Community-based fruit and vegetable prescription programs: a scoping review

Abstract

Identify and categorise different models of community-based fruit and vegetable prescription programs, to determine variation in terms of methodology, target population characteristics, and outcomes measured. Applying the scoping review methodology, ten electronic databases were utilised to identify community-based fruit and vegetable incentive programs. Results were evaluated by two independent reviewers, using Covidence software. All full-text reviews were completed and documented using the PRISMA-ScR guidelines. Search results were stored and reviewed within the Covidence software. Thirty full-text articles were utilised from the 40 206 identified in the search. Target populations were predominantly female, non-white, and low-income. Considerable heterogeneity was found in both study design and quality. Fruit and vegetable vouchers were utilised in 63 % (n 19) of the studies. Prescriptions were primarily provided by community health centres (47 %; n 14) or NGOs (307 %; n 9) and could be redeemed at farmers’ markets (40 %; n 12) or grocery stores (27 %; n 8). When measured, diet quality significantly improved in 94 % (n 16), health outcomes significantly improved in 83 % (n 10), and food security status improved in 82 % (n 10) of studies. Providing financial incentives to offset the cost of fresh fruits and vegetables can increase consumption, improve health outcomes, and improve food security status. The majority of studies showed significant improvements in at least one outcome, demonstrating the effectiveness of community-based fruit and vegetable prescription programs. However, the diversity of measurement techniques and heterogeneity of design, dosage, and duration impeded meaningful comparisons. Further well-designed studies are warranted to compare the magnitude of effects among different program methodologies.

Introduction

It has been well documented that individuals residing in the United States do not consume adequate amounts of fruits and vegetables (F&Vs).^(1–3) According to the Dietary Guidelines for Americans (DGAs), adults should consume approximately 1⋅5–2 cup-equivalents of fruits (approximately 120–160 g) and 2–3 cup-equivalents of vegetables (approximately 160–240 g) daily.⁽³⁾ Presently, 10 % of adults meet the vegetable recommendation and 12⋅3 % of adults meet the fruit recommendation.⁽²⁾ Adequate intake of F&V may reduce the risk of developing obesity, cardiovascular disease, and diabetes.⁽⁴⁾ Schwingshackl et al. found that consuming up to 300 g of vegetables per day decreases the risk of mortality by 11 % and consuming up to 250–300 g of fruit per day decreases the risk of mortality by 10 %.⁽⁵⁾

Many factors contribute to the inadequate intake of F&V among Americans, however the most prominent barrier cited is cost.⁽⁶⁾ It has been shown that a 10 % decrease in price of healthful foods increases consumption by approximately 12 %, and specifically increases consumption of F&V by 14 %.⁽⁷⁾ Multiple programs have been created at the state and municipal levels to increase consumption of F&Vs by reducing costs.^(8–19) The most widely used program focuses on increasing F&V dollars for Supplemental Nutrition Assistance Program (SNAP) recipients by doubling or matching^{(8,9,11–13)} or providing rebates⁽¹⁰⁾ on dollars spent on F&Vs. These programs often require participants to purchase F&Vs at farmers’ markets.^{(8,9,11–13)} The findings are encouraging as all have shown an increase in F&V purchases and/or consumption among participants.^(8–13) Similar results have been seen in programs aimed at increasing consumption of F&Vs among Supplemental Nutrition Program for Women, Infants, and Children (WIC) participants.^(14,15) Additionally, programs have been developed to offer direct vouchers of varying amounts^(16–18) or rebates⁽¹⁹⁾ to participants. These programs have resulted in an increase of F&V consumption across the United States.^(16–19) These data support that decreasing costs may increase purchases and consumption of F&Vs.

The U.S. government has allocated $25 million to pilot additional produce prescription programs which provide F&Vs as well as education.⁽²⁰⁾ At the present time, there are multiple models, but no clear consensus on which produce prescription programs are the most effective. This in part may be due to the different settings in which programs are administered. Engel and Ruder conducted a scoping review to identify F&V incentive programs for SNAP participants, documenting the different program structures.⁽²¹⁾ Whereas Veldheer et al. conducted a scoping review of the different types of program healthcare organisations implement to improve F&V access.⁽²²⁾ These reviews provide great insight into two important populations; however, there are still many other produce prescription programs that are not classified into one of these two groups. Therefore, the objective of this scoping review is to identify and categorise the different models of fruit and vegetable prescription programs offered in a community-based setting, in order to understand how they vary in terms of methodology, population characteristics, and outcomes measured.

Methods

This study utilised the framework for scoping reviews presented by Arksey and O'Malley⁽²³⁾ along with the recommended enhancements from Levac et al.⁽²⁴⁾ The five stages of the framework include: (1) Identifying the research question; (2) Identifying relevant studies; (3) Study selection; (4) Charting the data; and (5) Collating, summarising, and reporting the results.⁽²³⁾ The research objective was defined prior to data collection; however, following Levac et al. ⁽²⁴⁾ recommendations, the objective was slightly altered after the relevant studies were identified (stage two), as the original objective was deemed too broad.

Data collection

Data collection was completed between June 2021 and September 2021. To ensure a robust review, the researchers sought the assistance of the university research librarian. The librarian assisted in developing search streams using text terms and Medical Subject Heading (MeSH) terms. The text terms included: fruit, vegetable, produce, veggie, prescription, health promotion/s, health campaign/s, wellness program/s, incentive*, food voucher/s, food assistance, food stamp/s, school meal/s, school lunch/es, food pantry/ies, food bank/s, and the MeSH terms included: fruit, vegetable, crops, agricultural, health promotion, prescriptions, and food assistance. The researchers used these terms in seven electronic databases including PubMed (Medline), Scopus, Google Scholar, CINAHL, USDA website, and Cochrane Database. Additionally, the Gray Literature Database, OAIster, ProQuest, and MedNar were used to ensure a comprehensive search. In order to capture all possible literature, date parameters were not set for the search.

Due to the nature of a scoping review, the inclusion and exclusion criteria were broad. To be included, the article must describe a community-based program that incentivised F&V consumption, must be written in the English language, and the full-text article must be available. Community-based programs were defined as any F&V prescription being distributed or redeemed within the community setting. During the screening process, F&V prescription programs were defined as an intervention delivering a repeated prescription for F&V to address a diet-related health risk including food insecurity. There were four key elements that determined if a ‘program’ qualified: there must be a prescription for F&V, must require redemption or receipt of F&Vs, there must be a repeated dosage, and participants must have a diet-related health risk including food insecurity. Exclusion criteria included duplicate articles, non-intervention studies, and F&V prescription programs administered exclusively by a healthcare organisation as defined by Veldheer et al. ⁽²²⁾

All articles from the original search were imported into the Covidence software, which is used to manage systematic reviews.⁽²⁵⁾ This software removes duplicate articles, maintains accurate count of articles, and allows inclusion and exclusion criteria to be used to categorise articles. Covidence allows for multiple researchers to complete title and abstract screening and full-text review independently of each other by categorising articles as relevant or irrelevant. Additionally, any discordant vote is categorised as a conflict. The initial step of this review consisted of the title and abstract screening, the second step included full-text review, and the final step consisted of data extraction. The two independent researchers met on a weekly basis to discuss all conflicts and reach a consensus on the relevance of the disputed article(s), as well as determine that all necessary data had been extracted.

Data analysis

There were eleven unique variables collected, most of which were categorical. The variables included duration of intervention (weeks), number of participants, model of F&V prescription program (F&V vouchers, cashback rebates, F&V delivery or collection, or combination), education provided during intervention (none offered, cooking skills/classes, with nutrition professional, handouts, or combination), F&V prescription provider (community health centre, school, federal/state program, NGO or charity, or other), F&V produce provider (food bank/pantry, farmers market, grocery store, or combination), targeted age group of intervention (children, adults, or families), target population's health status (overweight/obese, diabetes, hypertension, multiple health conditions, or no health condition), recipient of government assistance (yes or no), diet quality measured (yes or no), health outcomes measured (yes or no), and food security status measured (yes or no). Extracted data was compiled in a Microsoft Excel spreadsheet. Duration of community-based F&V prescription program and number of participants were reported as means (sd), and frequency distribution was used for the remainder of the variables.

Results

The search identified 27 302 unique records after the removal of duplicates by Covidence software, see flow diagram in Fig. 1 detailing the study selection process for community-based F&V prescription programs. A further 27 056 were removed following title and abstract screening against the inclusion criteria, leaving 246 for full-text review which yielded 67 articles for data extraction. During the review, researchers further refined the search parameters to include the definition of ‘prescription program’ as an intervention delivering a repeated prescription for healthy produce to address a diet-related health risk including food insecurity. Following this refinement, thirty articles were included in the final review. A summary of the study characteristics of the community-based F&V prescription programs is presented in Table 1.

Fig. 1. PRISMA flow diagram detailing the study selection process for community-based fruit and vegetable prescription programs (separate document).

Table 1. Summary of community-based fruit and vegetable prescription programs

F&V, fruit and vegetables; FI, food insecurity; FPL, federal poverty level; Hb, haemoglobin; BMI, body mass index; FM, farmers’ market; NGO, non-governmental organisation; T2D, type 2 diabetes mellitus; HbA1c, glycated haemoglobin; BP, blood pressure; F2F, face to face; SNAP, Supplemental Nutrition Assistance Program; FMNP, farmers’ market nutrition program; WIC, women, infants, children special supplemental nutrition program; FS, food security; WC, waist circumference; CGM, continuous glucose monitor; HIV, human immunodeficiency virus; VLFS, very low food security; ART, antiretroviral therapy; HH, households; FBG, fasting blood glucose; HTN, hypertension; EBT, electronic benefits transfer.

^abcSuperscript letters denote studies carried out on the same intervention.

The thirty studies included in this review comprise twenty-seven unique interventions taking place in the USA (n 25), France (n 1), and Australia (n 1). The mean (sd) F&V program duration was 35⋅2 (25⋅0) weeks, with the most common length being 26 weeks. The number of study participants ranged from 10⁽³⁴⁾ to 2,859,^(31,37) with a mean (sd) number of participants of 366 (713). Table 2 shows the frequency distribution of categorical variables within program methodology, population characteristics, and outcomes measured in the community-based F&V prescription programs.

Table 2. Distribution of variables within program methodology, population characteristics and outcomes measured in community-based F&V prescription programs

Program methodology

The most common F&V prescription intervention model was F&V vouchers (n 19), which provide vouchers to participants to be redeemed for fresh fruit and vegetables.^{(27,31–33,35–42,44,47,50–54)} The second most frequent model was delivery and/or collection (n 8), with fresh F&V delivered to participants or made available for collection.^{(26,28,29,34,43,45,46,55)} Lastly, two studies used a cashback rebate on F&V purchases, via EBT card and associated with enhancement to SNAP benefits.^(48,49) In addition, one study employed a combination of these models.⁽³⁰⁾

Another variable within program methodology is provision of education, ranging from simple distribution of information, cooking tips and recipes via handouts^(26,31,37) (n 3), to cooking classes⁽³²⁾ (n 1), to regularly scheduled education sessions led by nutrition professionals^(38,48,49) (n 3). The most common approach (n 14) was to employ a combination of formats.^{(28–30,33–36,40,42,44,45,51–53)}

Providers of both prescriptions and produce varied depending on the program. The most common provider of F&V prescriptions (n 14) was community health centres,^{(26,28,29,32,33,38,41–47,50,52)} followed by non-governmental organisations, other non-profit, and charitable foundations^{(30,39,40,46,48,49,51,54,55)} (n 9). Other providers included federal/state programs^{(27,31,37,53)} (n 4) and schools⁽³⁴⁾ (n 1). The produce mainly came from one of three sources: Farmers’ markets^{(30,32,33,39,41,47–53)} (n 12), grocery stores^{(27–29,31,37,40,44,54)} (n 8), or food banks and pantries^(26,45,46) (n 3), with several (n 7) utilising a combination.^{(34–36,38,42,43,55)}

Population characteristics

The majority of F&V prescription interventions exclusively included adults^{(26,27,32,33,38–41,43,44,46,47,51–55)} (n 17), 30 % targeted families^{(30,31,34–37,45,48,49)} (n 9), and children made up the target population for the smallest proportion of the studies^{(28,29,42,50)} (n 4). Exactly half of the studies targeted populations with a specific health status, most commonly type 2 diabetes^{(32,33,41,43,52,54,55)} (n 7), followed by overweight and obesity^(34,48,49) (n 3), others included hypertension⁽⁵³⁾ and HIV-positive status.⁽⁴⁶⁾ Participation in a government assistance program was required in 27 % (n 8) of studies,^{(28–31,34,37,39,54)} although all of the studies described one or more indicators of low income or food insecurity measurement.

Other trends were seen within the target populations; however, these could not be collated as frequencies due to the differences in reporting between studies. Target populations were majority female,^{(26,27,30,32,33,35,36,39,40–44,47–55)} and majority non-white,^{(26–30,32,33,38,39,41–44,48,49,50,53–55)} where specified.

Study designs

There was considerable heterogeneity in the design of the different studies, with three main study designs used. The first being non-randomised observational study designs (n 24);^{(26–30,32,34,38,39–50,52–55)} the second, randomised controlled trials (n 5);^{(31,33,35–37)} and the third, non-randomised clinical trial (n 1).⁽⁵¹⁾

Outcomes measured

Change in dietary quality was measured in the majority of studies^{(27,29,30,34,35,37,40–46,49,50,51,53)} (n 17), only one of which reported no statistically significant improvement in diet quality.⁽³⁴⁾ The number and type of instruments used to measure diet quality varied widely, as did their levels of reliability, these included food frequency questionnaire (FFQ)⁽³⁵⁾ (n 1), validated 24-hour recall methods^{(27,29,40,51)} (n 4), HEI scores^(27,51) (n 2), and nutritional biomarkers⁽²⁹⁾ (n 1). Several of the studies used externally validated screeners to measure F&V intake, including NCI dietary screener^(37,41,49) (n 3), the Block screener⁽⁵⁰⁾ (n 1), and the Behavioural Risk Factor Surveillance System (BRFSS)⁽⁴²⁾ (n 1), in addition to other instruments not often cited in the literature^(44,46,53) (n 3). Internally designed, non-validated surveys were used to assess F&V intake in the remainder^{(30,34,43,45)} (n 4).

Outcomes related to health status were also measured in eleven of the studies,^{(28,32,33,38,42,43,46,47,52,54,55)} with all but one⁽⁵⁴⁾ reporting a statistically significant improvement in at least one health measure. All outcome measures were conducted by a trained medical or research staff, or retrieved from medical records, none relied on self-report. The different health outcome measures included frequency of sick visits to healthcare provider and/or hospital attendances⁽²⁸⁾ (n 1), number of oral antibiotic prescriptions⁽²⁸⁾ (n 1), iron status⁽²⁸⁾ (n 1), weight^{(32,43,47,55)} (n 4), BMI^{(28,38,42,46,47,54,55)} (n 7), blood pressure^{(32,33,43,47,54,55)} (n 6), glycated haemoglobin (HbA1c) level^{(32,33,43,46,52,54,55)} (n 7), waist circumference^(43,47,55) (n 3), glycaemic control (continuous glucose monitor and fasting blood sugar measurements)^(43,46,47) (n 3), and antiretroviral therapy adherence⁽⁴⁶⁾ (n 1).

Food security status was also an outcome of interest in eleven of the studies,^{(26,31,34,36,42,43,45,46,48,50,55)} with only one study not reporting a statistically significant improvement in food security status.⁽⁵⁰⁾ There was less variation in the measurement instruments compared to diet quality measurements, with the majority using the widely used USDA 18-item Household Food Security Survey^{(31,36,46,48,55)} (n 5). Others used the 2-item⁽²⁶⁾ (n 1) and 6-item^(42,43,50) (n 3) versions which have been shown to have high sensitivity and specificity when compared to the original Household Food Security Survey tool.^(56,57) Two studies used internally designed, non-validated surveys.^(34,45)

Discussion

The results of this review show encouraging outcomes for individuals who receive and utilise produce prescriptions from community-based organisations. Similar to previous studies that showed increased consumption of F&Vs when financial incentives are provided,⁽⁷⁾ SNAP benefits are increased,^(8–13) WIC benefits are increased,^(14,15) or when direct vouchers^(16–18) or rebates⁽¹⁹⁾ are provided, the majority of community-based produce prescription programs showed an increase in F&V consumption when a financial incentive was offered.^{(27,34,35,37,41,42,44–46,49,51,53)} A minority of the studies only showed increases in vegetable^(30,43) or fruit consumption,^(40,50) with even fewer showing no statistically significant increases in overall F&V intake.^(29,40) Also, two studies showed an increase in F&Vs purchased but did not report pre–post consumption data.^(39,54) These results further support the benefits of providing financial assistance to offset the often-cited barrier of purchasing F&V, which is cost.⁽⁶⁾ The community-based produce prescription programs are in fact assisting individuals increase F&V consumption moving them closer to meeting the DGA's recommendations.⁽³⁾

