^42,43 The present study advances what is known by demonstrating a positive association between the presence of written nutrition guidelines in food pantries and self-reports of more frequently offering healthier foods.

Only 19.3% of food pantries responding to the present survey reported offering client choice distribution; however, client choice distribution was associated with self-reports of more frequently offering F2E in general, as well as the specific Fruits and Vegetables, Grains, and Proteins categories. Many client-level benefits of client choice distribution have been reported previously.^9,10,12–15 However, the present study is the first to our knowledge to associate the presence of client choice distribution with increased self-reported offering of healthier foods across a large sample of food pantries.

The cross-sectional positive associations between client choice distribution and self-reported frequency of offering healthier foods — and between written nutrition guidelines and self-reported frequency of offering of healthier foods — are relevant to the current pantry-based nutrition intervention context. Intervention trials relying on behavioral economics principles to nudge clients to choose healthier foods have shown that when healthier food is available, the interventions can lead clients to select foods with higher nutritional quality.^16,44,45 Client choice distribution is already a prerequisite for many of these interventions,^46,47 but it is up to future research to determine whether or not nutrition guidelines can enhance the effectiveness of behavioral economics interventions.

Adequate refrigeration was reported by 65.0% of food pantries responding to the survey. Perhaps because adequate refrigeration appears more prevalent than written nutrition guidelines or client choice distribution, it has not been a primary focus of prior studies seeking to improve the nutritional quality of foods distributed by food pantries. In the present study, the presence of adequate refrigeration was associated with self-reports of more frequently offering F2E in general, as well as the specific Fruits and Vegetables and Dairy categories. Although adequate freezer storage was more frequently reported than was adequate refrigerator storage, adequate freezer storage was not associated with self-reported frequency of offering F2E in general or the specific categories, suggesting there is a difference between food pantries that have adequate refrigerator storage versus those that have adequate freezer storage. In fact, the issue of refrigerator storage has been mentioned by food bank and food pantry staff as a barrier to providing clients with healthier foods.^23,48–50 To better understand the association between adequate refrigerator storage and more frequently offering F2E identified in this study, future research should examine changes in food pantry offerings as a function of increasing refrigerator storage.

In contrast to written nutrition guidelines and adequate refrigerator storage, client choice distribution was associated with increased self-reported frequency of offering Other Foods (prepared desserts, salty snacks, etc.) in the aggregate, although this association was not significant in the GSEM analysis. Being in the highest tertile of food pantries with respect to the number of clients served per month was also associated with increased self-reported frequency of offering Other Foods. The cross-sectional data in the present study make it difficult to explain why this pattern of associations was found. Future research is needed to explain why client choice distribution is associated with increased self-reported frequency of offering both F2E and Other Foods, and why food pantries that serve the most clients per month report more frequently offering Other Foods, but not F2E.

One might have expected written nutrition guidelines to have a clear negative association with self-reported frequency of offering Other Foods, but it is important to note that written nutrition guidelines were not associated in either direction with offering Other Foods. This could be due to the specific content of the nutrition guidelines (e.g., focusing on obtaining more healthy foods, rather than explicitly restricting less healthy foods). The structure of Feeding America’s F2E framework itself focuses on encouraging the distribution of healthier foods, rather than discouraging less healthy options. In fact, in prior research, food bank and food pantry staff have expressed discomfort at the idea of restricting the types of foods that can be donated.^8,9,42

Limitations

There are several limitations of the present study that are common to survey research in general and, as such, the findings should be interpreted with caution. First, despite a large sample of 357 food pantries across the state of Arkansas, and a 51.4% response rate for eligible agencies, respondents may not have been representative of the entire population of food pantries in the food banks’ service areas. Second, because these data were reported at a single point in time, it is not possible to determine whether specific food pantry characteristics led to specific patterns of offering F2E and Other Foods or instead simply co-occurred with these patterns (e.g., as a function of certain styles of food pantry management). Third, respondents’ self-reported descriptions of their food pantries and estimations of their inventories may have been inaccurate (e.g., due to bias to present their food pantries in a positive light). Survey items were not validated in the field against actual conditions in the food pantries (i.e., research staff did not compare survey responses to actual frequencies with which a sample of food pantries offered F2E). Items were instead developed by food pantry researchers alongside staff from five food banks in order to maximize face and content validity and to ensure the food banks received useful data about their agencies. During survey development, the food banks elicited feedback from their partner food pantries (although there is no record of the specific number of food pantries who provided feedback). Survey methodology relying on expert assessment of content validity is often used in food pantry research and formative evaluations,^{8,17,23,30–32,51,52} and is a necessary step toward characterizing opportunities to improve the nutrition of foods offered in the food bank/food pantry system.

As described previously, the survey development process prioritized brief, simple items in order to facilitate participation of smaller agencies with few to no paid staff. This process emphasized response options with which key agency personnel would be readily familiar and could quickly complete. This approach precluded the inclusion of items intended to characterize any one food pantry characteristic in great detail. For this reason, the survey did not ask about every food category that may appear in a food pantry (e.g., sugar-sweetened beverages, condiments/dressings), but prioritized healthier foods (F2E) as well as less healthy foods (Other Foods) that are commonly found in food pantries in the region. In the case of nutrition guidelines, the survey was designed to identify whether or not the food pantries had written guidelines that addressed nutrition. These types of guidelines can vary greatly in complexity and scope, and future research should investigate the differences of food pantry nutrition policies. The results of the present study have identified characteristics associated with self-reported frequency of offering F2E, but it will require longitudinal research to determine whether implementing these characteristics in food pantries will tend to increase the frequency of offering F2E.

Similarly, the analytic approach for the present study focused on evaluating the associations for each food pantry characteristic while holding the associations of the other characteristics constant. Future research may benefit from adopting an approach that explores how these characteristics co-occur and interact in existing food pantries, and whether additional benefits or detriments for specific combinations of characteristics can be identified. Likewise, future research should explore the effects of implementing these characteristics in existing food pantries. These approaches would provide insight into whether these particular characteristics are associated with actual (rather than self-reported) frequencies of offering F2E and Other Foods, and whether encouraging food pantries to implement these characteristics is likely to lead to offering more F2E.

Conclusions

This study addresses a gap in knowledge by documenting food pantry characteristics associated with the frequencies of offering specific types of healthy foods. Collectively, the results indicate that self-reported written nutrition guidelines, client choice distribution, and adequate refrigerator storage were each positively associated with self-reports of more frequently offering F2E among a cohort of Arkansas food pantries. These findings will generate useful ideas to be evaluated in future projects by the Arkansas partnership, as well as others advocating for changes in food pantry policies and operations in order to increase their offerings of healthier foods. For example, written nutrition guidelines, client choice distribution, and adequate refrigerator storage were reported in a minority of food pantries in the current sample. Future research by the Arkansas partnership will evaluate whether implementing these characteristics results in more frequent offerings of healthy foods.

RESEARCH SNAPSHOT.

Research Question

What characteristics are associated with increased frequency of offering healthier foods in Arkansas’s food pantries?

Key Findings

Written nutrition guidelines, client choice distribution, and adequate refrigerator storage were associated with more frequently offering Foods to Encourage in a large sample of Arkansas food pantries. Specifically, having written nutrition guidelines was associated with more frequently offering fruits and vegetables, grains, proteins, and dairy. Offering client choice distribution was associated with more frequently offering fruits and vegetables, grains, and proteins. Having adequate refrigerator storage was associated with more frequently offering fruits and vegetables and dairy.