Pediatric produce prescription initiatives in the U.S.: a scoping review

Background To describe pediatric Produce Prescription (PRx) interventions and their study designs, outcomes, and opportunities for future research. Methods A scoping review framework was used to describe PRx interventions published between January 2000 and September 2023. Articles from online databases were uploaded into Covidence. Data on study characteristics, outcomes of interest (health, food insecurity (FI), nutritional and culinary efficacy, and fruit and vegetable (F/V) consumption), and feasibility were extracted. The Mixed Methods Appraisal Tool (MMAT) was used for quality assessment. Results 19 articles met inclusion criteria. Ten studies were quantitative, five were qualitative, and four used mixed-methods. Interventions included food vouchers (n = 14) or food box/pantries (n = 5). Four studies allowed food items in addition to F/Vs. Six studies measured changes in FI and five reported a statistically significant decrease. Seven studies measured changes in F/V consumption and five reported a statistically significant increase. One study reported a statistically significant reduction in child BMI z-score. Most studies reported high feasibility. Few studies used high-quality methods. Conclusions Pediatric PRx interventions show promising potential to reduce FI and improve diet quality and health-related outcomes. Future studies should utilize rigorous study designs and validated assessment tools to understand the impact of pediatric PRx on health. Impact This work offers a summary of programmatic outcomes including retention, redemption, incentives, nutrition education, study design and quality limitations to help inform future work. We found positive impacts of pediatric produce prescriptions (PRx) on FI, F/V consumption, and nutritional knowledge and culinary skills. More high-quality, rigorous studies are needed to understand the best delivery and design of PRx and their impact on child behavior and health outcomes. This work provides support for the need for rigorous studies and the potential for PRx to play a role in multi-pronged strategies that address pediatric FI and diet-related disease.


INTRODUCTION
2][3][4] However, few children and adults achieve the recommended daily F/V intake. 5,60][11] In 2021, the prevalence of FI in the United States (U.S.) was 10.2%, but higher in households with children (12.5%) and in Black (19.8%) and Hispanic (16.2%) households. 12In adults, FI is associated with increased risk of chronic disease such as hypertension, diabetes, and stroke. 13While its association with risk of obesity, hypertension, and diabetes has been mixed in children, 13 FI is associated with higher health care utilization and cost in children and families. 14,15It is also hypothesized that childhood FI could contribute to chronic diseases in adulthood, although there is not yet a clear understanding of this relationship.Given the disproportionate burden of FI in children and the rise in chronic conditions, such as diabetes, hypertension, and obesity in children over the past decade, 16,17 FI and nutrition focused interventions are an important area for research.
9][20] However, as FI and diet-related chronic disease are core medical concerns, interventions integrated within the healthcare setting are critical.According to the National Produce Prescription Coalition (NPPC) Produce Prescriptions (PRx) are "a medical treatment or preventative service for eligible patients due to diet-related health risks or conditions, food insecurity, or other documented challenges in access to nutritious foods, and are referred by a healthcare provider or health insurance plan.These prescriptions are fulfilled through food retailers and enable patients to access healthy produce with no added fats, sugars, or salt, at low or no cost to the patient.When appropriately dosed, PRx interventions are designed to improve healthcare outcomes, optimize medical spending, and increase patient engagement and satisfaction." 21PRx interventions fall under the Food is Medicine framework (or Food as Medicine). 22hese interventions work within the healthcare systems to offer patients with a diet-related chronic disease risk factor, such as prediabetes or obesity, and who may be at risk or experiencing food insecurity, greater access to produce.Low-or no-cost fresh, frozen, or canned produce and sometimes other food items such as non-perishable healthy staples (i.e.legumes/beans and whole grains) are offered through "incentives" such as redeemable vouchers or directly through provision of food by self-selection or pre-selection via pick-up or delivery.The goal of PRx is to help to prevent, manage, or treat diet-related disease. 21,23,242][33] The potential long-term health benefits of pediatric-focused PRx have not yet been well-studied given the relative novelty of this field.
To our knowledge, there is no published literature that summarizes pediatric-focused healthcare based PRx interventions.Furthermore, the rapid growth of the PRx interventions over the past 2-3 years renders a need for critical review of publications in the field.In particular, evidence for the impact of these initiatives on household FI and family dietary-related behaviors is lacking.This scoping review aims to describe the range of studies, interventions, and outcomes that exist in pediatric PRx interventions in healthcare settings and identify gaps and future directions for the field.