Lessening the financial burden associated with purchasing groceries may directly lead households or individuals towards improved food security status. This trend was consistently seen in ten of the articles reviewed,^{(26,34,36,37,42,43,45,46,48,55)} especially among individuals classified as very low food secure and low food secure.^(43,46,48) Interestingly, the article that did not find statistically significant improvement in food security status also did not see an increase in F&V consumption.⁽⁵⁰⁾ Whereas, all other studies that measured both food security status and F&V intake showed statistically significant improvements in both indices.^{(34,42,43,45,46)} These results indicate a possible correlation between the two variables.

As previously discussed, increasing F&V intake may reduce the risk of developing certain chronic diseases and decrease overall mortality risk.^(4,5) Although many of the studies reviewed measured certain health-related indices,^{(28,29,32,33,38,42,43,46,47,52,54,55)} only four measured health indicators and actual F&V intake.^{(42,43,46,54)} Additionally, only Jones et al. ⁽⁴²⁾ (BMI z-score) and Kerr et al. ⁽⁴³⁾ (waist circumference, systolic blood pressure, HbA1c, glucose, and weight in women) showed statistically significant improvements in health outcomes with an increase in F&V consumption. Diabetes management was a focus for many of the studies,^{(32,33,43,47,52,55)} with five showing improvement in HbA1c^{(32,33,43,52)} and fasting blood glucose levels⁽⁴⁷⁾ during the produce prescription intervention timeframe. Though encouraging, it is difficult to determine what specific factors contributed to improved glycaemic control during these interventions, as dietary data was not reported.

There are clearly opportunities to utilise community-based incentive programs to improve both dietary quality, food security status, and health outcomes, which has been supported by previous scoping reviews exploring the application of F&V prescription programs in other areas.^(21,22) However, they too have made similar comments regarding the variety and heterogeneity in design and quality of programs. Focusing specifically on healthcare organisations, Veldheer et al. ⁽²²⁾ found most studies assessing dietary quality showed improvements, although health-related outcomes were more mixed. Engel and Ruder found some measure of positive impact in all but one of the studies in their review of F&V incentive programs for SNAP participants.⁽²¹⁾ The current review adds to the literature that providing produce prescriptions in a variety of community settings, outside of healthcare organisations and SNAP, improves the overall intake of F&V, food security status, and certain health outcomes. The similarities among the programs could provide a standardised produce prescription blueprint to be implemented across different settings.

This scoping review has some definite strengths, it was conducted using the methodological framework from Arksey and O'Malley⁽²³⁾ with enhancements from Levac et al. ⁽²⁴⁾, and was conducted following the PRISMA-ScR guidelines and checklist.⁽⁵⁸⁾ The comprehensive database search was guided by a specialist, and all screening, data extraction, and mapping were conducted by two reviewers independently. There are some limitations, the search revealed several abstracts that met the criteria but were unable to locate the full-text article and therefore not included, this may have meant some relevant data could have been missed. In addition, following the broad format for a scoping review may have meant that search was too wide, a more targeted research question may have helped reduce the heterogeneity found across the interventions. Finally, no meaningful statistical analysis could be completed due to the lack of experimental design of most of the studies, plus the limited comparison groups in the non-experimental study designs.

Conclusion

This scoping review of community-based interventions delivering a repeated prescription for healthy produce to address a diet-related health risk found that most studies showed a significant improvement in one or more of the outcomes measured. However, the diversity of measurement tools made meaningful comparisons of the effectiveness of the different programs impossible. In addition, the heterogeneity of program design, dosage and duration, and the format and frequency of education, also make meaningful comparisons difficult. So, while most outcomes showed statistically significant improvements, these limitations raise questions as to the strength of the evidence.

In summary, further study is warranted to compare the magnitude of effects with the different program methodologies, and the education component requires further investigation to understand its contribution to program effectiveness. Most of the studies included in this review were non-randomised and did not have a control group, more rigorously designed and adequately powered RCTs are required to accurately evaluate these types of intervention. Additionally, there are other barriers to the intake of fresh fruit and vegetables besides cost, including access and acceptability, these also require further research. After completion of this review, recommendations for next steps would be to conduct larger randomised control trials to determine the effectiveness of produce prescription programs.