Fruits and vegetables are important for the prevention and management of chronic diseases( Reference Nicklett and Kadell 1 , Reference Boeing, Bechthold and Bub 2 ), yet US national data show that in 2013, fewer than 18 % of adults consumed the recommended amount of fruits and fewer than 14 % consumed the recommended amount of vegetables( Reference Moore and Thompson 3 ). Food environment research indicates that availability of fruits and vegetables in neighbourhood stores and living in close proximity to a supermarket are associated with increased fruit and vegetable consumption( Reference Gustafson, Hankins and Jilcott 4 , Reference Morland, Wing and Roux 5 ), while fast-food availability is associated with decreased consumption( Reference Fraser, Edwards and Cade 6 ). Research across the USA shows that low-income and minority neighbourhoods have disproportionately fewer supermarkets and healthful food outlets, while surrounded by a larger presence of fast-food restaurants and consequentially greater availability of energy-dense foods( Reference Larson, Story and Nelson 7 , Reference Hilmers, Hilmers and Dave 8 ). Deficiencies of healthful food environments among low-income and minority populations contribute to their susceptibility to poor health and health disparities( Reference Black, Moon and Baird 9 ).
Studies of the food environment–diet relationship most commonly use the availability of healthful food as an indicator of access; either through macro-level measures such as supermarket presence or micro-level measures such as in-store food audits( Reference Caspi, Sorensen and Subramanian 10 ). However, many nuances related to how people access healthful foods are often missed, such as perceptions of access, the temporal availability of food outlets (e.g. store hours)( Reference Chen and Kwan 11 ), and the pricing, placement and promotion of foods( Reference Kelly, Flood and Yeatman 12 ). The five dimensions of access, first proposed as a framework for measuring access to health care( Reference Penchansky and Thomas 13 ), has recently been proposed to guide more comprehensive measures of access to healthful foods( Reference Caspi, Sorensen and Subramanian 10 ). Previous studies have applied this framework to examine food access based on perceptions and interactions with the food environment( Reference Flint, Cummins and Matthews 14 , Reference Andress and Fitch 15 ), but to our knowledge, none have applied this framework to evaluate a food environment intervention. We apply the five dimensions – availability, accessibility, affordability, acceptability and accommodation – to better understand the complexity of access to fruits and vegetables in the context of a food environment intervention, a mobile produce market (MPM).
MPM, which are portable fruit and vegetable markets, contribute an important yet often overlooked source of food in food environments nationally and internationally. For example, mobile food carts in the colonias along the South Texas border with Mexico( Reference Sharkey, Dean and Johnson 16 , Reference Valdez, Dean and Sharkey 17 ), fruteros or fruit vendors in Los Angeles( Reference Rosales 18 ) and Green Carts in the Bronx( Reference Lucan, Maroko and Shanker 19 ) provide an alternative market for purchasing fruits and vegetables to supermarkets, which may be less accessible because of travel distance( Reference Black, Moon and Baird 9 ), and local convenience stores, which may have few options for fruits and vegetables( Reference Glanz, Sallis and Saelens 20 ). In developing countries, informal markets and mobile vendors serve as a primary source of fruits and vegetables among vulnerable populations in some urban and rural areas. One study found that 95 % of low-income households in Hanoi, Vietnam purchased fruits and vegetables several times per week from informal stationary markets and 32 % from mobile vendors, citing as advantages proximity to home, low prices and possibilities to buy on credit( Reference Moustier, Loc and Son 21 ). However, it should be noted that the percentage of purchases of fruits and vegetables varies greatly in different areas. For example, fruit as a percentage of total purchases from mobile vendors was 58·6 % in South Africa( Reference Steyn and Labadarios 22 ), 18·5 % in Haiti( Reference Webb and Hyatt 23 ) and 4·4 % in Burkina Faso( Reference Drabo, Savadogo and Tarnagda 24 ), and vendors sold a mix of food types with many vendors predominantly selling prepared foods( Reference Steyn, Mchiza and Hill 25 ).
In the USA and UK, MPM have received attention in recent years as a strategy to increase access to and consumption of fruits and vegetables, particularly among populations with greater health risk( Reference AbuSabha, Namjoshi and Klein 26 – Reference Zepeda, Reznickova and Lohr 31 ). A range of target populations have been described in MPM intervention studies, including elementary-school students( Reference Tester, Yen and Laraia 29 ), older adults( Reference AbuSabha, Namjoshi and Klein 26 , Reference Jennings, Cassidy and Winters 28 ), low-income families( Reference Jennings, Cassidy and Winters 28 ), predominantly Hispanic and Black neighbourhoods( Reference Evans, Jennings and Smiley 27 ), areas of low fruit and vegetable consumption( Reference Lucan, Maroko and Shanker 19 , Reference Jennings, Cassidy and Winters 28 , Reference Fuchs, Holloway and Bayer 32 , Reference Li, Cromley and Fox 33 ), areas of low produce availability( Reference Lucan, Maroko and Shanker 19 , Reference Fuchs, Holloway and Bayer 32 , Reference Li, Cromley and Fox 33 ) and areas with high rates of heart disease, stroke and diabetes( Reference Jennings, Cassidy and Winters 28 ). Older adults are an important target population for MPM because older adults are at increased vulnerability to the cumulative health impediments of chronic diseases( 34 ) and face additional barriers to obtaining fruits and vegetables, such as being on a fixed income and having physical limitations that may impede their ability to reach and shop at distant markets( Reference Keller and McKenzie 35 , Reference Munoz-Plaza, Morland and Pierre 36 ).
The purpose of the present study was to apply the five dimensions of access framework to examine MPM influences on fruit and vegetable access within an urban, low-income population and to examine differences in perceptions of access between older (≥60 years) and younger (18–59·9 years) adults. We compared older and younger adults because older adults often have increased vulnerabilities such as chronic disease management and less mobility that increase the challenges to access fruits and vegetables, and may have different shopping behaviours, perceptions and experiences related to MPM. To our knowledge, the present study is the first to apply a structured assessment, the five dimensions of access framework, to examine an intervention aiming to increase fruit and vegetable access.
Methods
Setting
Springfield is the third largest city in Massachusetts, USA, and in 2015 had an estimated population of 154 341( 37 ). The latest statistics on race/ethnicity from 2015 showed that the Springfield population has a much higher population of colour compared with Massachusetts overall: Black/African American (21·7 v. 6·9 %), Hispanic/Latino (38·0 v. 9·3 %), Asian (2·2 v. 5·6 %) and non-Hispanic White (35·4 v. 75·7 %)( 38 ). Older adults ≥60 years represented 15·9 % of Springfield’s population in 2015( 39 ).
The current study examined eight Springfield neighbourhoods that were part of an MPM intervention. The largest racial/ethnic group in all but one of the eight neighbourhoods was Hispanics (35·1–80·8 %, one neighbourhood at 14·0 %); other ethnic/racial groups in the eight neighbourhoods were as follows: non-Hispanic Whites (6·8–32·2 %, one neighbourhood at 62·0 %), Blacks (11·5–33·1 %) and Asians (0·1–4·8 %)( 38 ). Older adults aged ≥60 years ranged from 9·5 to 20·1 %, and younger adults aged 18–59·9 years ranged from 52·9 to 61·2 %, of the eight neighbourhoods’ populations( 39 ). Poverty rates showed that approximately half of the population lived under the poverty line (range 48·9–63·6 %) in four neighbourhoods and poverty rates ranged from 15·1 to 29·6 % in the other four neighbourhoods( 40 ).
Food access maps showed more than half of Springfield, MA census tracts as food deserts, defined as having a substantial share (i.e. ≥500 residents or 33 % of residents in census tracts) of low-income residents living >0·5 miles (>0·8 km) away from a supermarket( 41 ). Portions of all eight neighbourhoods in the MPM intervention were classified as food deserts, and five neighbourhoods were entirely food deserts( 41 ).
Go Fresh mobile produce market
The Go Fresh MPM in Springfield, MA aims to improve access to locally grown fruits and vegetables and is a not-for-profit multi-organizational collaboration that includes city departments, a public health institute, three farms, a community action organization and a non-profit community loan organization. Launched in 2011, Go Fresh sells locally grown fresh fruits and vegetables (average of twenty varieties per week), such as berries, apples, tomatoes, leafy greens and squash, from June to October in neighbourhoods with limited access to fruits and vegetables. The project has consistently provided 50 % discount for all Supplemental Nutrition Assistance Program (SNAP) participants to assist low-income shoppers to purchase produce. In 2014, Go Fresh operated three days per week and was scheduled for weekly stops, each lasting 1 to
$${\rm 1}\,{\raise0.5ex\hbox{$\scriptstyle 1$}\kern-0.1em/\kern-0.15em\lower0.25ex\hbox{$\scriptstyle 2$}}\,{\rm h}$$
, at twelve locations in eight of Springfield’s seventeen neighbourhoods. Go Fresh locations included subsidized housing complexes, senior centres and other sites of congregation. For purposes of the present study, data were collected at all twelve Go Fresh locations operated during the 2014 season.
Study participants and design
Cross-sectional survey data were collected from a convenience sample of Go Fresh shoppers, recruited during the last two weeks of the Go Fresh 2014 season, from 22 to 31 October. All shoppers, estimated to be 180, were asked to complete the questionnaire and 147 shoppers agreed to participate. Questionnaires missing data on age or sex were excluded (n 4) leaving a final sample of 143 participants.
Study questionnaire
A twenty-item questionnaire was self-administered and included information on demographics, shopping behaviours, and perceptions and experiences about Go Fresh. Demographics included age (12–17, 18–24, 25–29, 30–39, 40–49, 50–59, 60–69, >70 years), sex (male/female), race/ethnicity (Hispanic, White, Black, Asian, American Indian, Native Hawaiian or Pacific Islander, other – write in), neighbourhood residence, and living alone or with others. Shopping behaviours included amount of money spent at Go Fresh on the day of the survey ($US: 0–5, 5–10, 11–20, 21–30, 31–40, 41–50, 51–75, >75), shopping frequency at Go Fresh (weekly, twice/month, once/month, once or twice, first time), primary shopper for the household (yes/no), the number of people purchasing food for (1, 2–3, 4–5, >5), number of Go Fresh locations shopped this year (1, >1 location), duration of shopping at Go Fresh (first year, second year, started 2–3 years ago) and plans to shop at Go Fresh next year (yes, no, don’t know). There were two open-ended questions: ‘Are there any vegetables or fruits missing from the Go Fresh mobile market that you wish were available?’ and ‘Please share any other comments you have’. The questionnaire was offered in three languages (English, Spanish and Vietnamese) to reflect the most common races/ethnicities of shoppers.