METHODS
A protocol was developed based on the five-stepwise scoping review framework, 34,35 which included defining the research question, creating search criteria, developing data gathering and analysis procedures, and reporting findings.
Relevant articles were assessed for eligibility using preset inclusion/exclusion criteria.Eligible articles described empirical studies, implementing a PRx in households with children <18 years old (yo); delivered its PRx within the U.S. healthcare system; reported child or adult FI, health, nutritional efficacy, or behavioral outcomes; and were published between January 2000 and September 2023.We also included PRx interventions that were primarily focused on F/ V access but offered additional nutritious food items to provide a comprehensive review of interventions targeting F/V intake.Exclusion criteria were as follows: not in English; published before 2000; review articles; conducted outside of the U.S.; exclusively in adults; studies conducted outside of a healthcare system; studies without interventions or interventions that did not include a PRx component; and studies without measured outcomes.
Articles were obtained from online database searches on Pubmed, Scopus, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane, and Medline.A research librarian assisted in the creation of search terms, which included "food OR produce prescription OR vouchers OR programs OR pharmacy", "pediatrics OR children OR adolescents", "health outcomes", and "food insecurity".A full list of search terms can be found in Table 1.Articles were uploaded into Covidence (Veritas Health Innovation Ltd, Melbourne, Australia) 36 for title and abstract screening, full- text review, and final determination of eligibility for data abstraction by two independent reviewers based on the pre-set inclusion/exclusion criteria.Disagreements between screeners were resolved with review by a third screener.
A data abstraction form and data reporting were guided by the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) guidelines. 37ata were abstracted to describe the study characteristics (study and intervention design, setting, inclusion/exclusion criteria.When reported by articles, the maximum incentive amounts offered per intervention were calculated using the reported amount, frequency, and duration of incentive offered), feasibility, and changes in health, FI status/severity, nutritional efficacy, and behaviors.Two independent reviewers extracted data for each article and an independent third reviewer reconciled discrepancies.A quality assessment of included studies was conducted using the Mixed Method Appraisal Tool (MMAT), 38 which uses five criteria related to the appropriateness of the study design, representativeness of the population, adherence to the stated intervention, completeness of outcome data, and interpretation of the results to assess the methodological soundness of studies.Thus, each study received a rating on a scale of 0 to 5 corresponding to the quality of the study with respect to the stated primary outcome.Two independent reviewers evaluated each article and met regularly to come to a consensus on the quality rating for each study.
(FM) or retailers while one used an online produce market. 43All voucher programs allowed redemption for produce, though two programs allowed redemption for other food sold at FM locations, like meats, cheeses, baked goods, 54 or culturally significant foods like blue cornmeal or dried steam corn. 45Among box programs, two allowed self-selected produce and non-perishable items 48,56 and three programs offered pre-selected F/Vs only. 47,49,50One provided delivery services 47 and the rest were pick-up only.

Distribution of incentives
There was wide variability in the amount, duration, and frequency of incentive distributions.Among voucher programs, the lowest distribution amount was a one-time voucher of $10 or $20. 46 The highest incentive amount across all voucher interventions was $300 per month for 6-9 months, which would equal a theoretical maximum value of $2,700. 51The remaining voucher interventions were calculated to have a range of maximum incentives from a one-time distribution of $30 54 to $900 over 6 months 43,45 per household.Four interventions varied the voucher amount by family size 43,45,51,53 and one varied voucher amount by FI status. 46mong box programs, three studies 47,49,56 reported pounds of produce offered, which ranged from 8 to 30 pounds per distribution.One study described a "box" of produce per distribution 50 while another study indicated the amount as "12 meals per household" per distribution. 48The range of calculated maximum amounts of food offered by box programs was 192 pounds over 12 months 47 to 360 pounds over 6 months 56 per household.One box program varied the food amount by family size. 480][41][42] Duration of programs ranged from 1.5 months 57 to one year, 39,47,49 with an average duration of six months.

