Plate Waste in USDA Summer Food Service Program Open Sites: Results from the Project SWEAT Sub-Study

ABSTRACT The study objectives were to pilot approaches to examine plate waste and explore plate waste patterns by delivery style (pre-assembled trays versus multi-offering tray line with Offer Versus Serve (OVS)) at United States Department of Agriculture (USDA) Summer Food Service Program (SFSP) sites. Mean plate waste in the USDA SFSP was highest for vegetables (70.9% wasted), followed by meat/meat alternate (56.0%), fruit (47.9%), grains (48.9%), and milk (41.0%). The amount of food wasted was consistently higher when meals were delivered on pre-assembled trays versus multi-offering tray line with OVS. Findings from this study provide initial insights into overall plate waste patterns at USDA SFSP sites and potential implications for waste by delivery style.


Introduction
Food insecurity, defined as "whenever the availability of nutritionally adequate and safe foods or the ability to acquire acceptable foods in socially acceptable ways is limited or uncertain, 1 " is an ongoing public health crisis in the United States (U.S.).Low-income households with children are at particularly high risk. 2 In 2021, 10.2% of households in the U.S. identified as food insecure, and the rate was higher for households with children (under the age of 18; 12.5%). 2 Additionally, other groups for which rates were higher include households with reference persons who are non-Hispanic Black (19.8%), Hispanic (16.2%), with incomes below 185% of the poverty threshold (26.5%), and located in principal cities (12.2%). 2 While not all those who experience food insecurity live in poverty and not all who live in poverty experience food insecurity, these two situations often co-exist. Children who are Black and Hispanic children are disproportionately affected by poverty as a result of systemic racism and discrimination, and are thus significantly more likely to experience food insecurity than White children. 3,4 While White children have not experienced the same social disadvantages as Black and Hispanic children, there are still many White children suffering from poverty and food insecurity. 5 The child burden of food insecurity in the United States remains a crisis, particularly given the short-(hunger, headaches, stomach pains, poor diet quality, increased hospitalization admissions) and long-term health consequences (aggression, anxiety, depression, and developmental outcomes). 6 Federal food assistance programs were established in the U.S. during the mid-1900s to address hunger and food insecurity. Child-specific nutrition programs include the United States Department of Agriculture (USDA) School Breakfast Program (SBP), National School Lunch Program (NSLP), and the Summer Food Service Program (SFSP). 7 The SFSP was established to address the summertime window of risk due to the loss of access to healthy and nutritious foods provided by schools. 7 The execution of the SFSP includes sponsors, sites, and vendors. Sponsors are organizations that operationalize the SFSP at sites. The sites are the physical location where the SFSP meals and snacks are distributed. Vendors are the organizations that, through the coordination of the sponsors, provide the meals and snacks to the sites. SFSP sites can be open or closed. Open sites are often referred to as "drop-in" sites where children do not have to register to receive food; whereas closed sites require advance registration and may include a fee to fund programming at the site.
Meals served at SFSP sites are required to meet nutritional standards to assure proper nutrition for participants. 8 Sites must offer at least minimum amounts of four food components, including two fruits/vegetables (this can be in the form of one fruit and one vegetable, two fruits, or two vegetables), one grain, one meat/meat alternate, and one milk. 9 For meals to be reimbursable, participants must take all four food components with each lunch. However, school food authority site operators have the option to use an alternative approach called Offer-Versus-Serve (OVS). Under OVS, sites that offer all required food components in the minimum amounts through at least five different food items can allow participants to decline certain items if participants take at least three different food components with their lunch. This optional menu planning approach is intended to help reduce food waste and administrative cost by letting participants select foods they intend to eat while still ensuring nutritional adequacy of the meal. 9 As the U.S. continues to face the growing problem of food waste parallel to the ongoing issue of food security and heightened periods of risk for food insecurity, specifically the summertime, establishing sustainable food system management practices and recommendations for child nutrition programs like the SFSP is critical. According to the USDA, in 2018 the SFSP served nearly 2.7 million children accounting for approximately 150 million meals and snacks. 10,11 The number of children served during July of 2020 and 2021 were reported at 5.7 and 5.1 million, respectively. 6,12 It is plausible that the observed increase in SFSP utilization occurred due to a variety of waivers allotted during the COVID-19 pandemic public health emergency, e.g., site eligibility, allowance of meals to be taken offsite, etc. Given the large volume of food served over a relatively short time frame in the SFSP, it is important to understand consumption patterns and how much food is consumed or wasted. [13][14][15] Understanding how much and which foods are wasted by USDA SFSP participants, as well as the potential influence of delivery styles on food waste can provide insight into potential changes to SFSP food offerings and programmatic policies. This information is necessary for policy makers and program staff to ensure that children are receiving adequate nutrition while maintaining or increasing program efficiency by reducing food waste. Unfortunately, relative to the SBP 16,17 and NSLP, 18 food waste in the SFSP has been highly understudied and it is unknown to what extent children are or are not consuming the meals and snacks served at the SFSP sites. While the SFSP continues to be understudied, a recent systematic review of universal school meal programs -SBP and NSLP -found that most studies of school meal programs observed positive impacts on diet quality, food security, and academic performance. This is particularly notable if the same positive impacts are observed among SFSP participants given that children are at risk for hunger during the summer months and declines in diet quality, but also, declines in overall health, specifically unhealthy changes in weight status. [13][14][15] Therefore, it is imperative that child opportunities to consume nutritious foods during the summer are maximized.
Given the potential protective impact of federal meal programs on food security and diet quality and the heightened window of risk for declines in food security and diet quality during the summer months, it is imperative to understand the foods being consumed by children at SFSP sites. Therefore, the aims of the current study were to: 1) pilot methodological approaches for examining plate waste in the USDA SFSP setting; 2) examine plate waste patterns at urban USDA open SFSP sites; and 3) compare plate waste by meal delivery style.

