Beyond clinical food prescriptions and mobile markets: parent views on increasing healthy eating in food-insecure families

Children in food-insecure families face increased barriers to meeting recommendations for fruit and vegetable consumption. Hospitals and pediatric healthcare institutions have attempted to alleviate food-insecurity through various internal programs like food prescriptions, yet little evidence for these programs exist. Consistent with a patient-centered perspective, we sought to develop a comprehensive understanding of barriers to fruit and vegetable consumption and a parent-driven agenda for healthcare system action. We conducted six qualitative focus group discussions (four in English, two in Spanish) with 29 parents and caregivers of patients who had screened positive for food-insecurity during visits to a large pediatric healthcare system in a midwestern U.S. city. Out iterative analysis process consisted of audio-recording, transcribing and coding discussions, aiming to produce a) a conceptual framework of barriers to fruit and vegetable consumption and b) a synthesis of participant programmatic suggestions for their healthcare system.


Abstract Background
Children in food-insecure families face increased barriers to meeting recommendations for fruit and vegetable consumption. Hospitals and pediatric healthcare institutions have attempted to alleviate food-insecurity through various internal programs like food prescriptions, yet little evidence for these programs exist. Consistent with a patient-centered perspective, we sought to develop a comprehensive understanding of barriers to fruit and vegetable consumption and a parent-driven agenda for healthcare system action.

Methods
We conducted six qualitative focus group discussions (four in English, two in Spanish) with 29 parents and caregivers of patients who had screened positive for food-insecurity during visits to a large pediatric healthcare system in a midwestern U.S. city. Out iterative analysis process consisted of audio-recording, transcribing and coding discussions, aiming to produce a) a conceptual framework of barriers to fruit and vegetable consumption and b) a synthesis of participant programmatic suggestions for their healthcare system.

Results
Participants were 90% female, 41% Black/African American and 41% Hispanic/Latino. Barriers to fruit and vegetable consumption in their families fell into three intersecting themes: affordability, accessibility and desirability. Participant-generated intervention recommendations were multilevel, suggesting healthcare systems focus not only on clinic and community-based action, but also advocacy for broader policies that alleviate barriers to acquiring healthy foods.

Conclusion
Parents envision an expanded role for healthcare systems in ensuring their children benefit from a healthy diet. Finding offers critical insight on why clinic-driven programs aimed to address healthy eating may have failed and healthcare organizations may more effectively intervene by adopting a multilevel strategy.

Background
According to the Centers for Disease Control, increasing fruit and vegetable consumption (FVC) among children is a national public health priority (1). FVC not only provides immediate health and nutritional benefits in childhood (2), but is associated with adult dietary patterns and reductions of chronic diseases, including obesity, cardiovascular disease, type 2 diabetes, and some cancers (3,4).
National guidelines recommend children consume five or more servings of fruits and vegetables daily, however, recent data suggests only about 10% of children ages 2-18 years meet these guidelines (5).
Children experiencing food insecurity face major barriers to meeting FVC recommendations. Food insecurity is a household-level condition of limited or uncertain access to adequate food, resulting in disrupted eating patterns or reduced food intake (6). Nationally, about 12% of American households were food insecure for at least some time during 2017, even though most reported participating in one or more nutrition assistance programs (e.g. Supplemental Nutrition Assistance Program (SNAP), Women, Infants, and Children (WIC) & National School Lunch program) (7). Food-insecure families are two to four times more likely to report barriers to accessing fruits and vegetables (8), and severity of food insecurity has a strong, negative association with fruit and vegetable intake (9).
Consistent with increased efforts to address social determinants of health, hospitals and healthcare systems across the country have started investing in interventions to screen for and address food insecurity among their patients (10,11). Patients identified as food-insecure may be linked to services in the community, or to programs initiated within healthcare facilities, such as onsite food pantries, home delivered meals, or food prescription (12). Most of these programs, however, aim to increase general food access, missing the opportunity to target FVC (13)(14)(15). One exception is a fruit and vegetable prescription model, in which health care providers dispense "prescriptions" in the form of coupons or vouchers (16). However, published evidence on the impact of these prescription programs, particularly within a primary care setting, is lacking.
Our experience suggests that increasing FVC among children in food-insecure families may not be as simple as dispensing food prescriptions. In our pilot program, primary care providers at a large, urban clinic dispensed fruit and vegetable prescriptions to food-insecure families during a clinic visit. The prescription consisted of a $5 voucher, redeemable at a community mobile market that operated from April-October 2017 and exclusively sold low-cost, high-quality fresh produce. During the pilot, 462 coupons were distributed along with a mobile market schedule and educational brochures; however, only 5% of vouchers were redeemed. Following this unsuccessful pilot, we sought to conduct an indepth examination of barriers and facilitators to fruit and vegetable consumption among foodinsecure families.
Consistent with a patient-centered model of care, we also aimed to elicit a parent-driven agenda for interventions pediatric health care facilities can implement to address fruit and vegetable consumption for food-insecure families. Previously, while studies have identified high cost and limited access as barriers (17)(18)(19) to fruit and vegetable consumption in food-insecure families, none to our knowledge have sought to generate solutions from the caregiver perspective. Understanding the perspectives from families experiencing food insecurity can help identify opportunities to improve program acceptability, feasibility and effectiveness.

