Report

Bringing Produce to the People: Implementing aSocial Marketing Food Access Intervention inRural Food Deserts A. Susana Ramirez, PhD, MPH1; Lillian K. Diaz Rios, PhD, RD1,2; Zulema Valdez, PhD3;Erendira Estrada, BA1; Ariana Ruiz, BA1

1Departmof Califo2Divisio3DepartmConflict owith thiAddressSchool oNorth Lucmerce�2016 Sreservedhttp://dx

166

ABSTRACT

This study describes and evaluates the process of implementing a social marketing food access interventionfor food desert communities in rural California. A case study approach used mixed-methods data fromnationwide market comparisons, environmental assessment, and community informants. Lessons learneddemonstrate room for improvement in implementing such strategies and underscore the importance ofinvolving community in decision making; the strategic importance of operational decisions relating tointervention design, site and product selection, and distribution models; and the need to reconsider theproblem of access in rural areas.KeyWords: community development, rural health, food access, social marketing (J Nutr Educ Behav.2017;49:166-174.)

Accepted October 28, 2016. Published online December 9, 2016.

INTRODUCTION

Rural food deserts (places located > 10miles from a supermarket) often lack ac-cess to fresh produce; cluster in low-resource, low-income, ethnic minoritycommunities; and are associated withdisproportionate rates of poor healthoutcomes and chronic disease amongresidents.1 Interventions aimed at elim-inating food deserts have includedbuilding permanent structures (eg, full-service supermarkets) and remodelingexisting small stores to sell fresh pro-duce,2-5 as well as temporary solutions(eg, farmers' markets).6-8 Such interven-tions have had mixed success and havebeen conducted mostly in urban areas.The nature of rural areas (sparsely

ent of Public Health, School of Socialrnia, Merced, CAn of Agriculture and Natural Resourceent of Sociology, University of Califf Interest Disclosure: The authors’ conflis article on www.jneb.org.for correspondence: A. Susana Ramirezf Social Sciences, Humanities, and Arake Rd, Merced, CA 95343; Phond.eduociety for Nutrition Education and Beh..doi.org/10.1016/j.jneb.2016.10.017

populated, with minimal developmentand commercial areas)means that someof these interventions are not trans-ferable to rural settings.1,9 For example,larger chain stores often select areasbased on their potential for profit-abilityandwhere insuranceandsecuritycosts are low; rural areas typically donotmeet those criteria.10 Innovative strate-gies are needed to increase access tonutritious foods in rural areas.11,12 Onerecent intervention that gained tractionis the development of mobile farmers'markets that deliver produce for sale ata reasonable price. Although mobilefarmers' markets first emerged in urbanareas, public health and policy makersconcerned about access to food in ruralregions focused on whether this inter-

Sciences, Humanities, and Arts, University

s, University of California, Merced, CAornia, Merced, CAct of interest disclosures can be found online

, PhD, MPH, Department of Public Health,ts, University of California, Merced, 5200e: (415) 580-0952; E-mail: sramirez37@

avior. Published by Elsevier, Inc. All rights

Journal of Nutrition Education and Beh

vention might provide fresh produce torural residents in low-resource areas. Thepurpose of this report was to describeand evaluate the process of implemen-tation for 1mobile farmers' market by anonprofit organization that aimed toincrease access to fresh produce in ruralfood deserts. Specifically, it details thesocial marketing–based interventionplan, its implementation, and lessonslearned.

DESCRIPTION OF THEINTERVENTION

From November, 2013 through May,2015, a nonprofit organization devel-oped Produce on theGo (POTG), amobilefarmers' market, to improve food accessin a rural central California county byproviding locally grown fruit, vegeta-bles, and nuts for purchase throughweekly visits by a mobile grocery truck.Merced is a largely rural county in cen-tral California with high poverty(25.4%) and unemployment (17.5%),and where 1 in 6 households (15.6%)is food insecure.13 Despite its agricul-tural bounty, this county is home tomany food deserts, defined as areasthatareat least10miles fromthenearestsupermarket or chain grocery store, inpredominately low-income Latino-ma-jority communities.14 Disparities in ac-cess to fresh produce because of the

avior � Volume 49, Number 2, 2017

Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Ramirez et al 167

lack of supermarkets and large grocerystores in food deserts are associatedwith residents' dietary intake and ratesof obesity and chronic disease.

Produce on the Gowas developed by anonprofit organization and funded bythe Merced County Human ServicesAgency, California FreshWorks, and theUS Department of Agriculture (USDA).The mobile food access interventionincluded 2 vehicles (a cargo van, pur-chased with funding provided by theseorganizations, and an 18-wheel refriger-ated semi truck, donated by a drinkdistributor) to sell fresh produce atseveralMercedCounty sites on aweeklybasis. Two vehicles allowed access todifferent types of locations. From 10sites launched in November, 2013, theintervention expanded to 19 sites,mostly in Latino-majority, low-incomecommunities.

Social marketing theory, which in-corporates commercial marketing prin-ciples in the planning and executionof behavior change interventions,

Figure. Logic model for social marketing fo

underpinned the POTG interventiondesign. Social marketing takes into ac-count: (1) exchange theory: consu-mers must perceive a benefit in exchangefor their participation/purchase/behaviorchange; (2) audience segmentation:subgroups similar in some way relatedto the target behavior or may respondsimilarly to intervention; (3) marketingmix: a combination of price, product,place, and promotion; (4) customerorientation; and (5) continuous moni-toring.15-17 Exchange theorywas fundame-ntal to the intervention design: POTGsought to provide the benefit of foodaccess to participants and customers.Audience segmentation was place-based, with each truck site servingdistinct audiences. Promotion activitiesincluded postcard mailers, roadsidesignage, community festival participa-tion, recipe cards and produce descrip-tions from the Supplemental NutritionAssistance Program–Education, and thehiring of community liaisons to pro-vide nutrition education and opportu-

od access intervention.

nities for physical activity at selectsites. Other components of the mar-ketingmix are further discussed underLessons Learned, along with customerorientation and continuous moni-toring. The Figure shows a logicmodeldetailing the intervention compo-nents and expected outcomes.

DESCRIPTION OF THEEVALUATION

Process evaluation of the POTG inter-vention followed a parallel-convergentmixed-methods case study design con-sisting of 3 separate research activities(Table 1).18 The aim of the evaluationwas to assess the implementation ofthe intervention as it was conceptual-ized, focused on the inputs, activities,and short- and mid-term outcomes:namely, residents' awareness of themo-bile food vehicles and potential forsustainability (Figure), and to deter-mine whether the intervention was

Table 1. Evaluation Approach and Results Summary (Parallel Convergent Mixed-Methods Design)

Component Objective Methods Sampling n Key FindingsIntegrated

Interpretation

Environmentalassessment

Understand contextualfeatures that affectPOTG’s effectivenessand sustainability

Visual inspection,checklist,photographs

4 sites identified forhigh-intensityintervention, includingnutrition education

– � Proximity to retail, ser-vices

� Schedule� Site accessibility� Competition

� Importance of com-munity engagement,ownership, involve-ment in decision mak-ing

� Strategic importanceof operational deci-sions

� Definition of access� Challenges to imple-mentation and in-tended effects

National marketassessment

Identify and comparebarriers and facilitatorsto mobile food trucksas food accessinterventions

Interviews withmanagers

Inclusion: All mobilegroceries/farmers’markets with missionto increase foodaccess

5 � Community engage-ment

� Sustainability chal-lenges (cost vs price,management, oper-ating schedule)

Communityinformantsa

Assess awareness andperceptions of POTG

Understand perceptionsof community foodaccess and solicitsolutions

Depth interviews

Focus groups

>17 yearsLive in interventionneighborhood.

30

7

� High awareness;shopping barriers

� Selection, quality� Pricing� Schedule/freq� Food access ¼ cost

Obtain demographicdata and perceptionsof food access fromstandardizedquestions

Surveys All interview and focusgroup participants.

37 � Healthy foods areavailable in neighbor-hoods

POTG indicates Produce on the Go.aRefer to Table 2.

168Ram

irezet

alJournalofNutritio

nEducatio

nandBehavior�

Volume49,

Number

2,2017

Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Ramirez et al 169

appropriate and accessible.19 Three sepa-rate evaluationactivitiesdelivereddistincttypesof information thatwere then in-tegrated to provide a more compre-hensive understanding of the processof implementation and consequentoutcomes thancouldbe achieved throughany singlemethod.20,21 These activitiesincluded (1) an environmental assessmenttounderstandhow structural factors atdifferent types of sites affected thesustainability of the intervention andhow residents perceived of and interactedwith the intervention (ie, the consumerexperience); (2) a national marketassessment that consisted of interviewswithmanagers of similar mobile produceinterventions and served to comparebarriers and facilitators to implementation,best practices, and structural featuresat intervention sites that might conditioneffectiveness or sustainability; and (3)focus groups, interviews, and surveyswith community informants, to assessresidents' perspectives on food accessin their communities, and their thoughtson and experiences with the interventionas appropriate and accessible. Surveysalso helped in understanding thedemographic profile of participants toassess whether the intervention reachedthe target population. Integration wasachieved in the design and inter-pretation,20 revealing how the variouscomponents of the intervention fittogether. The University of California,Merced, Institutional Review Boardapproved the components involvinghuman subjects.