Study Design, Participants, Recruitment, and Setting
Project Summer Weight and Environmental Assessment Trial (SWEAT) was an observational, prospective study implemented at two elementary schools in predominantly low-income urban neighborhoods of Columbus, Ohio. 19 The Project SWEAT main study examined trends in child weight and health status during summer. Project SWEAT also included a sub-study focused on examining the food, physical activity, and social behavioral and environmental determinants of unhealthy weight gain during the summer months. The data presented in this article are from the Project SWEAT sub-study. More information regarding the Project SWEAT study design is published elsewhere. 19 Purposive sampling was employed whereby sites were selected based on the number of children attending the sites (i.e., more highly attended sites), as well as meal delivery style. Additionally, feasibility factors were considered, i.e., menu cycles during scheduled data collection periods and staffing. SFSP open sites located in the Project SWEAT neighborhoods were invited to participate and two sites enrolled. The community partner sponsoring the SFSP was operating n = 20 sites in the participating neighborhoods. Plate waste data were collected from any child who received a lunch at the participating site. A verbal assent script was read to all participants at the start of the lunch period. Parental consent was exempt for this portion of the Project SWEAT sub-study.

Data Collection
Data were collected in July of 2017 over the course of two weeks. The determined length of time was based on the rotating menu cycle. Data were collected for n = 5 days at Site 1 and n = 6 days for Site 2. Data were not collected for all n = 10 days over the two-week period due to children eating off site at field trips, days sites were closed, etc.
Field Notes: One trained researcher observed and took daily field notes of site characteristics, changes to menus, observed patterns of consumption (e.g., many children seeking seconds for a particular food item, etc.). Additionally, while the participants consumed their lunches, the researcher unobtrusively observed the participants from the back of the room and recorded field notes.
Plate Waste: Data collectors underwent a half-day standardized calibration training with a senior researcher in the field. The training consisted of education on methodological approaches, as well as observed application of the methodological approaches on three trays. On each data collection day, a trained researcher arrived at the site approximately 30 minutes prior to the start of the lunch hour. The researcher first checked the site attendance sheet and recorded how many participants were present. Next, the researcher reviewed the cafeteria menu for the day and visually observed the lunch food options before pre-populating a data collection spreadsheet with item names and serving sizes for each food that would be offered or served. During this time, the researcher visually and tactilely observed a full portion size for each lunch item, which later served as the reference size for assessing plate waste. At the start of the lunch hour, the research staff reminded child participants to not trade or discard any contents of their lunch and to leave the remaining contents of their lunch tray at their seat when they were finished eating. All child participants sat together and ate lunch at the same time. Trashcans were removed from the immediate eating area and stored in the back of the room away from the participants.