Five dimensions of access
We adapted the five dimensions of access defined by Capsi and colleagues( Reference Caspi, Sorensen and Subramanian 10 ). ‘Availability’ refers to the supply of food, including amount, type and variety, and the presence of certain store types; ‘accessibility’ refers to the location of the food supply and the convenience of reaching it; ‘affordability’ refers to prices of food or perceptions of cost; ‘acceptability’ refers to the quality of food; and ‘accommodation’ refers to the extent that food sources adapt to consumer needs, with attention to store hours and types of payment( Reference Caspi, Sorensen and Subramanian 10 ). For the purposes of our study, the five dimensions were assessed by the following: availability by perceptions of produce variety; accessibility by perceptions of location and distance travelled to the MPM; affordability by perceptions of produce price; acceptability by perceptions of produce quality; and accommodation by perceptions of market hours, safety of market location, community-led activities at the market and use of electronic benefit transfer (EBT) cards at the market.
We used indicators of perceived access because previous research shows that perceived measures are more strongly related to dietary behaviours than objective measures( Reference Caspi, Kawachi and Subramanian 42 , Reference Caldwell, Kobayashi and DuBow 43 ) and are instrumental to explain nuances influencing individuals’ self-efficacy to intake of healthful foods( Reference Usher 44 ). In the current study, two main questions captured measures of perceived access: (i) ‘What do you like about the Go Fresh mobile market?’ and (ii) ‘What would you like to see changed about the Go Fresh mobile market for next year?’, with response options to capture perceptions of the following: produce variety (availability); market location (accessibility), produce cost/pricing (affordability); ‘produce is locally grown’ as a proxy for quality( Reference Martinez, Da Pra and Pollack 45 ) (acceptability); hours of operation, safety of location and activities led by community organizations, such as food demonstrations and visit by a nutritionist from the Special Supplemental Nutrition Program for Women, Infants, and Children (accommodation). Participants were asked to ‘circle all that apply’ and there was an ‘other’ open-ended write-in response option. Two questions captured more objective measures of access. For accessibility, ‘How far did you travel to get to the Go Fresh mobile market today?’ (<1, 1–5, 6–10, 11–25, 26–50 miles (<1·6, 1·6–8·0, 9·6–16·0, 17·7–40·2, 41·8–80·5 km)) captured an objective measure of distance travelled. For accommodation, ‘Have you ever used an EBT card to purchase your produce at the Go Fresh mobile market?’ (yes/no) captured an objective measure of payment type. In addition, write-in responses to ‘Please share any other comments you have’ were categorized into one of the five dimensions of access.
Statistical analysis
Frequencies and percentages were calculated to describe participant characteristics and shopping behaviours. Age was collapsed into two categories, older adults (≥60 years) and younger adults (18–59·9 years); younger age categories had relatively few participants per category (18–24 years, n 7; 25–29 years, n 11; 30–39 years, n 19; 40–49 years, n 18; 50–59 years, n 27) compared with older adults (n 61). A χ 2 analysis was used to identify significant differences in participant characteristics and shopping behaviours between older and younger participants.
Separate logistic regression models were used to examine four shopping behaviours: EBT use (yes/no), amount of money spent at Go Fresh (≤$US 10/>$US 10), shopping frequency at Go Fresh (<weekly/weekly) and distance travelled to Go Fresh (<1 mile/≥1 mile (<1·6 km/≥1·6 km)), with the following variables in the models: age (18–59·9 years/≥60 years), race/ethnicity (non-White/White), sex (male/female) and living situation (living alone/living with others); EBT use (yes/no) was added to models examining money spent, shopping frequency and distance travelled. Each logistic regression model included only participants with complete data for that model.
Frequencies and percentages were used to capture perceptions and experiences of the MPM. Each dimension of the five dimensions of access in influencing fruit and vegetable access was assessed through the absolute percentages of responses. For each dimension, a χ 2 analysis was used to identify significant differences between older and younger participants.
Statistical significance was set at P < 0·05. Statistical analyses were conducted using the statistical software package Stata release 14.
Results
Participant characteristics are provided in Table 1. Participants were more likely to be women (69·9 %) and were nearly evenly distributed between older (42·7 %) and younger (57·3 %) age categories. Older participants were Black (32·1 %), White (25·0 %), Hispanics (21·4 %) and Asian (19·6 %), while Hispanics represented a larger proportion of younger participants (46·2 %; P=0·02).
Table 1 Characteristics of participants responding to the Go Fresh mobile produce market survey, Springfield, MA, USA, October 2014: comparison of younger v. older adults
* χ 2 tests were used to assess significance between younger and older adults.
Participant shopping behaviours are provided in Table 2. Most participants were their household’s primary food purchaser (86·6 %) and reported purchasing food for two or three people (58·5 %). Nearly half (44·2 %) of the participants shopped at Go Fresh weekly and most (84·2 %) shopped only at one Go Fresh location. For the majority of participants, 2014 was the first year of shopping at Go Fresh (72·1 %), and nearly all participants (92·9 %) planned to shop at Go Fresh the following year. The median range of money spent was $US 5–10, with 36·6 % of participants spending less than $US 5 and 25·4 % spending more than $US 10 at Go Fresh on the day of the survey. The most common means that participants found out about Go Fresh was through a community organization (41·3 %), followed by friends (28·0 %), flyers/postcards (19·6 %) and passing by the market (14·7 %).
Table 2 Shopping behaviour of participants responding to the Go Fresh mobile produce market survey, Springfield, MA, USA, October 2014: comparison of younger adults v. older adults
EBT, electronic benefit transfer.
* χ 2 tests were used to assess significance between younger and older adults.
† Respondents were asked to identify all that apply, so total percentages do not add up to 100 %.
Compared with younger participants, older participants were more likely to live alone (P < 0·001; Table 1), purchase food for one person (P < 0·001), be long-term shoppers (for 2–3 years; P=0·002) and use EBT (P=0·01; Table 2). Older participants shopped more regularly at Go Fresh – weekly and once/twice monthly – compared with younger adults, one-third of whom shopped there only once or twice in the season (P=0·09; Table 2).
EBT use was associated with living alone (OR=2·5, P=0·03) and being older (OR=2·1, P=0·09). Weekly shopping at Go Fresh was associated with living alone (OR=2·5, P=0·03). Travelling <1 mile (<1·6 km) to Go Fresh was associated with living alone (OR=2·6, P=0·02) and being non-White (OR=0·9, P=0·09). Money spent was not associated with any independent variables in the regression model (Table 3).
Table 3 Logistic regression models of participant shopping behaviour at Go Fresh mobile produce market, Springfield, MA, USA, October 2014
EBT, electronic benefit transfer.
Five dimensions of access
Below the results are organized into overall findings of the five dimensions of access and within subheadings of each dimension: availability, accessibility, affordability, acceptability, and accommodation. Differences between older and younger participants for each dimension are noted under the subheadings.
Overall, location was the dimension that most participants liked about Go Fresh (72·7 %), followed by price (47·6 %), locally grown produce (46·2 %), hours (37·1 %), variety (32·2 %), safety (30·1 %) and activities, such as food demonstrations and visits by nutritionists (10·5 %). Regarding desired changes, wanting more variety was the highest response (39·5 %), followed by hours (21·1 %) and lower prices (17·7 %).
Availability
Younger participants indicated greater approval of produce variety (41·5 %) compared with older participants (19·7 %; P=0·01; Table 4). In response to the question, ‘Are there any vegetables and fruits missing from the Go Fresh mobile market that you wish were available?’, fifty-four unique food items were provided by participants. These items included twenty-seven vegetables, fifteen fruits, and two other items, bread and nuts. Hot peppers, bell or sweet peppers, cilantro (coriander), green beans, cucumbers, avocados, oranges, peaches and grapes were mentioned three to five times. Tomatoes, bananas and ‘fruit’ were mentioned at least six times. Fruit, unspecified and including bananas, peaches and grapes, was mentioned most often at sixteen times.
Table 4 Survey participants’ perceptions and experiences of Go Fresh mobile produce market, Springfield, MA, USA, October 2014: comparison of younger v. older adults
SNAP, Supplemental Nutrition Assistance Program; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children.
* Respondents were asked to identify all that apply, so total percentages do not add up to 100 %.
† Current activities included food demonstrations and visits by WIC nutritionists.
Accessibility
More than half of participants lived within one of the neighbourhoods served by MPM (56·6 %; Table 1) and most were Springfield residents (97·7 %). Almost all participants travelled ≤5 miles (≤8·0 km) to the market (94·8 %), with the majority travelling <1 mile (<1·6 km; 72·2 %; Table 2). There was a high response to liking the market location (72·7 %), particularly among older participants (82·0 %; P=0·03); 7·3 % of younger participants wanted changes in location (Table 4). Nine participants wrote in comments about enjoying the convenience of location (Table 5).
Table 5 Responses to open-ended survey question ‘Please share any additional comments’, categorized into the five dimensions of access, among participants responding to the question (n 45) in the Go Fresh mobile produce market survey, Springfield, MA, USA, October 2014
SNAP, Supplemental Nutrition Assistance Program.
* Each participant’s response may have multiple comments that fit into more than one category.
† Write-in response overlaps with discrete response categories from other questions.
Affordability
Nearly half (47·6 %) liked and 17·7 % of participants wanted changes to produce price (Table 4). Older compared with younger participants were more likely to want changes in pricing (24·2 v. 13·4 %; P=0·09; Table 4). Of those who used EBT, 70 % indicated satisfaction with price.
Acceptability
Nearly half of participants (46·2 %) indicated liking the produce quality of being locally grown, with there being no significant difference in perceptions between older and younger participants (P=0·16; Table 4). ‘Other’ responses to aspects liked about the market included ‘freshness’ (Table 4). Nine participants provided write-in comments about enjoying the produce freshness, and three more commented on the good produce quality in general (Table 5). No participants indicated dissatisfaction with the produce quality.
Accommodation
EBT was used at the MPM by about one-third of the participants (29·9 %). Older participants were more likely to use EBT (41·8 %) compared with younger (21·5 %; P=0·01; Table 2). Two write-in comments expressed gratitude that SNAP was accepted at the MPM (Table 5).