Environmental Assessment

An assessment of 4 target neighbor-hoods, representing5of the19 interven-tion sites, aimed to understand theenvironmental or contextual featuresthat may have influenced the effective-ness of the produce truck as a strategyto improve food access in MercedCounty. Each location was physicallyvisited by 2 trained research assistantsat different times to allow for verificationof the observations. The data collectioninstrument (coding sheet) was based onan assessment of prior literature2-5 andthe parallel assessment of nationalproduce trucks, to capture contextualcomponents hypothesized to affectsuccess: parking; proximity to publictransportation and walk/bike paths;retailers/services; and homes/offices/other

locations where people might gather.Two researchers systematically codedthe 5 site assessments independently;intercoder reliability was near-perfect(Cohen's kappa ¼ 0.99).

National Market Assessment

A total of 5mobile interventions sellingproduce in food deserts in any region ofthe US were identified through the Cal-ifornia FreshWorks Web site and a gen-eral Internet search. Managers wereinterviewed to solicit informationregarding the process of implementingthe intervention, the appropriatenessand accessibility of the intervention(eg, the degree to which the interven-tion successfully engaged the targetcommunity), and the potential for thesustainability of the intervention. Theinterviewer recorded extensive inter-view notes, which were deductivelyanalyzed for themes regarding commu-nity engagement and sustainability.22

Interventions were excluded if: (1) theintervention was not yet active; (2)managers could not be identified owingto insufficient contact information; and(3) the intervention operated exclu-sively outside the US.

Community Informants: DepthInterviews, Focus Groups, andSurveys

Depth interviews (n ¼ 30) and focusgroups (n ¼ 7 participants; 2 focusgroups) were conducted by trainedbilingual (English/Spanish), bicultural(Mexican American) research assistantsto obtain feedback about the appropri-ateness and accessibility of POTG andfood access issues within the commu-nity. Informants were recruited nearintervention sites. The interview guidewas pretested in mock interviews withcollege students. Informants alsocompleteddemographicquestionnaires(Table 2 lists the data). Each focus grouphad a dedicated moderator and note-taker.23 Audio recordings were tran-scribed verbatim. To preserve contentintegrity, Spanish transcripts were nottranslated; all analyses were conductedby bilingual, bicultural researchers. Thecodebook was based on domains ofinterest to the intervention managers,as per the logic model (Figure), whichincluded the following domains: aware-ness, perceptions, and expectations of

the intervention; barriers and facilita-tors to shopping POTG; and food accesschallenges.22,24 Transcripts were sepa-rately coded by 2 researchers. Res-ponses were organized first into thedomains of interest; themes and patt-erns were then identified within eachdomain, and codes were systematicallyassigned to quotations representingthemes or domains. A third researcherreviewed the coding scheme and resolveddiscrepancies. Saturation occurred atthe seventh interview for awareness/perceptions, and the 11th interviewfor food access challenges.

LESSONS LEARNED

Lessons learned are syntheses of out-comes fromeachof the3 separate evalu-ation activities, each of which provideddistinct types of information; inte-grating these data through interpreta-tion assisted with understanding howthevariouscomponentsof the interven-tion fit together. Integration involvediterative comparisons of the resultsfrom each of the data streams toexaminetheextent towhichthevarioustypes of data confirm, contradict, orexpandunderstandingof thephenome-non in question.

Importance of CommunityEngagement and Involvement inDecision Making

Fostering relationships basedonmutualrespect and trust is a critical first step todeveloping successful interventions toincrease foodaccess in low-resource,mi-nority communities. Building relation-ships and partnerships creates thefoundation to engage with communitymembers regarding their perceptionsof the neighborhood food environ-ment, including identifying their needsand input on potential solutions.Engaging neighborhood residents re-quires a strongcommunity-basedmeth-odology from the planning to theimplementation stages of the interven-tion.25,26 In this mixed-methods casestudy, the authors found evidence thatinterventions conducted with engagedcommunities were more successfulthan those in which residents (ie, inter-vention targets) were not fully engaged,including POTG.

Table 2. Characteristics of Community Informants

Characteristics % (n ¼ 37)

Age, y (mean [range]) 39.0 (18–75)

GenderFemale 84Male 16

LanguageSpanish 62English 38

Aware of POTG 32

Ever shopped at POTG 30

Education, highest levelElementary 30Grades 7–8 14High school diploma/General Educational Development 24Some college/technical training 24Some graduate/professional school 3

EmploymentFull-time 37Part-time 14Not employed 49

Receiving food assistance 44

Children in household, n<5 496–18 49

Body mass index, kg/m2

<25 1925–29 32$30 46

POTG indicates Produce on the Go.Note: Sample included residents from 2 high-priority communities served byPOTG. Although demographically similar, the 2 communities were selected torepresent 2 ends of the urbanization spectrum in the county: 1 was suburban(South Merced) and the other rural (Planada). Information for residents whocompleted depth interviews (n ¼ 30) and those who participated in focus groups(n ¼ 7) is combined; there were no differences in group characteristics.

170 Ramirez et al Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017

Governance structure. The nationalassessment revealed consistency in thegovernance structures of the interven-tions: With the exception of POTG, allhad advisory boards that were com-posed of representatives of the commu-nity; were managed, operated by, orotherwise affiliated with an existingnonprofit community organization;and had advisory boards that wereactive participants in the interventiondesign and decision-making processes.The advisory boards of successful inter-ventions included representatives fromthe community; boards were convenedregularly and frequently, and/or indi-vidual members were consulted, asmanagers thought appropriate. Produceon theGo's advisoryboardwascomposedof administrative officials from local

governmental organizations, funders,and academics, and met only sporadi-cally. Thus, decision making was left tothe operations manager.

Perceptions of POTG. Nearly onethird of community informants wereaware of POTG, and nearly all whowere aware had shopped there(Table 2). Community informants hadpositive perceptions of POTG initially;however, they reported a decline inquality over time: ‘‘When it first startedit had excellent fruit . lately . theyweren't getting the produce as fresh’’(Participant [P]5). Informants reportedthat the quality of POTG produce wasnot good; specifically, produce was

‘‘not as fresh as some at the grocerystores’’ (P3) or just ‘‘old’’ (P12).

Strategic Importance ofOperational Decisions

Successful food access interventionsgrounded in a strong community-based methodology25,26 demonstratedcustomer orientationby soliciting inputfrom community members to makedecisions regarding all aspects of day-to-day operations, including the loca-tion and timing of community visits,the produce selected, price, and accessi-bility. Involving community membersin operational decisions helps to meetthe foodaccess needsof the community(as identified by the members them-selves), facilitates trust, and encouragesbuy-in from community members. Incontrast, most of the operational deci-sion making for POTG was made by asingle individual—the operations man-ager—with little input fromcommunitymembers or members of the advisoryboard.

Location. Produce on the Go interven-tion sites varied in proximity to retailand services. For example, 1 site wasnear train tracks; when trains passed,dirt was kicked up and the noisedrowned out conversation. Other siteswere where people visited on a regularbasis (eg, health clinic, elementaryschool, post office, Special SupplementalNutrition Program for Women, Infants,and Children clinic). In contrast, all ofthe sites described by other interven-tions in the national assessment weredeliberately selected to be part of some-thing larger than the intervention.

Timing. The day of the week and timeofdayvariedacrossPOTGsites.A schoolsite visit was scheduled in the earlymorning, allowing children andparentsto browse the selection before school;the visit was long enough to allowteachers to bring students for morningsnack. In contrast, the visit to the sitenext to the train tracks occurred on aSunday morning, on the opposite endof town from the Catholic Churchwhere most residents attendMass regu-larly. In addition, community infor-mants indicated that infrequent visitsmade it hard for them to includePOTG in their regular shopping. Man-agers of other interventions included

Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Ramirez et al 171

in the national assessment had madesimilar observations and reported thatthey had addressed this issue by visitingfewer sites with greater frequency, andhad decided which sites to add or drop inpartnership with community residents.

Accessibility. Spaces suitable for park-ing POTG's 18-wheel semi truck wereusually distant from public transporta-tion. This was less of a problem for thePOTG van and for the comparison pro-duce trucks in the national assessmentthat operated out of converted buses orother, smaller vehicles that could ma-neuver more easily in tighter quarters.

Competition. Some intervention siteswere near neighborhood markets orvendors that soldproduceatprices com-parable to POTG. Informants ment-ioned that the quality and selection ofthe competitors' products, as well asthe presentation (eg, 1 competitor wasa Mexican man selling produce out ofa weathered pickup truck), negativelyaffected perceptions of POTGcomparedwith alternatives and also may haveaffected shopping behaviors.