Education
50 Four studies reported on frequency and/or duration of education, which corresponded to 24 h total, 47 16 h, 52 monthly sessions, 53 and four total sessions. 55 Incenive redemption was explicitly tied to education session attendance in two studies, 45,48 otherwise, education was not mandatory for incentive redemption. Helthcare clinic staff or providers (Clinician, Nutritionist or Health Educator) were involved in delivering nutrition education in four 47,48,52,53 out of the 12 programs that offered nutrition education.

Behavioral outcomes
44][45]47,[49][50][51][52][55][56][57] Twelve studies reported F/V intake with qualitative (n = 3) or quantitative (n = 9) tools.Among the quantitative assessments of F/V intake, two studies 43,49 used nonvalidated tools, and seven 39,41,45,47,51,52,57 used validated instruments.Statistical testing was conducted in five studies. 39,41,45,47,51All five showed significant increases in child F/V intake, with increases in fruit intake of 30% from 0.6 to 0.8 servings per day (PD) 41 and 43% (from 0.8 to 1.3 cups PD), 47 a 33% increase in vegetable intake (from 0.7 to 0.9 servings PD), 39 and a 31% and a 7.5% increase in total F/V intake (from 5.2 to 6.8 cups PD 45 and from 3.47 to 3.73 cups PD, 51 respectively).Two studies statistically tested and reported the change in adult caretaker F/V intake.One found a significant increase in adult vegetable intake (from 2.22 to 2.44 servings PD) and fruit intake (from 2.05 to 2.46 servings PD), 49 while the other found a non-significant increase in adult F/V intake. 47One study reported families were more likely to shop at an FM in the month following the intervention compared to non-participants. 40Four studies reported changes in physical activity outcomes, 44,45,52,57 in one study the increase was statistically significant. 52

Food insecurity outcomes
56 Redeemable at specific retailers or Farmer's Markets ).Thirteen voucher programs utilized farmers' markets (FM) or retailers and one used an online produce market.Two voucher programs allowed redemption for other foods in addition to produce.Among box programs, two allowed self-selected produce and non-perishable items and three programs offered pre-selected F/Vs only. One provided dlivery services and the rest were pick-up. Of these five,ur 39,45,47,53 measured FI severity using variations of the USDA Household Food Security Survey 58 while the fifth 56 used the Hunger Vital Sign (HVS) screener.59 Overall, these studies reported increased access to food, including F/V, and improved FI status of households.Among the five that utilized validated tools and conducted statistical testing, four reported significant improvementstwo in FI status 45,56 and two in FI severity.39,53 There were reported reductions in household FI status by 94% (100% at baseline to 5.9% at 6 months) 56 and 17% (82% at baseline to 65% at 6 months) 45 and reductions in FI severity by 55% (1.96 at baseline to 0.87 at 12 months) 39 and 12.5% (from 0.72 at baseline to 0.81 at 6 months, this tool scored food security, so an increase is a reduction in FI severity).53 One study reported a "dose"-response effect on FI severity, indicating those with higher intervention participation (5-6 visits out of 6) had a greater reduction in FI severity than those who only attended 1-2 visits out of 6 (effect size β = 0.07).53 One reported non-significant improvements in FI severity.47 Only one study measured longitudinal change in FI as reported by the child and found a statistically significant reduction in their modified FI score from 1.88 at baseline to 1.04 at 12-months.39 Hager et al. reported household FI but it was not possible to disaggregate FI data in households with children from all households.51 Nutritional & culinary efficacy outcomes Nutritional and culinary outcomes, including confidence, skills, knowledge, and attitudes towards the preparation, storage, consumption, and purchasing of foods, were reported by six studies.43,44,47,50,56,57 Three studies utilized qualitative measures 43,44,50 and three utilized mixed measures, 47,56,57 with two using validated questionnaires.47,57 There were no statistical analyses of nutrition and culinary outcomes, though all studies reported improved cooking skills and increased confidence in cooking, utilizing produce, and following recipes.