Satisfaction with MPM hours was indicated by nearly two-fifths of participants (37·1 %) while about one-fifth (21·1 %) wanted a change in hours, with no significant differences in perceptions of hours between older and younger participants. ‘More hours’ and ‘sell on Thursdays’ were ‘other’ responses to aspects wanted changed about the market (Table 4). Four write-in comments expressed a desire for more hours (Table 5).
Safety in market location was liked by almost a third (30·1 %) of participants and was much higher than desires for changes (2·7 %). There was no difference in perception of safety between older and younger participants (Table 4).
Participants showed an interest in the MPM offering taste sampling (44·2 %), cooking classes (23·8 %), nutritional counselling (30·6 %) and health-care information (27·9 %). Younger participants were more likely to want more activities at the MPM (P=0·04) and were more interested in taste sampling (P=0·05) and nutritional counselling (P=0·01) than older participants (Table 4).
Forty-five people wrote in responses to the request, ‘Please share any other comments you have’. These responses captured all five dimensions of access: availability (i.e. enjoying or wanting more variety); accessibility (i.e. enjoying the convenience of the location, or not finding the location convenient, wanting additional locations); affordability (i.e. prices affordable or prices not affordable); acceptability (i.e. enjoying produce freshness and quality); and accommodation (i.e. enjoying the market setup, appreciation of EBT payment option and incentives, and desiring more hours and a winter MPM). Ten participants provided comments about the enjoyment of market staff. A quarter of the comments provided expressions of gratitude and positive experiences (Table 5).
Discussion
Our findings suggest that MPM influence access to healthful food choices within deficient food environments and that older and younger adults appear to have differing perceptions on some dimensions of access. Ours is the first study to apply the five dimensions of access framework to capture a market intervention influence on access to fruits and vegetables. We applied the framework to the perceptions of consumers in an urban environment with residents living within food deserts. The MPM captured all dimensions of access.
Go Fresh adds a source of produce to neighbourhoods, influencing availability, defined as the supply of food and presence of stores( Reference Caspi, Sorensen and Subramanian 10 ). Our data showed mixed perceptions of the variety of produce offered, with a greater satisfaction in variety among younger compared with older adults. Overall study participants indicated a desire for more of the items that Go Fresh offered (such as cilantro and tomatoes), cultural varieties such as recao (a long-leafed variety of cilantro), and items that are not grown locally such as bananas, oranges and avocados. This might reflect cultural preferences given that Hispanic/Latinos comprised a significant proportion of our participants. Particular preferences for ethnic options among African Americans and Latinos( Reference Fish, Brown and Quandt 46 ) as well as limited availability of culturally specific fruits and vegetables in many neighbourhoods( Reference Grigsby-Toussaint, Zenk and Odoms-Young 47 ) have been highlighted as important considerations for improving the food environment to meet the needs of residents. The desire for MPM to include staple food items such as bread and milk for one-stop shopping was expressed in another MPM study( Reference Zepeda, Reznickova and Lohr 31 ). While only two participants in our study indicated a desire for non-produce items, our results suggest that sourcing more produce, particularly fruit, and culturally specific produce such as recao, may satisfy and draw additional customers. Taste sampling may also improve shopper satisfaction with the current variety of produce because it might offer shoppers the opportunity to taste unfamiliar fruits and vegetables( Reference Zepeda, Reznickova and Lohr 31 ).
Accessibility of MPM locations and close proximity to residences was highlighted as an important reason for shopping at Go Fresh. Our findings that most MPM shoppers were Springfield residents, lived within one of the neighbourhoods served by MPM, travelled less than 1 mile (1·6 km) to market and liked the market locations underscore the mobility of MPM as an important asset to this marketing system and support the intention that MPM improve local access to produce for people living in food deserts( Reference Zepeda, Reznickova and Lohr 31 , Reference Fuchs, Holloway and Bayer 32 , Reference Marina, Blum and Bettisworth 48 , Reference Widener, Metcalf and Bar-Yam 49 ), addressing food deficits and diet-related health issues related to living in food deserts( 50 ). Despite the intention of MPM to improve accessibility to fruits and vegetables by making the location more convenient and by serving areas lacking fresh produce options, studies on Green Carts, a special permitted MPM in New York City, have found some MPM vending locations straying from areas with limited access and clustered around areas with greater commercial activity, sometimes competing with other stores selling fruits and vegetables( Reference Lucan, Maroko and Shanker 19 , Reference Li, Cromley and Fox 33 ). These findings are contrary to the current study. Go Fresh maintained its route throughout the duration of the 2014 season. In subsequent years, Go Fresh locations were modified to better serve vulnerable populations such as increasing the number of affordable housing sites and locations that were frequented by older adults. One reason for the difference may be that Go Fresh is a non-profit that is subsidized by grants and business sponsorship, while the Green Carts are for-profit businesses and therefore may gravitate to areas of higher sales. Similar to our findings, the need for more MPM locations( Reference Zepeda, Reznickova and Lohr 31 ) and transportation options to MPM( Reference Haynes-Maslow, Auvergne and Mark 51 ) have been expressed in previous MPM research.