Product selection. Community infor-mants indicated that the mobile foodintervention did not meet their needs;as a result, participation in the food ac-cess intervention declined over time.For example, some expressed thatPOTGdidnot offer the types of producetheywereused toeating;offering staples(eggs,milk, beans, cilantro, and tropicalfruits such as avocados and mangos)would make them more likely to shopthere regularly. Informants also obser-ved a problem with an inconsistencyof available produce and a lack of clarityabout why certain products were some-times unavailable. Research on food ac-cess interventions in low-resource,ethnic minority communities under-scores the critical need for a strongcommunity-based strategy at all stagesof planning and implementation. Solic-iting input from community residentsat theplanning stagewouldhave identi-fied key products to stock; input duringthe implementation likely would haverevealed neighborhood residents' con-cerns and provided an opportunity fora correction. Managers of comparisoninterventions reported a commitmentto the social marketing principle ofcontinuous marketing that generated

this typeofongoingfeedbackfromcom-munity residents, which formed the ba-sis for decision making about productselection and increased their success.

Definition of Access

Researchon foodaccess typicallyunder-scores physical or spatial disparities inthe distribution of food outlets acrossdifferent communities (eg, food desertsor not; rural vs urban). Yet, recentstudies reveal that the cost of food alsolimits food access in areas designatedas rural fooddeserts. Food access in ruralfood deserts requires a consideration ofphysical access in terms of proximityto food outlets, and price,27,28 and thiswas borne out in this study.

Perceptions of food access. All com-munity informants lived in USDA-designated food deserts, yet only 14(37.8%) reported that they could notpurchase healthy foods near theirhomes. When the issue of accessibilitycameupasabarrier toeatinghealthfullyin the interviews/focus groups, itwas related to financial and time-dependent access rather than physicalaccess.Thesedefinitionsshapedthesug-gestions that informants had forimproving access to healthy foods intheir communities, and their percep-tions of POTG.

Price and time. The issue of reducingthe high costs of healthy foods toimprove access came up repeatedlyand was often paired with timing. As 1informant remarked,

The accessibility to get them. Someofthe veggies are high cost or they're notin season,. they have to bring themout of different states and when itcomes to that it’s a higher price .and such is not in the budget .(P13)

Similarly, another informant indi-cated, ‘‘Time! Time and maybe money,too, cause all the healthy and good stuffis usuallymore expensive than the junkfood’’ (P4). Some informants discussedtime in terms of food preparation:

[Fresh produce doesn't] have preser-vatives like junk food does, and usu-ally when you are busy and by thetime you get to something later onin the week, it's already rotten .

but you pretty much have to maketime to do that stuff. (P6)

For others, high cost was related tocost in terms of time:

. Sometimes when there is gooddeals on stuff there [are] a lot of peo-ple there and sometimes people don'thave time to be waiting in line allday. (P6)

Pricing. Consistent with their defini-tions of food access, the informantsevaluated POTG based on pricing.Some thought prices were competitive;others thought they were high. Stillothers reported inconsistency in pric-ing, which was perceived as a barrier toshopping POTG.

Challenges

The original intent of POTG was to in-crease food access so as to improve die-tary behaviors and ultimately tocontribute to preventing and reducingobesity among the rural populationthrough a social marketing approach(Figure). Although no other studiedinterventionhadsuchanexplicitmodelof effects, the national assessment re-vealed similar expectations across allgroups. The next sections describesome challenges to implementationand sustainability and a comparisonwith managers' experiences from thenational assessment.

Marketing mix. An important compo-nent of POTG's social marketingapproachwas the4P's: thecombinationof price, product, place, and promotionthat comprise the marketing mix. Thenational assessment revealed that suc-cessfulmobile food access interventionsreported substantial engagement withcommunities and spent considerableeffort conducting a strategicmixofmar-keting activities that included residentsin decision making about price, prod-uct, and place (ie, location). In contrast,POTG had less transparency about thefirst 3 P's and only minimally engagedin the fourth promotion. Nationally,successful trucks promoted themselvesthrough coupons, targeted advertise-ments, and social media. Nutrition edu-cationwasacorecomponentofmarketingactivities formost, althoughexecutionvar-ied substantially. All provided recipes.