Study Setting and Participants
We conducted qualitative focus group discussions with parents and caregivers of patients at a large pediatric health care system. According to 2016 data in the hospital's four-county catchment area, 19.5% of caregivers "often" or "sometimes" worried that their food would run out and 16.0% "often" or "sometimes" ran out of food and did not have money to buy more (20). We recruited participants who screened positive for food insecurity and were given a food prescription during the 2017 pilot program. Participants were considered food insecure if they answered "yes" to one or both of the Hunger Vital Sign™ questions sourced from the Safe Environment for Every Kid (SEEK) questionnaire (21): "In the past 12 months, (1) did you worry that your food would run out before you could buy more? (2) did the food you bought just not last and you didn't have money to get more?" Participants were eligible if they spoke English or Spanish and were at least 18 years old. We forecasted that 5-7 focus groups (6-8 participants each) would be adequate to reach thematic saturation (22), and aimed to conduct 3-5 groups in English and 2 in Spanish. This study was approved by the Institutional Review Board at Children's Mercy Kansas City.

Data Collection
Focus groups were conducted at various locations within the pediatric facility campus and were led by trained moderators in English or Spanish. An assistant moderator took notes and recorded non-verbal cues such as body language, emphasis and facial expressions (23). Each discussion was audiorecorded and lasted approximately 90 minutes. Participants gave verbal consent prior to participation and were offered a $35 gift card, healthy meal, and on-site childcare.
The moderator used a semi-structured guide to lead the discussion. The guide was composed of questions and probes that reflected our aim to identify barriers and facilitators and elicit a parentdriven agenda. Discussions began with an open exploration of parent's general concerns regarding food for their family. The latter half of the discussion used community-based participatory research methods (24) to guide participants in generating solutions for what a healthcare institution could do to increase FVC among families. Ideas were summarized on chart paper, ranked by participants in order of preference and top ideas were discussed further.

Data Analysis
Focus groups were transcribed and uploaded into Dedoose version 8.2.14 (25). Spanish transcripts were transcribed into English by Spanish-speaking moderators. Analysis methods were informed by grounded theory (26) and began with line-by-line inductive coding of the first English and Spanish transcripts by two independent coders (27). Preliminary codes were refined through discussion and consensus and organized thematically into a hierarchical codebook. Iterative restructuring of the codebook by group consensus continued as new themes and relations between themes emerged in subsequent transcripts. Code outputs were reviewed and summarized, and the structure of the codebook became the basis of two final analytic goals: a) a conceptual framework of barriers to FVC and b) a synthesis of participant programmatic suggestions, which thematically aligned within a social-ecological framework (28).

Results
A total of 29 participants attended the six focus group discussions (four in English and two in Spanish). Participants represented various age and racial groups, with the majority (26) female.

Predominant Barriers to Fruit and Vegetable Consumption
Participant-reported barriers to FVC fell within three interrelated categories: affordability, accessibility, and desirability ( Fig. 1).

Affordability
Affordability concerns encompassed lack of money to buy food as well as the high price of fruits and vegetables. When asked about their biggest worries regarding food in general, participants' most salient response was overwhelmingly lack of money. When asked specifically about consuming fruits and vegetables, caregivers maintained affordability as the biggest barrier.
"If parents could find cheaper vegetables, we would be able to buy them" (Spanish-speaking Female,

Group 4).
Participants considered "healthy foods" (fruits and vegetables) more expensive than other foods and could only consider purchasing them after paying for other necessities. As one participant explained: Expense of fruits and vegetables was a concern for participants trying the stretch their food budgets to feed large families:

"Not only is it expensive, when you got 5 kids […] A bag of oranges are gone in a day, an hour"
(English-speaking female, Group 1).

Accessibility
Participants also talked about difficulties in accessing affordable, quality fruits and vegetables, including frustration with the lack of transportation and lack of stores selling healthy foods within their neighborhoods.
"Nothing is close, might have to take several buses to get to the grocery store or buy food at a [gas] station where they mark food up 500%" (English-speaking female, Group 3). Access to quality fruits and vegetables was especially challenging to families relying on food pantries.
One participant described the produce at pantries as "on the edge of expiration" (English-speaking female, Group 2). While they did not prefer canned goods, several participants saw them as the only option for obtaining fruits and vegetables at food pantries. Families who did not have easy access to pantries faced additional barriers related to affordability and desirability, a choice that one participant explained was impacted by the low quality of pantry food:

Participant-Generated Recommendations for Healthcare-based Initiatives
Participants believed the healthcare system could have a wide-ranging role in helping children of food-insecure families eat more fruits and vegetables. Major themes included addressing affordability through direct assistance with foods and other basic needs and through advocacy; increasing accessibility through integrating services into their routines while using multiple channels of communication; and promoting desirability through the involvement of families and the influence of clinic providers. Beyond a focus on food, participants discussed other factors that impact a family's ability to acquire healthy food, such as one Spanish-speaking focus group that emphasized the need to increase access and education surrounding family planning. Table 2 summarizes participants' intervention ideas across a socio-ecological model (29), and their potential impact to increase affordability, accessibility and/or desirability. Address Affordability: (1) Offer direct assistance and linkages to external resources: As affordability was the most significant barrier discussed, parents suggested healthcare institutions offer direct assistance in the form of coupons or onsite food (i.e. bags of fruits and vegetables, onsite pantry). Coupons were most preferred, but participants thought a higher amount, redeemable at more convenient locations would make the program more successful than the pilot food prescription program.
"$5 would be okay, but if they were to give $20, I would say it's worth taking and not losing out on it.
And I could use it at whatever store I go to, or where I buy the most produce" (Spanish-speaking female, Group 4). Increase Accessibility: (1) Integrate services into "my routine": Participants recognized the ease of using programs like WIC and SNAP where they could access benefits at times and places that were already "within the routine […] it wasn't an extra trip" (English-speaking female, Group 3).
Participants recalled positive experiences with healthy vending options at hospitals and with programs that were tied to routine activities such as healthcare or grocery store visits that gave free pieces of fruit to children.
(2) Communicate programs opportunities often, through multiple channels: Participants felt they would be more likely to access programs if they were publicized in multiple ways and times. Some recalled existing community gardens that were underutilized because people did not know how to get involved. Participants suggested that hospital-based programs (like gardens, educational classes, and support groups) be publicized through their provider during clinic visits, calls to make appointments, appointment reminders, as well as through clinic memo boards, flyers, websites, email and postal mailings.
Influence Desirability: (1) Involve children or whole family: Participants emphasized the importance of involving children in programming, particularly in discussions about community gardens, workshops, or rewards programs.
In talking about a community garden at the healthcare institution, one participant said,  (19) and the lack of resources to purchase them (affordability) (18). Other studies have also identified accessibility barriers, including lack of stores selling healthy foods and transportation difficulties (17,19) to those that do, as well as desirability barriers, such as low within-family demand for fruits/vegetables (31), high time investment to prepare healthy meals (30), and cultural traditions which may emphasize less healthy foods. Using grounded theory methods, our study advances the holistic understanding of barriers among food-insecure families and offers a comprehensive conceptual model to guide future inquiry and program planning.
Although affordability was the most salient barrier discussed in our focus groups, we found that the three main barriers were often interconnected. For example, participants who may not be able to afford food at grocery stores may visit pantries, which were often difficult to access or did not carry fruits and vegetables of a desirable selection or quality. Further, families are concerned about wasting money on vegetables that children may not want to eat or that perish quickly, highlighting the intersection of affordability and desirability. Other studies have identified some intersectionality of barriers, including the low desirability of food options at more accessible locations (32,33), and the links between affordability and access (17,30,34). In examining these three intersecting barriers alongside participant programmatic recommendations, our results highlight the need for hospitalbased initiatives to go beyond addressing single barriers and consider affordability, accessibility and desirability factors in their programming.
Our study fills critical gaps in guiding health systems to offer patient-driven, effective programmatic solutions to increase FVC among food-insecure families. Families themselves identified that healthcare institutions can play a distinct role in interventions across the socio-ecological model ( was not likely to alleviate food insecurity nor increase FVC, one aim of the pilot "prescription" was to introduce families to a community resource for low cost fruits and vegetables that serves food deserts in the urban core. This aim was not achieved as the mobile market was not easily accessible for families in the pilot. Additional implementation science research is needed to determine if fruit and vegetable prescription interventions can be adapted to meaningfully address barriers related to food insecurity in primary care populations.
Our participants also recognized that interventions at the individual/family level alone are not likely to be enough. Programs at the community level may be beneficial but need to be designed according to family needs and accompanied by targeted outreach. Participants spoke positively about programs like mobile markets, group classes and community gardens and emphasized them when generating ideas for new programs. However, contrary to wide utilization of programs like WIC, SNAP and schoolbased food services, very few had reported participating in these existing community-level programs.
Low reported utilization of existing community-based programs may be due to access barriers (such as transportation (17) or time constraints (30)), thus, additional effort may be needed to ensure these interventions are compatible enough with participants' daily realities to become "part of my routine".
At the organizational level, parents suggested that healthcare providers connect families to community resources, an idea that has also been endorsed by nutrition and research experts (35).
Consistent with existing literature, participants in our study described how their food budgets were limited after paying for "priority" items such as housing, childcare, and medical bills (17). Healthcare institutions are testing models to employ community navigators or community health workers to connect families to programs for food security and broader services, with promising results (12,36).
On the policy level, our study highlights that patients see health care providers and institutions as advocates for their families for issues from food security to safe affordable housing. Healthcare providers should consider ways to advocate for maintenance and expansion of effective federal and community programs like WIC and SNAP (consistent with American Academy of Pediatricians recommendations) (37) as well as increasing the nutrition quality of existing emergency food programs.
This study has several limitations. Its focus on families accessing primary, pediatric care an urban center may not be representative of food-insecure families accessing care in suburban or rural areas, or those who are not connected to formal healthcare services. While males were not excluded from participating in a focus group, only three participants were male, thus, gender influence in perspectives and solutions may not have been fully captured. Solutions generated were solely of the parent perspective and will need to be further explored to determine their feasibility and effectiveness if implemented. Further, while saturation was reached in terms of barriers and facilitators, some new program ideas were generated at each focus group and additional focus groups may have produced additional ideas. Lastly, although participants were encouraged to share both positive and negative experiences, research staff conducting the focus groups were employees of the healthcare center, which may have led to social desirability bias in questions related to the facility's programs.

Conclusion
Food insecurity and FVC are complex challenges, affecting both short and long-term health outcomes for children. According to families, healthcare institutions have a role in addressing these challenges on multiple levels. Our study offers a conceptual framework and parent-driven solutions to guide hospitals in defining a strategic, comprehensive role in increasing FVC and its health benefits for children facing household food-insecurity.

Consent to publication
Not applicable Figure 1 Predominant barriers to fruit and vegetable consumption among food-insecure families