The importance of affordability as an impetus for shopping at Go Fresh, indicated by produce price being the second most favoured characteristic of Go Fresh, supports findings from studies in New York that surveyed Green Cart customers( Reference Fuchs, Holloway and Bayer 32 ) and older adult shoppers of the Veggie Mobile( Reference AbuSabha, Namjoshi and Klein 26 ) that found affordable prices to encourage shopping at MPM. Go Fresh aims to offer produce at prices that reflect current grocery store prices so there is no price differential, and to incentivize produce purchases through 50 % discounts for all EBT users. An MPM study in Syracuse, NY demonstrated affordability by showing that seven of nine items offered by the MPM were lower than grocery store prices( Reference Robinson, Weissman and Adair 52 ). Our data suggest that EBT use and accompanying incentives facilitate affordability among low-income participants. Although only a third of participants used EBT, close to three-quarters of those who used EBT at Go Fresh indicated satisfaction with produce price. Our study extends the research showing SNAP benefits as an important facilitator for purchasing fresh fruits and vegetables among people with low incomes( Reference Karakus, MacAllum and Milfort 53 ) into another, perhaps more accessible market, the MPM. The importance of leveraging food assistance programmes to increase MPM use was identified in an MPM study in which 40 % of their participants expressed lack of affordability as a barrier to MPM use, particularly at the end of the month when people ran out of money or SNAP benefits( Reference Zepeda, Reznickova and Lohr 31 ). Our data suggest that SNAP and accompanying discount for EBT users were important for facilitating affordability particularly among older adults and those who lived alone. Our data also reflect contradicting experiences of affordability, with some participants identifying the need to improve pricing. Improvements in perceived affordability might be achieved by providing more information on SNAP benefits, which was desired by over a quarter of participants, and by posting comparison prices from supermarkets.
Study participants’ appreciation of produce quality being fresh and locally grown demonstrates acceptability of the MPM as a source of fruits and vegetables and supports other research findings that low-income customers of farmers’ market and MPM have enjoyed the fresh quality of produce( Reference AbuSabha, Namjoshi and Klein 26 , Reference Jennings, Cassidy and Winters 28 , Reference Karakus, MacAllum and Milfort 53 ).
While we did not set out to measure the quality of the relationship between customers and market staff as an indicator of acceptability( Reference Penchansky and Thomas 13 , Reference Usher 44 ), the importance of social interactions and relationships emerged from our data. Ten participants wrote in comments about positive experiences with market staff, suggesting that social interactions positively influenced shopper experience. Social networks were also important for initiating MPM use, with the highest two responses of finding out about Go Fresh being through coordinator/organizations and friends. In a study using focus groups to interview shoppers and non-shoppers of four MPM in different parts of the country, the theme of trust emerged, and participants expressed the importance of building trusting relationships with market staff and the preference for staff being from their own community( Reference Zepeda, Reznickova and Lohr 31 ). MPM acceptability may be improved through intentionality around community building, which was an important objective of the Community Development Corporation Mobile Market in Syracuse, NY that has become a vital social space particularly for older adults( Reference Robinson, Weissman and Adair 52 ). The importance of social relationships between consumers and food providers has been discussed in literature on local food systems through the conceptualization of social embeddedness( Reference Sage 54 , Reference Hinrichs 55 ) and should be examined in greater depth as a factor influencing food access in food environment research. Considering that perceptions of staff were the most often unsolicited comment from our participants, future research could benefit from examining the quality of social relationships as a factor influencing food access and an indicator of acceptability.
Perceived safety of MPM locations was one indicator of accommodation. Crime and other issues with safety in neighbourhoods can discourage people from walking to nearby food outlets( Reference Bader, Purciel and Yousefzadeh 56 , Reference Odoms-Young, Zenk and Mason 57 ). A study that examined food access disparities in New York City found that when crime and poor traffic safety were considered in food environments, disparities in access to a supermarket widened and negatively impacted food access in the poorest neighbourhoods and neighbourhoods with higher proportions of Black, Hispanic and Asian residents( Reference Bader, Purciel and Yousefzadeh 56 ). By travelling directly to residential sites such as apartment buildings and to community centres where consumers may be comfortable, MPM can mitigate safety issues that may otherwise be of concern. While neighbourhood safety has not been explicitly defined as an indicator of accommodation in previous five dimensions of access frameworks( Reference Caspi, Sorensen and Subramanian 10 , Reference Penchansky and Thomas 13 , Reference Usher 44 ), future applications of the framework should consider its inclusion.
Perceptions of market hours, another indication of accommodation, was mixed in our study. While over a third of our study participants liked the market hours, almost a quarter wanted changes in hours. Unlike supermarkets that are open most days and evenings, MPM usually are open for short periods of time (1–2 h) at each location that may not accommodate individual schedules(
Reference Zepeda, Reznickova and Lohr
31
,
Reference Gorham, Dulin-Keita and Risica
58
). Hours for each of Go Fresh’s market stops in 2014 were on weekdays, lasting from 1 to
$${\rm 1}\,{\raise0.5ex\hbox{$\scriptstyle 1$}\kern-0.1em/\kern-0.15em\lower0.25ex\hbox{$\scriptstyle 2$}}\,{\rm h}$$
, between 10 am and 5 pm. Interviews from a study of Fresh to You, an MPM in Rhode Island, found that working adults had a difficult time attending market due to work conflicts(
Reference Gorham, Dulin-Keita and Risica
58
). In subsequent years, Go Fresh increased the hours of operation by adding weekend hours in an effort to accommodate more people.