Health outcomes
Only three studies measured biometrics and health outcomes such as weight, body mass index (BMI), 45,51,52 and health status 51 in children.
One study showed a statistically significant decrease in BMI z-score (95.6 to 73.1) in children who were classified as having overweight or obesity at baseline. 45The other studies found no significant effect of the intervention on child weight. 51,52One study measured change in health status on a five-point scale (poor, fair, good, very good, excellent), and found a statistically significant likelihood of improving one level from baseline after program participation (for example going from poor to fair, or fair to good). 51One study measured change in depression screening scores in adult mothers and found a reduction in scores in women in the intervention arm and no significant change in scores in women in the control arm. 52asibility Feasibility was measured primarily by enrollment of target population, perceived positive impact, and satisfaction in interventions both by adult participants and medical providers.,56 Retention and redemption Retention rates ranged from 27% 44 to 4.5%. 41Redemption ranged from below 20% 52,56 to 80% or greater. 43,55Retention and redemption reporting methods were not standard across studies and some studies did not explicitly report on retention or redemption rates of their participants.

Summary of qualitative
Qualitative responses revealed decreased financial hardship and increased access to healthy affordable food. 42,44,47,50,54,5654,55 In addition, household-level attitude changes towards healthy eating, increased family time, and increased involvement of children in the cooking process were reported. 43,44,47,50,54Qualitative statements also captured sitespecific barriers to participation, concerns over sustainability of perceived impacts after the program ends, 57 and the importance of healthcare-based delivery of programs as a motivation for participation. 44[57] Quality appraisal and limitations The majority of studies were feasibility or pilot studies with small sample sizes, which limits their generalizability and ability to assess the impact of an intervention.MMAT results, as presented in Table 2, reflect acceptable quality with regard to feasibility outcomes.However, with respect to outcomes related to FI and F/ V intake, there was a lack of high-quality methods, including not using validated assessment tools, having incomplete data, not controlling for confounders in the design and analysis, and lacking or inappropriate statistical methods.
Lessons Learned.We have highlighted some considerations for implementation of pediatric PRx interventions in Table 3.