172 Ramirez et al Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017

Some had regular food demonstra-tions including tastings; others part-nered with other organizations forcommunity health fairs and blockparties to engage community mem-bers with produce on offer and morebroadly, with the truck as food retailer.The national assessment managersmentioned several of these advancedsocial marketing strategies as instru-mental for engaging communitiescredited them for those interventions'successes. Promotion activities thatwere enacted for POTG included directmailings to residents of interventioncommunities, signage near the sitesor directing traffic to the sites, andparticipation in community festivals.These activities may have served to in-crease residents' awareness of the pro-gram. However, a major challenge, asdescribed earlier, was that residentsdidnot remember that the truckwouldbe visiting on the actual day andconsequently failed to integrate shop-ping at the site within their regularfood shoppingplans. Promotion activ-ities that might have served as moretimely reminders to residents, such asa Web site with an updated calendaror social media postings, were not im-plemented. For a largely Latino, socialmedia–savvy, and mobile phone–dependent29 community, these un-used communication strategies mayhave been more useful. Other socialmarketing aspects of the interventionnot implemented included nutritioneducation and food preparation dem-onstrations to address knowledge andself-efficacy as mechanisms for behaviorchange.30 Even concepts that wererelatively more straightforward tocommunicate, such as those to in-crease awareness of the interventioncomponents, appeared to have failed.For example, although nearly half ofthe informants received food assis-tance, almostnoneknewthat the truckaccepted Electronic Benefits Transfer.In social marketing terms, this failureto promote the product effectively(ie, to advertise Electronic BenefitsTransfer andaffordability) contributedto potential misperceptions regardingprice and lack of participation.

Sustainability. All produce trucks inthe national assessment began with as-pirations to effect sustainable, structural

change in fooddistribution systemsandincrease food access among vulnerablecommunities; yet, the national compar-ison assessment revealed challenges tosustainability. Challenges reflected thefinancial realities of operating a retailfoodbusiness:Thehighcostsofproduceand operations, combinedwith low pri-ces, made breaking even challenging.All trucks required external/grants fund-ing to operate. To increase the potentialfor self-sustainability, managers negoti-ated serving those who most need theintervention and thosewho could affordto pay prices required for sustainabil-ity. Financial tensions also influencedoperating decisions (ie, locations, fre-quency of site visits). Lack of traininginfinanceandmanagementwas achal-lenge. Yet again, the researchers observedthe importance of community relation-ships:Thosewithstrongcommunityrela-tions developed pricing and outreachstrategies to navigate tensions successfully.

DISCUSSION

Using a parallel-convergent mixed-methods case study design, this reportdetails a process evaluation of the im-plementation of a mobile food accessintervention, POTG (Figure), to in-crease access to fresh produce in ruralfood deserts. Lessons learned from theevaluation revealed that perceivedben-efits may not outweigh the challengesof using the intervention.16 The inter-vention was not more convenientthan existing food retailers in locationaccessibility, price, or availability.Although distance is commonly usedto define food access,31 a recent USDAreport32 revealed that proximity is notnecessarily a decisive factor for low-income people, who may travel toshop for produce at preferred stores.In line with those findings, the major-ity of informants—residents of USDA-defined food deserts—expressed theywere able to purchase fresh producenear their neighborhoods. Likewise, arecent study on food access in Califor-nia's Central Valley found thatalthough rural communities in thislow-resource, Latino-majority areacould be characterized as food deserts,the agriculturally rich setting offeredseveral nontraditional outlets (ie, fruitand vegetable stands; dollar stores)that providedaccess to freshproduce.28

Among low-income groups, access topersonal vehicles is increasing,32whichmay explain why distance was notperceived as a major issue. With time,isolation from services,which is a high-ly salient barrier, may represent a largerfactor than lack of proximity to publictransportation, because isolation forcespotential consumers to make addi-tional plans to shop at POTG. Infor-mants reported that physical access tofresh produce was less of a concern forthem than was cost; price and afford-ability were more likely to influencewhere they shopped. Quality, variety,and availability of familiar foods wereperceived as better at other retailers;competition inthe formof small cornermarkets or other mobile vendors wasalso identified.

Nutrition education including fooddemonstrations were part of POTG so-cial marketing activities but were notimplemented. These activities couldelicit positive brand recognition17whileraising awareness and improvinghealthy eating attitudes among resi-dents of target communities, and mayrepresent an added value that couldmitigate personal barriers such as costand time. Food access interventionsthat include nutrition education haveshownpromising results amongvulner-able populations.12

Providing access to healthy foodsto residents of food deserts is mademore challenging in rural areas wherefull-service grocery stores or othertraditional food access interventionsmay not make financial or practicalsense. This mixed-methods evalua-tion assessing the implementation ofa mobile food access social marketingintervention demonstrated room forimprovement in implementing suchstrategies, and provided suggestionsfor sustainability and impact on ruralfood access disparities. Specifically,findings suggest the strategic signifi-cance of operational decisions relatingto site and product selection and dis-tribution models, and the criticalrole that effective implementation ofsocial marketing principles may havein successful interventions. Findingsunderscore the importance of com-munity involvement in decision mak-ing and nutrition education to ensurethe sustainability and efficacy of mo-bile interventions to improve accessto produce.

Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Ramirez et al 173

IMPLICATIONS FORRESEARCH ANDPRACTICE

Community involvement is critical todeveloping and successfully imple-menting strategies to improve healthyfood access. Consultationwith commu-nity members should start at the plan-ning phase and continue throughoutimplementation, involving residents indecision making including identifyingsites and schedules, product selectionand pricing, and distribution channels.For example, community members canprovide critical insight regarding thebest day of the week and time of theweek to visit. This may mean servingfewercommunities (if therearecommonlypopular and unpopular times), orvisiting more infrequently than onceper week, or itmaymean providing ac-cess to produce indirectly (eg, by servingas a supplier or distributor to small localstores rather than serving end con-sumers directly). In a related manner,a produce truck as a food access inter-ventionmayhave a stronger likelihoodof success if it is seen as an extension ofservices already provided by a trustedorganization rather than as an inde-pendent, standalone program.

Future mobile interventions shouldconsider alternate distribution models.The POTGsiteswere largely inaccessibleby public transportation, limiting po-tential clientele to those who had pri-vate transportation and could thereforereach other, nonmobile food vendors.Thus, an intervention distributionmodel must be flexible enough to beeffective in reducing barriers inherentin isolated communities, includingtransportation and price. Most POTGintervention sites made the producetruck the destination, because theywere not proximal to other retail andservices. This is a risky model; POTGand national comparisons revealedthat the sites in which produce truckswere integrated into an active commu-nity site weremore successful. In addi-tion, competition existed in the formof small corner markets or other mo-bile vendors; opportunities could beexplored to synergizewith these ratherthan compete. For example, the pro-duce truck may serve as a supplier offresh produce to local markets and aconsultant on storage and vendingproduce. This natural extension of

healthy corner store projects2-5 wouldwork well in rural settings.

ACKNOWLEDGMENTS

Theauthors thankAlexandraLemonforassistance with the national marketcomparison component of this projectand acknowledge the members of theAdvisory Board of the Produce on the Gointervention for assistanceand feedbackonthis researchproject.Themanuscriptwas supported by the National CancerInstitute of the National Institutes ofHealth under Award No. K01CA-190659. The content is solely the re-sponsibility of the authors and doesnot necessarily represent the officialviews of the National Institutes ofHealth.

REFERENCES

1. Walker RE, Keane CR, Burke JG. Dis-parities and access to healthy food in theUnited States: a review of food desertsliterature. Health Place. 2010;16:876-884.

2. Martin KS, Havens E, Boyle KE, et al.If you stock it, will they buy it? Healthyfood availability and customer purchas-ing behaviour within corner stores inHartford, CT, USA. Public HealthNutr. 2012;15:1973-1978.

3. Cavanaugh E, Green S, Mallya G,Tierney A, Brensinger C, Glanz K.Changes in food and beverage environ-ments after an urban corner store inter-vention. Prev Med (Baltim). 2014;65:7-12.

4. Dannefer R, Williams DA,Baronberg S, Silver L. Healthy bodegas:increasing and promoting healthy foodsat corner stores in NewYork City.Am JPublic Health. 2012;102:e27-e31.

5. Song H-J, Gittelsohn J, Kim M,Suratkar S, Sharma S, Anliker J. Acorner store intervention in a low-income urban community is associatedwith increased availability and sales ofsome healthy foods. Public HealthNutr. 2009;12:2060-2067.

6. Gillespie G, Hilchey DL, Hinrichs CC,Feenstra G. Farmers’ markets as key-stones in rebuilding local and regionalfood systems. In: Hinrichs CC, LysonTA, eds. Remaking the North AmericanFood System: Strategies for Sustainability.Lincoln, NE; University of Nebraska;2007:65–83.

7. Dimitri C, Oberholtzer L, Nischan M.Reducing the geographic and financialbarriers to food access: perceived bene-fits of farmers’ markets and monetaryincentives. J Hunger Environ Nutr.2013;8:429-444.

8. Schmidt M, Kolodinsky J, DeSisto T,Conte F. Increasing farm income andlocal food access: a case study of acollaborative aggregation, marketing,and distribution strategy that linksfarmers to markets. J Agric Food SystCommunity Dev. 2011;1:157-175.

9. Anderson TJ, Saman DM, Lipsky MS,Lutfiyya MN. A cross-sectional studyon health differences between ruraland non-rural U.S. counties using theCounty Health Rankings. BMC HealthServ Res. 2015;15:441.

10. PloegMVer,BrenemanV,DutkoP, et al.Access to affordable and nutritious food:updated estimates of distance to supermar-kets using 2010 data. USDA Econ ResServ—ERR143. https://www.ers.usda.gov/webdocs/publications/err143/33844_err143_summary.pdf. Accessed November 27,2016.

11. Valdez Z, DeanWR, Sharkey JR. Mo-bile and home-based vendors’ contribu-tions to the retail food environment inrural South Texas Mexican-origin set-tlements. Appetite. 2012;59:212-217.

12. Kennedy BM, Champagne CM,Ryan DH, et al. The ‘‘Rolling Store’’;an economical and environmentalapproach to the prevention of weightgain in African American women.Ethn Dis. 2009;19:7-12.

13. US Census Bureau. State & CountyQuickFacts: Merced County, Califor-nia. http://www.census.gov/quickfacts/table/AGE765210/06047. AccessedDecember 2, 2016.

14. USDA Economic Research Service.The Food Access Research Atlas:Merced County, California. https://www.ers.usda.gov/data-products/food-access-research-atlas/go-to-the-atlas. Ac-cessed December 2, 2016.

15. Dearing JW, Maibach EW, Buller DB.A convergent diffusion and social mar-keting approach for disseminatingproven approaches to physical activitypromotion. Am J Prev Med. 2006;31(4suppl):S11-S23.

16. Kotler P, Zaltman G. Social marketing:an approach to planned social changeon JSTOR. J Mark. 1971;35:3-12.

17. Grier S, Bryant CA. Social marketingin public health. Annu Rev PublicHealth. 2005;26:319-339.

174 Ramirez et al Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017

18. Yin RK. Case Study Research: Design andMethods. 5th ed. Thousand Oaks, CA:Sage; 2014.

19. Parks RB, Helwig D, Dettmann J, et al.Developing a performance nutritioncurriculum for collegiate athletics. JNutr Educ Behav. 2016;48:419-424.e1.

20. Curry L, Nunez-Smith M. MixedMethods in Health Sciences Research: APractical Primer. Thousand Oaks, CA:Sage; 2015.

21. Creswell JW, Plano Clark VL. Designingand Conducting Mixed Methods Research.Thousand Oaks, CA: Sage; 2011.

22. Mayring P. Qualitative content anal-ysis. Forum Qual Soc Res. 2000;1:20.

23. Krueger RA, Casey MA. Focus Groups:A Practical Guide for Applied Research.Thousand Oaks, CA: Sage; 2009.

24. Patton MQ. How to Use QualitativeMethods in Evaluation. 2nd ed. NewburyPark: Sage; 1987.

25. Dubowitz T, Ncube C, Leuschner K,Tharp-Gilliam S. A natural experimentopportunity in two low-income urban

food desert communities: research design,community engagement methods, andbaseline results. Health Educ Behav. 2015;42(1 suppl):87S-96S.

26. Sweeney G, Rogers C, Hoy C, et al.Alternative agrifood projects in com-munities of color: a civic engagementperspective. J Agric Food Syst Commu-nity Dev. 2015;5:69-75.

27. Dean WR, Sharkey JR. Rural and ur-ban differences in the associations be-tween characteristics of the communityfood environment and fruit and vege-table intake. J Nutr Educ Behav. 2011;43:426-433.

28. Valdez Z, Ramirez AS, Estrada E,Nathan S, Grassi K. Community per-spectives on access to and availabilityof healthy food in rural, low-resource,Latino communities [published onlineahead of print December 15, 2016].Prev Chronic Dis. http://dx.doi.org/10.5888/pcd13.160250.

29. Lopez MH, Gonzalez-Barrera A,Patten E. Closing the Digital Divide:

Latinos and Technology Adoption. Wash-ington, DC; 2013. http://www.pewhispanic.org/2013/03/07/closing-the-digital-divide-latinos-and-technology-adoption/.Accessed November 27, 2016.

30. Fishbein M, Cappella JN. The role oftheory in developing effective healthcommunications. J Commun. 2006;56(suppl):1-17.

31. Economic Research Service, USDepartment of Agriculture. Food Envi-ronment Atlas. http://www.ers.usda.gov/data-products/food-environment-atlas/go-to-the-atlas.aspx. AccessedNovember 27, 2016.

32. Ploeg M Ver, Mancino L, Todd JE,Clay DM, Scharadin B. Where DoAmericans Usually Shop for Food andHow Do They Travel to Get There?Initial Findings From the NationalHousehold Food Acquisition and Pur-chase Survey. http://www.ers.usda.gov/publications/eib-economic-information-bulletin/eib138.aspx. Accessed May 18,2016.

Journal of Nutrition Education and Behavior � Volume 49, Number 2, 2017 Ramirez et al 174.e1

CONFLICT OF INTEREST

The authors have not stated any con-flicts of interest.