Having a higher EBT use rate among older participants may indicate Go Fresh’s capacity to accommodate low-income older adults despite potential stigma with SNAP participation that is often magnified in older adults( Reference Haynes-Maslow, Auvergne and Mark 51 , Reference Gabor, Williams and Bellamy 59 ). National rates of SNAP participation in 2014 show SNAP enrolment by only 42 % of eligible older adults (≥60 years) compared with 85 % of eligible adults aged <60 years( Reference Gray and Cunnyngham 60 ). A study examining EBT use at Green Carts in the Bronx, NY found that EBT users spent an average of $US 3·81 more per transaction compared with those who paid by cash, suggesting that EBT use may support increased fruit and vegetable purchases and subsequently increased fruit and vegetable consumption( Reference Breck, Kiszko and Abrams 61 ).
Our data suggest that Go Fresh accommodated the needs of those living alone, which is promising in the light of findings of greater nutritional risk among those living alone( Reference Nicklett and Kadell 1 , Reference Deierlein, Morland and Scanlin 62 ). Living alone was associated with EBT use, shopping weekly and travelling <1 mile (<1·6 km). Those living alone may benefit from the opportunity for social interaction offered by MPM( Reference Jennings, Cassidy and Winters 28 ), which could partly explain their weekly commitment to shopping at Go Fresh. Research shows that gathering places like community centres and congregate meals are important for breaking social isolation and lowering the risk of malnutrition, particularly among older adults living alone( Reference Walker and Beuchene 63 ).
Older adults, who are more likely to face challenges to food access due to transportation, mobility and health limitations( Reference Wolfe, Frongillo and Valois 64 ), may particularly benefit from MPM, which can bring healthful foods closer to home. Lack of transportation is a commonly perceived barrier to fruit and vegetable access among older adults( Reference Wolfe, Frongillo and Valois 64 , Reference Haynes-Maslow, Parsons and Wheeler 65 ) and older adults in urban areas rely on walking/walkable neighbourhoods and public transportation to acquire groceries( Reference Munoz-Plaza, Morland and Pierre 36 , Reference Chung, Gallo and Giunta 66 ). In our study, older participants had a higher response to liking the market location, suggesting accessibility as an important MPM facilitator of produce access particularly for older adults. Similarly, in an intervention study examining the introduction of the Veggie Mobile to two senior housing sites in Troy, NY, convenience of MPM location was identified as the most-liked aspect of the MPM by 77·8 % of older adult shoppers( Reference AbuSabha, Namjoshi and Klein 26 ). Our findings that older participants had a longer shopping history at Go Fresh and shopped at the MPM more frequently compared with younger participants may indicate greater accommodation among older adults, who may need to shop more frequently in small batches to avoid spoilage and the burden of carrying large groceries( Reference Munoz-Plaza, Morland and Pierre 36 ). A qualitative study that used interviews and observations to examine shopping behaviours of older adults in Brooklyn, NY found that older adults typically shopped at least once per week, bought a few items at a time, and often shopped at multiple stores to acquire food that met their desire for quality, fit their budgets, and where they trusted the business practices( Reference Munoz-Plaza, Morland and Pierre 36 ). Although our study did not find significant differences in perceptions and experiences of quality or price between older and younger participants, older adults’ appreciation and preference for fresh, non-processed, healthful, high-quality food at a low cost has been consistent in the literature on older adult shopping behaviours( Reference Munoz-Plaza, Morland and Pierre 36 , Reference Pettigrew, Worrall and Biagioni 67 , Reference Feldmann and Hamm 68 ). Being on a fixed budget with added medical expenses underscores the importance of low-cost food for many older adults( Reference Nicklett and Kadell 1 , Reference Wolfe, Olson and Kendall 69 ), but may be less pronounced (i.e. different from younger adults) when considered in the context of a low-income population in which the need for affordability is commonplace. Our finding that fewer older adults responded to liking the produce variety compared with younger adults may be related to older adults’ preferences for starchy vegetables( Reference Nicklett and Kadell 1 ), habitually consumed( Reference Pettigrew, Worrall and Biagioni 67 , Reference Pettigrew and Moschis 70 ) and culturally specific produce varieties( Reference Munoz-Plaza, Morland and Pierre 36 ), and staple food items( Reference Zepeda, Reznickova and Lohr 31 ) not offered at the MPM.
Overall, our study found that MPM captured the five dimensions of access to fruits and vegetables. Previous research has shown that fruit and vegetable purchases( Reference Evans, Jennings and Smiley 27 ) and consumption increased( Reference AbuSabha, Namjoshi and Klein 26 – Reference Jennings, Cassidy and Winters 28 ) in MPM intervention sites and that shoppers consumed more fruits and vegetables than non-shoppers( Reference Zepeda, Reznickova and Lohr 31 ). Our study sets the foundation for examining how MPM fit into the broader food environment nationally and internationally. Our findings may be used to better understand drivers of access to previous research that has shown MPM being used as a primary( Reference Jennings, Cassidy and Winters 28 ) and supplementary( Reference Marina, Blum and Bettisworth 48 ) source of fruits and vegetables. In developing countries and US urban centres where street food and other mobile vendors sell a variety of foods including prepared foods, mobile market users may face different challenges to healthful food access, such as food safety or competition with unhealthful foods( Reference Steyn, Mchiza and Hill 25 , Reference Lucan, Maroko and Bumol 71 ). In these settings, the five dimensions of access framework may be applied to better understand the nuances of these markets’ influence on healthful food access.