DISCUSSION
To our knowledge, this is the first scoping review to describe pediatric PRx interventions.There is a rising interest in pediatric PRx, which coincides with the recent national traction that FI, food inequity, and nutrition insecurity have received as major public health concerns and policy priorities in the U.S. 60 Our work adds to previous reviews of PRx interventions by incorporating recently published studies and focusing on pediatric and caretaker outcomes related to FI, food intake, nutrition and culinary efficacy, and the family perspective.Additionally, it offers a summary of programmatic outcomes of retention, redemption, incentive amounts, nutrition education, and study design and quality limitations to help inform future work.Overall, the studies reviewed reported positive impacts of PRx towards FI, consumption of F/V, nutritional knowledge and culinary skill in households with children.However, more high-quality, rigorous observational cohort and RCT studies need to be undertaken to understand the best delivery and design of PRx interventions and their impact on behavior and health outcomes.This scoping review provides support for the need for further rigorous study and the potential for PRx to play a role in multi-pronged strategies that address FI and diet-related disease in children.
The emergent nature of pediatric PRx is reflected in the prevalence of pilot feasibility and acceptability studies utilizing non-randomized experimental designs.These studies provide a foundational understanding of the various design and implementation strategies of PRx interventions and explore their potential impact on FI and nutrition-related behavior and knowledge.They also provide a rich qualitative perspective of PRx participant experiences, which can help to define the priorities, needs, and assets of the individuals and communities for whom such interventions are being implemented.There was a wide diversity of study settings, a small number of studies, and lack of rigorous methods, limiting the ability to make generalizations about optimal PRx delivery models from this data.The review explored the relationship between PRx intervention characteristics (incentive amount, type, duration; retention and redemption; offering nutrition education, and assessment methods) and FI and produce intake outcomes but found no obvious patterns.As such, further studies are needed to determine the most efficacious intervention design.The optimal delivery and design of PRx interventions are also likely to differ based on the clinical context and the needs of the target population.Future PRx interventions should determine feasibility and efficacy within specific populations, focusing on strengths, challenges and cultural norms that impact participation and outcomes.The research highlighted herein suggests that increasing accessibility to healthy food by overcoming barriers related to transportation, cost, knowledge, and skills is an important part of interventions that positively impact healthy eating and purchasing patterns.Other research suggests prioritizing cultural considerations in food provision and educational programming is also important. 61n important next step in the research continuum for pediatric PRx is utilizing more rigorous study designs such as RCTs that include robust program evaluation methods, validated measurable outcomes, and sufficient sample sizes to detect statistically and clinically significant behavioral and health changes.While there are important ethical considerations to randomizing families with children for critically needed nutritional resources, RCTs can be conducted equitably with thoughtful methodology.For example, RCTs could explore PRx interventions with different frequencies, durations, delivery methods, produce quantities, participant choice, and various nutrition education components to better understand the efficacy of various PRx modalities and moderating factors.Programmatic outcomes such as retention and redemption rates should be reported to better understand the programmatic success and "dose" of produce prescription being provided.
While food as medicine approaches such as PRx are often used to treat disease in adult populations, a sole focus on treatment of disease may limit the potential impact of these interventions in children.The studies reviewed support the use of pediatric PRx as primarily preventive of adverse long-term health outcomes in children who are at-risk of diet related disease, while also addressing acute nutritional needs.PRx interventions offer an opportunity to impact the developmental trajectory of children by exposing them to a wide variety of tastes and flavors early in life, which promotes healthy eating behaviors later in life. 62Children typically have a slower onset of observable clinical abnormalities and visible signs of disease, however preventing diet-related, chronic disease through PRx may provide cost savings within the healthcare system in these future adults. 63,64Pediatric population health has the potential for impacts on a national level, as these current students will soon become workforce adults and eventually become utilizers of Medicare, so behavior changes now are perhaps incredibly powerful for long-term healthcare implications.As such, prevention-related markers, such as microenvironmental, self-efficacy, and behavioral changes, including increased F/V intake, nutritional knowledge gained, and improved food security, which are predictive of long-term health outcomes, may be highly appropriate outcomes to assess the impact of pediatric-focused PRx within a relatively healthy, resilient, and young population and long-term, cohort follow-ups in increments of 5 years would be beneficial.However, the choice of measured outcomes to assess impact, whether clinical metrics or behavioral or environmental proxies of health, will depend on the target Although some of the articles reviewed measured these proxy outcomes, it is difficult to compare across studies due to the heterogeneity of study designs and assessment tools used.The articles examined utilized three different methods to assess household FI: qualitative interviews, HVS screener, or variations of the USDA food security screener.Studies that utilize the HVS or qualitative tools are not able to determine severity of household FI, making cross-study comparison of the impact of PRx on FI depth difficult.Additionally, PRx interventions may reduce the severity of household and child-level FI rather than eliminate FI, making longitudinal assessments of household and child-level FI severity an important consideration of their evaluation.Standardized measurements are also needed for diet-related outcomes.The studies in this review that reported diet-related outcomes utilized a mix of qualitative, non-standardized single items, and standardized validated questionnaires.Previous adult PRx systematic and scoping reviews have also described difficulty encountering heterogeneous outcome measurements. 32,65A major step toward rigorously evaluating PRx interventions came with the 2018 Farm Bill, which funds the Gus Schumacher Nutrition Incentive Program (GusNIP).GusNIP attempts to standardize evaluation outcomes with shared metrics across studies. 66Forthcoming data will help identify optimal evaluation strategies that balance participant burden with the need for highquality data and provide valuable insight into the impact of these programs.Future studies should consider using the free, validated, standardized tools identified by GusNIP to assess severity of FI and changes in F/V intake, in addition to any other stated primary outcomes, so comparison across studies can be done.Many PRx interventions are funded by philanthropic organizations in community health settings and may have little or no in-house technical support for evaluation.However, it is critical for PRx interventions to conduct program evaluation and research using standardized assessment methods and to disseminate results, which may require a multidisciplinary team of community, academic, and clinical collaborators and study budgets should reflect adequate funding for these efforts.
While interventions specifically addressing populations participating in FNPs were not included in this review, there is wellestablished evidence that these programs can improve FI, decrease poverty, and improve health outcomes. 67Although supplemental, these programs often comprise a major portion of family food budgets and remain inadequate for addressing the needs of families struggling with FI, 68,69 leading to early benefit exhaustion associated with reduced perceived healthfulness of diet 67,70 and maladaptive eating patterns. 71Furthermore, populations that face FI and diet-related chronic illnesses are often more likely to experience additional social risk factors, so effective strategies must continue to advocate for policies that comprehensively address social inequities related to the income and wealth gap that can co-occur in families with FI. 72 PRx interventions can be one part of a multi-pronged approach, in addition to other policies aimed at enhancing the purchasing power of families with limited resources, including assistance for better income, housing, and employment opportunities, to buffer complex and multifactorial social determinants of health. 73uture studies should focus on exploring the impact and sustainability of PRx interventions within healthcare systems.Some barriers to the sustainability of PRx interventions include the cost, complexity, and capacity of integrating PRx within existing clinical settings. 74Although these interventions may require a large initial investment, they may provide long term healthcare cost savings.Additionally, the clinician-patient relationship may be enhanced by clinicians offering PRx as a resource for families with FI.Offering resources in response to positive clinical screening may serve as a "trust catalyst" when engaging with patients about social determinants or sensitive diet-related chronic disease topics. 75The success of integrating medically tailored meals (MTMs) within major state and private insurance systems serves as an example for the potential future of PRx at-large. 76To demonstrate impact and sustainability of pediatric PRx interventions, the field must undertake hypothesis-driven studies, powered to measure clinically significant changes in FI, and health-related behaviors and health outcomes.Additionally, future studies should conduct cost-effectiveness analysis with respect to healthcare utilization to understand the benefit of PRx within a healthcare system.Ultimately, if proven effective, the route of sustainability would be for federal policy to align and insurance coverage of PRx as medical interventions.A question remains, whether PRx interventions can be short-term interventions or whether they need to be sustained for a longer period of time to see long-term impacts.The goal may be to intervene during a window of opportunity, to expose, educate, and establish longterm healthy behaviors that can alter an individual's trajectory of disease risk, putting them on a path of healthy lifestyle behaviors that will last a lifetime.
This study has some limitations.Eligible studies were limited to PRx interventions within a healthcare system, which excluded community and FNPs.20]77 Search criteria were limited to the U.S. and excluded studies conducted before 2000.However, most studies were published within the last two years and have broad geographic range, making this scoping review timely and somewhat generalizable.Lastly, being a scoping review, the study was unable to measure aggregated outcomes and effect sizes across studies.

