For over a decade, our research team has focused on healthy food access among low-income populations with a focus on farmers markets. Financial costs and physical access barriers limit the consumption of fruit and vegetables, especially among households with low income. While nutrition incentive programs aim to curb financial burdens and encourage farmers market use, awareness of local farmers markets and existing nutrition incentive programs is low among SNAP consumers [20,37-39]. The lack of repeat SNAP consumers at farmers markets is a national issue. Of the SNAP households shopping at a farmers market in 2021, roughly half made only one purchase throughout the year [40]. This underuse may be amplified by additional barriers to farmers market, including location, hours of operation, and transportation [41]. Face-to-face interventions aimed at improving engagement with farmers markets among low-income populations were not widely adopted because of the human capital investment required to support implementation [42], so technology may be an avenue to increase farmers market use.

Building on this evidence, researchers from our team along with community nutrition partners tapped into lessons we have learned from our ongoing use of technology to improve community nutrition [22,43,44]. Our goal was to explore technology-based solutions to bridge the awareness gap among SNAP populations related to the availability of nutrition incentive programming.

Our previous work found that the majority of people hear about food procurement opportunities (like farmers markets) and nutrition incentive programs through word-of-mouth [20,45]. We hypothesized that technology could provide a unique opportunity to spur word-of-mouth dissemination [46,47]. Additionally, in support of a technology-based intervention, several studies have found significant improvement in diets and other health behaviors through app use [48-51]. Thus, the idea for a mobile app was identified as a solution to the underuse of farmers markets and nutrition incentive programs among low-income populations.

Support for a digital technology intervention like the app was strengthened by our research team’s previous successful development and deployment of technology-based solutions to improve community nutrition, including the Produce Path Manager Portal. Produce Path Manager Portal is an app that allows farmers market managers to track transactions, nutrition incentive use, vendors, and general information about their farmers market [52]. In use at over 500 farmers markets across 7 states, Produce Path Manager Portal is designed for market managers and state-level organizations to track and report on nutrition incentive programming. Farmers market managers create individual accounts for their markets, but market information is not accessible publicly for market customers.

The creation of the new app enhances the market manager features (of Produce Path) as a manager-facing portal and merges those features with a customer-facing portal for market customers. Given their different perspectives and expertise, both managers and consumers were included in our design and development process. We contracted a software development team to help design the prototype for the app and build version 1 (V1). Through a bidding and request-for-proposal process, we invited 2 companies for interviews with our research team and community nutrition partners. We selected one company based on their described skills, previous app development, and ideas for the app.

The “ideation” phase included development and feedback iterations (3 stages) to inform prototype design and learn about interests in using the app as well as features and design preferences. Qualitative feedback was collected over 16 months (May 2021-October 2022) with key end users: SNAP consumers with children and farmers market managers. Data came through feedback sessions and structured web-based interviews conducted via Zoom. Informed consent was obtained from participants involved in structured interviews. A summary of the data collection process and goals is outlined in Table 1.

Similar to approaches adapted by Kruzan et al [53] and Chandler et al [54] we worked with software developers to create 2 initial prototype interfaces. One was a shopping-centered design with a grocery list option and the second was a game-centered design with rewards and points for using the app. Both prototypes featured recipes with produce from farmers markets, a map of farmers markets, and a digital wallet for managing SNAP benefits, as illustrated in Figures 2 and 3. The shopping-centered design included market inventory and the ability to add produce to a grocery list. The gamification design distributed points to users for using the app and its features, and accumulated points could be traded in for in-app rewards, such as ribbons and badges.

We held web-based feedback sessions via Zoom with SNAP consumers with children who had previous farmers market experience (n=4) and farmers market managers (n=3) to prioritize the prototype designs and understand the desired features of the front-facing app. These participants were purposively recruited by word-of-mouth by our community nutrition partners. During the individual interviews, all participants were presented with simple, low-fidelity versions of the 2 app models and shown the content of the app and how they would function. Participants were asked about their overall impressions of both prototypes (ie, “Were there specific features of this model that you liked? Why?”) and what features would motivate them to keep using the app. Participants were also asked if they thought either of the prototypes could potentially improve their farmers market experiences. Interviews were conducted by researchers COH and KP, and they lasted between 30 and 60 minutes and were not recorded or transcribed.

Following the prototype selection in stage 1, we conducted structured qualitative interviews with 20 SNAP consumers with children younger than 18 years to understand their healthy food shopping routines and features of the app that might align with those routines. These interviews were conducted digitally via Zoom by researchers COH, JS, OY, RG, and KP, and they lasted between 45 and 60 minutes. We collected information about their likes and dislikes of mobile app designs in general, as well as apps that they specifically used in their cooking and shopping routines. These participants were recruited from prior research studies and across the state of Ohio through phone calls and flyers at farmers markets, grocery stores, and program offices (ie, WIC, Head Start). Eligible participants had used SNAP in the last 12 months, had at least 1 child younger than 18 years at home half of the time, did at least half of the household grocery shopping, and had experience using a smartphone app.

The interview guide and our engagement with the potential end users were rooted in the diffusion of innovation (DOI) theory. DOI is one of the first social science theories to understand behavior change [55]. This theory considers both diffusion, meaning the passive spread of an idea or product over time, and innovation, which includes novel ideas, products, interventions, etc [56]. Recently, DOI has been applied to health-related technology and mobile apps to understand their adoption and scalability [57-61]. The interview questions were shaped by 3 DOI attributes [56]: (1) relative advantage (ie, “Can you tell me about any apps or websites that you use for preparing and planning for shopping and cooking for your family?”), (2) compatibility (ie, “How would you rate this statement on a scale of 1, meaning not a good fit, and 10, a perfect fit: Using an app that tells me what foods are available at the farmers market every week could help me add more fresh and healthy foods to my shopping list?”), and (3) complexity (ie, What do you dislike about the apps that you use, or have used in the past?”). Attending to these constructs of innovations can aid in adoption and diffusion [55,62].