Outcome Measures
The researcher collected data using the visual quarter-waste method for each item on individual trays after participants exited the lunchroom. 20 The visual quarter-waste method has been validated for use in child nutrition programs 21 and previous research has demonstrated that this particular scale produces the most accurate visual estimation results when assessing waste from prepackaged food items, similar to those served as part of the SFSP. 20 Using this method, each individual food item on each tray was given of a score based on the following scale: 0 for none wasted; 1 for ¼ wasted; 2 for ½ wasted; 3 for ¾ wasted; and 4 for all wasted. Data were only collected on foods that could be identified and an item was left blank if the researcher was unable to determine whether an item had been present.

Sample Size Calculation
Due to the formative nature of the research, a formal sample size calculation was not appropriate and therefore not performed. Previous plate and food waste research that was formative in nature, i.e., due to new meal standards, addition of supplemental education programs, etc., historically has not included or reported formal sample size calculations. [22][23][24][25]

Data Analysis
Field notes were electronically transcribed verbatim from original hard copy notes. The transcriptions were summarized by content area, i.e., site characteristics, changes to menu, and consumption trends. The daily summarized notes were compared to quantitative waste data to contextualize findings.
Each food item on collected trays was assigned a food group -fruit, vegetable, grain, meat/meat alternate, or milk. Frequency of amount wasted (0%, 25%, 50%, 75%, and 100%) for each food item was summarized and organized by food group (fruit, vegetable, grain, meat or meat alternate, or milk) and by site. Mean percent waste (i.e., average of percent wasted − 0%, 25%, 50%, 75%, and 100%) was also calculated for each food group. Many of the food items (e.g., turkey and cheese sandwich) served had components that counted toward two or more food groups (e.g., grain, and meat/meat alternate). To account for these mixed foods, each food item was coded for each food group for which it contained a full serving. For example, a turkey at cheese sandwich was coded as grain and meat/meat alternate. Additionally, it was possible for trays to contain more than one serving of a particular food group, such as two fruit servings. Therefore, total food item servings coded exceeded the number of trays analyzed for several food groups.
All study materials and procedures were approved by The Ohio State University Behavioral and Social Sciences Institutional Review Board (2016B0034).

Field Notes
Content of field notes included style of food (i.e., hot vs. cold), packaging of foods, meal components, service style, handling of leftovers, and consumption patterns (i.e., children returning for second servings of food items), and data collection issues (i.e., interruptions from site staff). Most food items served at the sites were served cold and prepackaged with the exception of whole pieces of fruit, such as apples. An example meal was a ham sandwich (3 oz. ham and bun), 3 oz. celery, 3 oz. canned pears, and 8 oz. fat-free chocolate milk. The two sites had different lunch menus as the meals were provided by two different SFSP vendors.
The two sites also had different meal service styles ( Table 1). The options consisted of the daily menu items, as well as leftovers from the previous day, as available. Site 1 had a multi-offering tray lane with OVS whereby participants could decline some food components in favor of selecting the foods they intended to eat, as long as they took at least three food components including a fruit or vegetable. Participants typically had three different options for fruits with the exception of one day in which only one fruit was offered. Except for whole pieces of fruit, most foods came in prepackaged form. Participants always had an option of 1% white milk or fat-free chocolate milk. The participants were permitted to return to the tray line for second portions of the meal components that remained after all participants had received a meal. At site 2, meal delivery style was a pre-assembled tray where children were given a pre-assembled lunch tray filled with prepackaged food items, including two fruits/vegetables, one milk, one grain, and one meat/meat alternate. Children at Site 2 could not decline any meal components. The default  food groups was consistently greater at Site 2 (pre-assembled trays) compared to Site 1 (multi-offering tray line with OVS) (Tables 2 and 3).