CONCLUSION
The studies explored here show that PRx interventions may support household food security and improve nutritional knowledge and F/V intake in children and adult caregivers.While quantitative health metrics are currently lacking, longer-term evaluation of pediatric PRx outcomes will help identify the impact of these programs on health outcomes across the lifespan.Additional research in this field should continue to explore qualitative experiences while also incorporating more rigorous study designs, larger sample sizes, quantitative analyses of behavioral and health outcomes, cost-effectiveness assessments, and tracking of healthcare utilization.This work will require a multi-disciplinary approach that includes the family unit, community resources, the healthcare team, payers, and trained evaluators with standardized tools to offer the most useful information to promote the advancement and integration of PRx interventions.

Fig. 2
Fig.2Description of produce prescription intervention types.There were a total of 19 interventions evaluated, intervention models fell into two categories: voucher programs (n = 14) or box/pantry programs (n = 5).Thirteen voucher programs utilized farmers' markets (FM) or retailers and one used an online produce market.Two voucher programs allowed redemption for other foods in addition to produce.Among box programs, two allowed self-selected produce and non-perishable items and three programs offered pre-selected F/Vs only.One provided delivery services and the rest were pick-up.

Table 1 .
Search strings used in database searches.

Table 2 .
Characteristics of Produce Prescription (PRx) articles included in scoping review.

Table 2 .
continued g FI, Food Insecurity; h F/V, Fruits and Vegetables; i BMI, Body Mass-Index; j FM, Farmers' Market; k NNE, No Nutrition Education; l NA, not applicable; m EFNEP, Expanded Food and Nutrition Education Program; n USDA, U.S. Department of Agriculture; o NR, Not reported; p HVS,

Table 3 .
Considerations for Implementation of a Pediatric Produce Prescription (PRx) Intervention.