Strengths and limitations
The current study had several strengths. We had a high response rate: 147 of 180 shoppers (81·6 %) participated in the study. Translating the questionnaire into Spanish and Vietnamese enabled the inclusion of customers with different language preferences. The questionnaire had closed-ended and open-ended questions; many closed-ended questions had an ‘other’ category that allowed participants to provide additional information. For two dimensions, both objective and subjective indicators of access were included to enhance analysis of the five dimensions of access, such as the objective measure of distance travelled to MPM as well as perceptions of the MPM locations.
There are several potential limitations to the current study. Our findings capture the perceptions and experiences of shoppers at one urban MPM and therefore may not reflect rural areas or other MPM that may have operational differences. The use of a convenience sample did not include non-shoppers or former shoppers who may have had poor experiences with the MPM. This may have increased positive perceptions of the MPM and may have missed important aspects of access. However, our findings included both satisfaction and dissatisfaction for four dimensions of access, found differences in perceptions between older and younger adults, and demonstrated that each dimension was able to be captured in a food environment intervention. This convenience sample allowed us to identify aspects of dimensions that will help refine the application of the five dimensions of access to the food environment. For example, inclusion of interactions between shoppers and market staff may facilitate understanding of the acceptability dimension and inclusion of neighbourhood safety may do the same for accommodation. Another potential limitation is the use of ‘locally grown’ as a proxy indicator of quality. Although superior freshness and taste are commonly perceived quality indicators attributed to locally grown food( Reference Feldmann and Hamm 68 ), our study participants may have interpreted ‘locally grown’ in different ways. In a review of consumers’ perceptions and preferences for local food, perceived attributes of local food included supporting the local economy, environmentally friendly growing practices and greater food safety( Reference Feldmann and Hamm 68 ). Among a sample of older Australians (>60 years) in a qualitative study, greater healthfulness and food safety were attributed to locally produced food( Reference Pettigrew, Worrall and Biagioni 67 ). Future studies may want to explore other measures of food quality and attributes of ‘locally grown’, such as direct questions on the quality of produce, or food safety, that may influence consumers’ perceived acceptability of produce, and to be more specific with the language and indicators used for quality. In addition, a checklist of characteristics that demonstrate produce quality could be added as an objective measure of acceptability.
While findings from the current study demonstrated variability between dimensions of access, our research design did not allow us to set a criterion or threshold to demonstrate changes in or facilitation of access. Food environment interventions may benefit from research that uses mathematical models applied to larger populations to set thresholds to indicate facilitation of access for each dimension.
Conclusion
We used a novel approach to assess an MPM intervention through five dimensions of access, demonstrating the framework’s applicability for researchers and public health organizations to assess food access, and expanding understanding of how MPM facilitate access to fruits and vegetables. We look beyond conventional measures of access that tend to focus on healthy food availability within a food outlet or geographic boundary, instead broadening access to consider the combinations of more indicators, including: perceptions of produce variety, price and quality; perceptions of market location, hours and safety; distance travelled to market and EBT use. By examining nuances of MPM influence on fruit and vegetable access, we examine the complexity of food access, involving social, cultural, economic and structural factors that influence individuals’ interactions with their food environment. Our study sets the foundation for evaluating an intervention through multiple dimensions of access. Future studies should continue to refine measures used to capture multiple dimensions, for example develop indicators to reflect an intervention/food environment’s accommodation of cultural food preferences. More studies capturing the five dimensions of access are needed to establish thresholds within each dimension for evaluating interventions.
The present study suggests that MPM may influence fruit and vegetable access in low-income urban neighbourhoods with limited fruit and vegetable access and may especially benefit older adults and individuals living alone. Our finding that living alone was associated with three shopping behaviours (i.e. EBT use, shopping at Go Fresh weekly, travelling <1 mile (<1·6 km) to Go Fresh) sets the foundation for future research about food acquisition behaviour related to living alone or household status, and the economic and social importance of food environment interventions. The role of social interactions in facilitating better nutrition and health outcomes is lacking in current MPM literature and would provide valuable insight on the impact of MPM.
Acknowledgements
Acknowledgements: The authors thank the study participants for their valuable contribution to this study. Financial support: This work was supported by the US Department of Agriculture, National Institute of Food and Agriculture, Multistate NC1196 and Hatch MAS00465 (grant number #1002626). The opinions expressed in this article are the authors’ own and do not reflect the view of funders. Conflict of interest: All authors declare no conflicts of interest. Authorship: B.H. conceptualized the study, led data analysis and interpretation, and wrote the manuscript. L.S. assisted in conceptualizing the study. K.W. contributed to the development of the survey instrument and led the data collection. L.M.T. oversaw the study, assisted in conceptualizing the study, contributed to data interpretation, and assisted in writing the manuscript. All authors contributed to revising the manuscript and approved the final version. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by The Institutional Review Board at University of Massachusetts Amherst. Verbal informed consent was obtained from all subjects. Verbal consent was witnessed and formally recorded.