Discussion
Food insecurity among children residing in low-income households is a persistent public health problem in the U.S., particularly during the summer months when school is out of session and families lose access to school-based federal programs such as the USDA SBP and NSLP. While the summertime counterpart does exist (USDA SFSP), it is historically underutilized. Even more, the USDA SFSP has been highly understudied, and nutrition standards for the SFSP lag behind other child meal programs. 7 There is a dearth of information available regarding SFSP meals. Given the trends observed in food waste in other child meal programs (e.g., SBP, NSLP) coupled with the lagging nutritional standards, 7 it is possible that significant food waste exists within the SFSP, the aims of the current study were to pilot approaches for examining plate waste at USDA open SFSP feeding sites, assess food waste of meals served at USDA open SFSP feeding sites and examine differences in food waste by meal delivery style. To our knowledge, this study is one of the first assessments of plate waste in the SFSP. Findings from the current study demonstrate that the greatest complete food waste (100% wasted) occurred for vegetables, with 56.6% of vegetables completely wasted. Mean percent waste for vegetables was also the highest among the food groups at 70.90%. These findings are consistent with a 2003 SFSP study and previous research on the NSLP that indicated vegetables are typically the most wasted food group by elementary school students participating in child nutrition programs. 23,[26][27][28] Del Rio-Rodriguez and Cullen also observed vegetables as the least consumed food group among SFSP participants. 29 In the current study, fruits were the second most completely wasted food group, with 33.2% of fruits completely wasted. Taken together, these findings on fruits and vegetables are in line with a review of plate waste in the NSLP from 1978-2015 that determined fruits and vegetables accounted for the greatest food waste across studies. 18 The current study also found milk to be the least wasted food group, which is comparable to the 2003 SFSP study, previous studies on the NSLP, and Del Rio-Rodriguez and Cullen that determined milk is often among one of the least wasted food groups. 18,23,27,29 School or child meal programs may be perceived by parents and children as the "place" where they consume milk or milk may not otherwise be offered in the home.
In the current study, complete fruit and vegetable waste was higher (43.4% and 62.0%, respectively) at Site 2, where children were served pre-assembled trays and could not decline food components. While waste patterns in the current study are consistent with studies that have examined waste in the SBP and NSLP, reasons for waste patterns (e.g., food preference) continue to be underexplored in food and plate waste research. 18 Types and presentation of vegetables may play a role in vegetable waste. For example, some vegetables served during lunch in the current study included plain celery or plain broccoli. Additionally, the findings from the current study may indicate that the delivery style of SFSP meals may play a role and need to be considered to maximize nutritional health during the summertime window of risk.
There were notable differences in the lunchroom set-up across the two sites. When comparing waste across the two sites, Site 2 had higher complete food waste and mean food waste for all food groups compared to Site 1. This suggests that a multi-offering tray line with OVS lunchroom set-up, which more closely mirrors the structure of a school cafeteria, may have a more favorable impact on food waste compared to serving pre-assembled trays. These findings also support the potential effectiveness of OVS in reducing food waste in the SFSP while still maintaining the nutritional integrity of meals.
The option to utilize OVS was granted to school food authorities (SFAs) participating in the SFSP in 1998 and later extended to non-SFA sponsors (e.g., community and recreation centers) in 2011. Unfortunately, the extension of the OVS flexibility to non-SFAs was rescinded in 2022, with USDA citing concerns about the effectiveness of OVS implementation by non-SFA sponsors. 30 Based on the findings of the current study, continuing to limit the OVS option to only SFA sponsors may increase food waste generated by the SFSP. More research needs to be done to confirm this study's findings as well as explore strategies to educate SFA sponsors on successfully implementing OVS in the SFSP. In the future, USDA should consider re-expanding the OVS option to non-SFA sponsors participating in the SFSP in an effort to reduce food waste.
In 2022, the USDA made permanent a flexibility for sponsors to allow SFSP participants to take one food item (a fruit, vegetable, or grain) offsite for later consumption. 30 The USDA began considering this option after program operators expressed concern about young children being unable to consume all meal components in one sitting as well as the corresponding plate waste generated by the SFSP. Some of the findings of the current study support the continuation of this flexibility given that complete (100%) food waste was high with 33.2%, 56.6%, 26.0% of fruits, vegetables, and grains wasted, respectively. Although, it is important to note that no food waste (0%) was greatest for fruit and grains, indicating that children are completely consuming or completely wasting these foods, which may be due to food acceptability or food preference. At present, the flexibility only allows participants to take home one item. However, the high amount of waste among these three specific food groups suggests that in the future, USDA may consider expanding this flexibility to allow participants to take home more than one item to ensure children are receiving the nutrients they need for healthy growth and development. Future studies or pilot tests should explore the potential benefits of expanding this flexibility and the associated effects on participant nutrient intake and food waste.
It is worth noting that the total vegetable servings were lower and total meat/meat alternate servings was higher compared to the other food groups. Relative low total vegetable servings was driven by Site 1. This is because they operated as a multi-offering tray line meal service with OVS to which many children refused the vegetable thus driving down the total servings for vegetables. Total number of servings for meat/meat alternate was likely driven by many factors. At Site 1 multiple kids on multiple days were able to return to the line and take more than 1 entrée. At Site 2 there were multiple days in which two separate meat/meat alternates were served in the same meal, such as turkey slices and pepperoni slices and cheese cubes or a sunbutter (i.e., nut butter made from sunflower seeds often served in child meal programs due to nut allergen concerns) sandwich with cheese stick on the side.
The USDA SFSP, while historically underutilized relative to the USDA NSLP and SBP, is a program that has significant potential to impact child health due to the time that it is offered -the summer months, an identified window of risk. The USDA SFSP was created to address concerns of child food insecurity during the summer months as children lose access to free and reduced-cost school meals. Not only are children at risk for hunger during the summer months, but also, they are at risk for declines in diet quality 11 and overall health, specifically unhealthy changes in weight status. 13,14,31 Serving foods that are healthy and liked by children may reduce food waste and support overall child nutritional health. 32,33 Click or tap here to enter text. Our findings indicate that there may be great potential to prevent excessive food waste based on the delivery style of SFSP meals. Consistently lower food waste observed at Site 1, the current study also indicates that format of meal delivery may play a significant role in child food consumption. Given the dearth of research conducted on food waste in the SFSP, it is imperative that future research measure food waste consistently to compare across studies. 18,18,20,21 The main strengths of the current study are its novelty and relevance. There is a dearth of peer-reviewed literature pertaining to the USDA SFSP, with an even greater scarcity of data pertaining to foods consumed by SFSP participants. Therefore, not only does the current study provide insight to a highly understudied, federally funded program, but also, the current study examines food wasted, which has relevance to economic and sustainability nutrition efforts. Additionally, the validated quarter-waste data collection method was employed. In the current study, plate waste data were collected utilizing the visual quarter-waste method. This approach was lowcost and unobtrusive, with minimal researcher training required. Previous studies have shown that the visual quarter-waste method is a validated and reliable approach to measuring food waste in school cafeteria settings, especially for prepackaged food items. 20,21 Beyond its use in traditional school year settings, this method was also a practical way to assess food waste at SFSP sites in which most items were prepackaged and uniform in size.
Several study limitations existed. First, generalizability of the findings is limited given the short timeframe (i.e., one to two weeks during the summer) and the small number of SFSP sites (i.e., two sites in one urban setting). Second, other factors, such as child food preference, food skills, can impact food consumption and waste, but were beyond the scope of the current study. Third, the study was also limited to only open SFSP sites. Open sites compared to closed sites can be fundamentally different, in that closed sites are typically more structured in nature. Children sign in and out of closed sites and may have more structure around their meal patterns. While children attending open sites can drop in and leave as they please. It is possible that a child dropped into an open site just before lunch and was served a lunch even if they consume food prior to coming to the site. Fourth, the number of trays analyzed per site varied greatly, and thus conclusions drawn comparing meal delivery type should be made cautiously. Finally, child characteristics (e.g., age) were not obtained and could have impacted the results, i.e., younger children may have more plate waste compared to older children.

Conclusions
There is need for further research to confirm findings from the current study and to identify underlying factors in order to properly address this food waste problem. Researchers should consider supplementing quantitative plate waste research with rigorous qualitative approaches to understand why children are consuming or not consuming various foods offered during the SFSP. Future research should explore the impact of meal service style on plate waste in the USDA SFSP setting. Finally, policies surrounding the SFSP meals should consider implications on food waste, as well as child hunger and food security.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
USDA North Central Nutrition Education Center for Excellence.