Fruit and vegetable consumption helps prevent chronic diseases responsible for the major causes of illness and death in the United States (1–3). Despite the known benefits of eating fruits and vegetables, the typical US diet fails to meet recommendations (1). Substantial disparities in fruit and vegetable consumption exist by region, race, and income (4); these are mirrored in disparities in health outcomes and disease prevalence (5,6).

Limitations in community food sources and lack of transportation to reach food outlets may be responsible for some disparities in fruit and vegetable consumption (7–10). Residents of minority communities experience more problems in accessing healthful foods than do residents of predominantly white neighborhoods (11–14).

Suggested solutions to disparities in food access include direct-to-consumer marketing efforts, such as farmers markets, mobile produce trucks, community gardens, and community-supported agriculture (CSA) shares (15). CSA allows community residents to purchase local food directly from a farmer by buying a “share” of the farm’s anticipated production at the beginning of the growing season. Shareholders receive a box of produce each week. The required lump sum payment often restricts CSA programs to high-income consumers. Although some variations on CSA programs developed for low-income families have been described (16,17), their evaluation has been limited (18).

The primary objective of this pilot feasibility study was to test whether a program providing a summer’s CSA share and supportive programming around food purchase and food preparation would be associated with increased fruit and vegetable household inventory variety and increased fruit and vegetable consumption in low-income, minority families. Data are being used to determine programming revisions necessary to design a larger trial.

Twenty-five women were recruited for each group (Table 1). At follow-up, we were able to contact and interview 23 control and 21 intervention group members. Participants ranged in age from 24 to 60; most were African American and unmarried. Educational attainment ranged from high school graduate (or general educational diploma) to graduate degree. No significant differences were observed between groups at baseline.

Intervention group participants picked up the produce box an average of 9.2 (standard deviation [SD] = 4.58; range, 1–16) of 16 weeks. They attended an average of 1.2 (SD = 1.32; range, 0–4) of the 5 events. We observed a significant increase in the number of foods in the total inventory of fruits and vegetables and in vegetables alone in the intervention group, compared with the control group (Table 2). We observed no significant increases in fruit and vegetable consumption in the intervention group, compared with the control group (Table 3), although results trended in favor of the intervention group having higher intakes.

All but 4 of the 21 intervention participants interviewed reported challenges to picking up the produce box. Work schedules (n = 8), transportation (n = 4), and forgetting (n = 4) were the most common challenges. Distance, health issues, and out-of-town travel were also challenges. Similar challenges were listed for attending evening skill-building sessions; night classes and family activities often conflicted with evening attendance. The community agency had problems contacting participants with reminders because participants “graduated” from agency classes, and contact information changed frequently.

All intervention participants reported that they would be willing to pay $10 per week for a produce box in the future. Many were willing to pay more: 15 (71%) were willing to pay $15, 8 (38%) were willing to pay $20, and 3 (14%) were willing to pay $25. Of the 9 who reported receiving WIC (Special Supplemental Nutrition Program for Women, Infants, and Children) or SNAP (Supplemental Nutrition Assistance Program) benefits, 7 said they would be willing to use these benefits for the produce.

In an overall evaluation of the Farm Fresh Healthy Living program, 20 of 21 participants stated that if the program were conducted again, they would participate. In an open-ended question asking “What did you think of the program?,” all 21 responded positively. Responses included mention of the variety provided weekly, the chance to eat foods that were too expensive to purchase at the grocery store, the chance to expose children to new foods, and the better flavor of local produce compared with grocery store produce.

Debriefing results identified problems and possible solutions (Figure). Clarifying agency expectations and siting the program in agencies that are natural hubs of family activity (eg, community centers, day- or afterschool care centers) could improve participation and maintaining participant contact. Adapting produce box contents and size to participant preferences could make the program more effective and sustainable. More sensitive and specific evaluation methods could provide better understanding of the program’s impact on the overall household food system.

This study was designed to evaluate the effects of supplying local food from a CSA program to low-income families on families’ reported household food inventory and fruit and vegetable consumption. Compared with a control group, the participants in the Farm Fresh Healthy Living program reported a greater variety of fruits and vegetables in their households at the end of the season than did the control group. There was no difference in fruit and vegetable intake, though a trend toward higher consumption in the intervention group was observed. The greater number of fruits and vegetables present in food inventories of intervention participants demonstrates that the intervention increased the diversity of foods available to families. A larger sample size or a more sensitive data collection instrument for dietary intake may have demonstrated a significant effect on fruit and vegetable consumption.

Participants picked up their food an average of 9 of 16 weeks. Although the pick-up schedule was more generous than is the case with other CSA programs and accommodations were made for participants’ schedules, these efforts may have been insufficient to adjust the program to the life situations of low-income women. Many women worked multiple jobs and lacked workplace flexibility. Some depended on multiple buses to travel to and from the pick-up point, and they struggled to transport the produce box on the bus. Such issues that prevented full participation are consistent with those observed by other researchers (17,24). Proposed program changes reflect the problems of promoting and maintaining participation, increasing sustainability, and targeting outcome evaluation. Although participants reported liking and using most foods, whether such foods supplement regularly eaten foods (as intended) or displace them should be investigated.

Nationally, the proportion of fruits and vegetables consumed fresh has increased over the last 2 decades from 40% to 46% (25). However, substantial disparities exist in fresh produce consumption; households with incomes of $100,000 or more average $712 spent per capita in 2009, compared with $254 per capita spent by those with incomes less than $15,000 (25). These disparities in consumption, as well as the growing literature on food access disparities, underlie attempts to improve the food environment for underresourced communities. These efforts have been diverse, including supporting urban agriculture, incentivizing new grocery stores, promoting healthy corner stores, and establishing new farmers markets. Expanding CSA programs to underresourced communities fits with such efforts. The limited literature on CSA programs as an alternative food source suggests that the traditional organization of CSA programs is designed primarily to support growers (26) and that most CSA members have high incomes (27). Improving food supplies to low-income communities requires factors — low prices, distributed rather than lump-sum payments, and liberal food pick-up opportunities — that run counter to a grower-first orientation. Although the participants in this study placed high value on the food provided and stated a willingness to pay per week for the food, the amount they were willing to pay was less than half what other CSA members were charged during the same summer.

Our study has limitations. This was a small feasibility study conducted with clients of a single agency, receiving food from 1 CSA program. Therefore, the results may not be representative of other populations or other CSA programs. The outcomes were measured by self-reports rather than by more detailed dietary intake measures or by household inventory observations. The fruit and vegetable consumption measure focused on the mother, so the effect of the CSA share on others in the family is unknown. Not all possible explanatory factors (eg, distances from home to pick-up) were measured.

Despite these limitations, the study has several strengths. The design incorporated a well-matched control group drawn from the same population. The outcome measures have been widely used, particularly that for fruit and vegetable consumption, which is derived from the BRFSS evaluation (23). The statistical analysis adjusted for potential confounders, including education and self-efficacy for fruit and vegetable consumption. Process measures were used to identify problems and possible solutions that will be implemented in a follow-up intervention.

Expanding access to healthful foods is an important step in reducing health disparities. This study shows that food from a CSA program has positive effects on recipient households. Altering some of the financial and operational aspects of traditional CSA programs will be necessary to improve the impact of CSA participation.