Following each interview, the interviewer and notetaker met to debrief and recorded key features of interest and any feedback relevant to the app development on a debrief form. Preliminary rapid analysis of debrief forms and discussions among the researchers who were involved in the interviews revealed features and design ideas mentioned by participants throughout the interviews. Informed by participants, research team members, community partners, and software developers, some suggested features were incorporated into the development of V1 in real time based on feasibility, available resources, and time. All interviews were recorded and transcribed, and formal qualitative analysis will be presented in future work.

Finally, individual feedback sessions were held with SNAP consumers (n=7) and farmers market managers (n=6) regarding V1 of the app. All SNAP consumers in these feedback sessions had previously participated in the stage 2 interviews. Two farmers market managers had previously participated in stage 1 interviews, and the remaining were 4 were purposively recruited by word-of-mouth by our community nutrition partners. SNAP users were selected based on the relevance of their responses in the stage 2 interviews related to app development and app use experience. During the debrief sessions after each interview in stage 2, the interviewer and note taker rated the relevance of participants’ responses to the research goals on a scale of 0 (low) to 10 (high). Participants who scored 8 or higher were invited to participate in stage 3 feedback sessions. In these feedback sessions, V1 was shared to test the functionality, design, and utility of the app.

During these individual interviews, researchers demonstrated the app; SNAP consumers were shown the consumer-facing portal and market managers were shown the manager-facing portal to market managers. After the demonstration, participants were given time to practice using the app on their own (ie, SNAP consumers added certain items to their grocery list and market managers created a new market event). All participants were asked about their likes and dislikes (ie, “How well will the app support farmers market use and experience for you and other families with children who shop with SNAP?”), pains and gains (ie, “What feature(s) was confusing and why? What will help make it less confusing?”), and ideas for additional features (ie, “From your perspective, what kind of information should be included in a marketing video about farmers markets and shopping with SNAP at farmers markets?”). SNAP consumers were asked about the functionality of the shopping list and market selection, whereas market managers were asked about the functionality of updating market inventory, events, and vendors. The goal of these pilot interviews was to inform future app updates.

Feedback sessions in stage 1 determined the prototype design. Core features of both prototypes included: market home pages with the address, hours, and website in addition to inventory, vendors, and events tabs; a map of nearby markets with the ability to “star” a home market; a recipe bank that could be filtered by ingredient; and, a web-based SNAP wallet that tracked both EBT and nutrition incentive use. The grocery shopping model uniquely included a customizable grocery list and the recipe ingredients could be directly added to the grocery list. On the other hand, the gamification version included an awards page to encourage app use, including challenges to visit multiple farmers markets and make recipes with leftovers, and in-app engagement to favorite recipes and review markets.

About 75% (n=3) of the SNAP consumers interviewed were highly in favor of a grocery model design compared with the gamification model (Figure 2 can be used for comparison). Their choice was largely informed by the utility of a grocery list, and not feeling motivated by in-app rewards or points in the game version. One participant for example shared that earning badges in the gamification version would not motivate them to use the app since they could not use those rewards in “real life.” Another shared that they felt farmers market users would like the grocery shopping model since it provides exactly what they need. Thus, we moved forward with the grocery model for the app development. Participants expressed general interest in the app, stating that app-available information could help them eat healthier, and they emphasized the potential for the app to save them time and money while grocery shopping.

All farmers market managers were similarly interested in the app and were shown the customer-facing portal. They particularly liked the feature that listed the available produce at their markets, along with the option to indicate if a product was limited in quantity at their market. They felt the app would help them engage with market customers, and that it would help customers prepare for visits to the market.

There was mixed feedback on the web-based SNAP wallet feature and the feasibility of linking EBT cards through the app: some SNAP consumers expressed distrust in digital finances, and some market managers were concerned about market vendors not having the technology to accept digital payments. Thus, we decided to not pursue this feature in V1 development. Additionally, participants expressed a desire for a clean, non-cluttered design. Participants were distracted by having too much information on the home screen.

Formal qualitative interviews with SNAP consumers in stage 2 provided feedback on the desired features and design of the app, as well as overall utility. Participants were mostly Black (n=10, 50%), women (n=19, 95%) with children between 0 and 5 years, and lived in urban counties (n=14, 70%) in Ohio (Table 2 contains demographic characteristics of participants). About 80% (n=16) of the participants used mobile apps daily in their routine. Many suggested features and functions from stage 2 interviews were implemented in real-time development based on participant consensus and feasibility from our software design partners, as illustrated in Figures 4-6. Cost and time prevented us from incorporating all suggested features, such as features to manage nutrition incentive balances.

Consumer suggestions included in V1:

After incorporating the previously mentioned suggestions from stage 2 interviews, we finalized the app V1 development and informally shared it with 7 SNAP consumers from the stage 2 interviews and 6 farmers market managers representing markets in rural, urban, and diverse racial populations. The SNAP consumers were all women, 43% (n=3) identified as Black and another 43% (n=3) identified as White, about 14% (n=1) were Hispanic, and 57% (n=4) were from urban counties in Ohio. SNAP consumers thought the app was clear and straightforward. They provided additional features and design suggestions to inform the app version 2 and future app updates, including:

Farmers market managers found the manager-facing portal of the app V1 simple, user-friendly, easy to navigate, and visually appealing. The manager-facing portal was created from Produce Path Manager Portal, which they were familiar with, and offered enhanced features. Market managers appreciated the messaging feature and discussed that it could replace some market newsletters and offer more streamlined communication between markets and customers. Farmers market managers expressed the following suggested updates: