Article Text

Download PDFPDF

Protocol
Evaluating the impact of a full-service mobile food market on food security, diet quality and food purchases: a cluster randomised trial protocol and design paper
  1. Melissa L Horning1,
  2. Kristen S Gorman1,
  3. Stephanie Wagner2,
  4. Jayne A Fulkerson1,
  5. Julian Wolfson3,
  6. Melissa N Laska4,
  7. Lisa Harnack4
  1. 1 School of Nursing, University of Minnesota, Minneapolis, Minnesota, USA
  2. 2 The Food Group, New Hope, Minnesota, USA
  3. 3 Division of Biostatistics and Health Data Science, School of Public Health, University of Minnesota, Minneapolis, Minnesota, USA
  4. 4 Division of Epidemiology and Community Health, School of Public Health, University of Minnesota, Minneapolis, Minnesota, USA
  1. Correspondence to Dr Melissa L Horning; horn0199{at}umn.edu

Abstract

Introduction Mobile food markets may help to mitigate diet-related and weight-related inequities by bringing low-cost, nutritious food directly to underserved populations. By stocking foods to meet a range of dietary needs, full-service mobile markets may improve multiple aspects of diet, food security and fruit and vegetable procurement with a convenient one-stop shop.

Methods and analysis This cluster randomised trial is evaluating the impact of a full-service mobile market, the Twin Cities Mobile Market (TCMM). The TCMM sells staple foods at affordable prices from a retrofitted bus that regularly visits communities experiencing low incomes. The trial’s primary outcome is participant diet quality. Secondary outcomes include intake of specific foods and nutrients, food security and servings of fruits and vegetables procured for the home.

Together with our partners, we enrolled four subsidised community housing sites in three waves (12 total sites), aimed to recruit 22 participants per site (N=264) and collected baseline data. Sites were then randomised to either receive the full-service TCMM intervention or serve as a waitlist control, and the full-service TCMM began implementing at intervention sites. Follow-up data collection is occurring at 6 months post-implementation. After follow-up data collection for each wave, the full-service TCMM intervention is being implemented at the waitlist control sites. Waves 1 and 2 are complete and Wave 3 is in progress.

At baseline and follow-up data collection, dietary quality and intake are being assessed through three, interviewer-administered, 24-hour dietary recalls, food insecurity is being assessed by the 18-item Food Security Screening Module and fruit and vegetable procurement is being measured by collecting one month of food procurement tracking forms.

We will use intent-to-treat analyses to determine if participant diet quality, food security and procurement of fruits and vegetables improve in the sites that received the full-service TCMM intervention relative to the participants in the waitlist control condition.

Ethics and dissemination Trial procedures have been approved by the University of Minnesota Institutional Review Board. We plan to disseminate main outcomes in Grant Year 5 in both scientific and community spaces.

Trial registration number ClinicalTrials.gov: NCT05672186.

  • Clinical Trial
  • Community Participation
  • Food Insecurity
  • NUTRITION & DIETETICS
http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • Evaluates a full-service mobile market with a strong, cluster randomised trial design.

  • Relies on strong community-engaged research partnerships.

  • Uses high-quality measures to assess diet quality and food security.

  • Uses a novel tool to assess the impact of the mobile market on food procurement.

  • Is limited to one geographical region.

Background

Diet contributes to a multitude of chronic health conditions (eg, diabetes, heart disease) and leading causes of death in the USA, which disproportionately impact individuals from racial and ethnic minority backgrounds and individuals with low incomes.1 2 Lack of affordable, nutritious food access and food insecurity are important contributing factors to these disparities, and mobile food markets (‘grocery stores on wheels’) have been proposed as a mitigating strategy because they bring low-cost, nutritious food directly to underserved populations.

Several small, produce-only mobile market research studies have demonstrated increases in fruit and vegetable access, purchase and/or intake among customers.3–7 Two cluster randomised trials have found produce-only mobile markets increased fruit and vegetable intake by as much as a half8 to one serving per day.9 Such findings are promising. However, effects on overall diet quality, food insecurity or foods procured for the home (a precursor to dietary intake) are not yet known.

Aligned with recommendations from research,5 8 full-service mobile markets may increase the number of individuals who want to shop at a mobile market by providing one-stop shopping versus only produce. Full-service mobile markets sell food from buses (or other vehicles) that regularly visit low-income urban neighbourhoods. Preliminary research on full-service mobile markets suggests promising outcomes (eg, mobile market use was quantitatively linked to associations with less food security and qualitatively linked to improved dietary intake),10–13 and full-service mobile markets serve high needs populations, with one study finding that 85% customers’ surveyed experienced food insecurity.12 These data warrant rigorous testing of the full-service mobile market model in communities experiencing low food access and incomes.

Trial design

The objective of this manuscript is to describe the trial design and protocol for the ongoing, cluster randomised trial evaluating the real-world impact of the Twin Cities Mobile Market (TCMM) on diet quality, intake of specific foods and nutrients, food security and servings of fruits and vegetables procured for the home. The trial uses a parallel group design with study sites (clusters) randomised in a 1:1 ratio and uses a superiority framework where we anticipate the TCMM will contribute to better outcomes. The trial grant was funded on 25 February 2022 with an anticipated end date of 31 December 2026.

Aims and hypotheses

The trial’s primary aim is to evaluate the impact of the TCMM on the diet quality of study participants. Our hypothesis is that improvements in diet quality between baseline and follow-up will be higher among participants at sites randomised to the mobile market relative to waitlist control site participants.

Secondary trial aims include evaluating the impact of the full-service mobile market on intake of specific foods and nutrients (servings of fruits and vegetables; energy; per cent of calories from added sugar; and sodium), food security over the past 6 months and fruit and vegetable servings procured for the household. Our hypotheses for these secondary aims are that change in each outcome will be more favourable among participants at sites that received the mobile market after randomisation relative to waitlist control site participants (eg, greater increase in fruit and vegetable intake, greater decrease in sodium intake and higher proportion with an improved level of food security or stable high food security).

Public involvement

This trial is possible based on a strong community partnership between MLH (first author and principal investigator (PI)) and leadership at the TCMM that has developed over nearly a decade. The trial began through small partnered research studies in which the Director of the TCMM at the time and MLH were co-creating the research base that generated preliminary data for this trial.10–13 This research specifically included focus groups with mobile market customers at the time to ensure the research was designed with their input and feedback; for example, participants wanted to have a waitlist control. Although the founding Director of the TCMM left the mobile market as it transitioned to a new organisation, the research partnership remained strong between MLH and The Food Group, with SW (coauthor), a named co-investigator on the trial and integral partner to the trial’s successful completion.

Methods: participants, interventions and outcomes

Study setting

The trial is taking place in the Twin Cities, the largest metropolitan area in Minnesota. To facilitate feasibility, we are conducting the trial in three waves of four community sites. Specifically, we recruited 12 public and private subsidised housing community sites (‘clusters’) who had not previously received mobile market services to receive the TCMM as randomised (ie, either after baseline data collection or after follow-up data collection is complete). Then, based on power analysis calculations, we targeted recruitment of 22–25 participants from each site. Waves 1 and 2 are complete and Wave 3 is in progress (ie, sites and participants are enrolled but the intervention period and follow-up data collection are yet to be completed).

Eligibility criteria

Site eligibility: Site eligibility criteria included the following for each site: (1) public or private subsidised housing not located within 0.5 miles of a full-service grocer or other trial or mobile market sites to minimise risk for contamination; (2) property management and resident interest in receiving mobile market services; (3) willingness to host the mobile market each week with a designated parking spot that is free from debris and vehicles; (4) willingness to allow flyers and posters to be hung and informational meetings, data collection and community engagement activities to be held; and (5) willingness to allow mobile market staff to use onsite bathrooms.

Participant eligibility: Participant eligibility criteria included the following: (1) 18 years of age or older; (2) identifying as the primary grocery shopper in their household; (3) being able to speak English or American Sign Language; (4) living within a half mile of the community site location; (5) reporting being likely or somewhat likely to shop at the market after hearing a brief description of the service from study staff (response options: likely to unlikely); and (6) being willing and able to participate in all study data collection activities. Participant exclusion criteria were: (1) planning to move in the next 12 months; (2) currently shopping at the mobile market; (3) not having a phone number or mailing address; or (4) having a condition that would prohibit participation in the study or the quality of the data.

Intervention

The TCMM is a programme of The Food Group, a non-profit organisation working to increase access to nutritious and culturally connected foods. The TCMM is a full-service grocery store on two buses that were each retrofitted to be single aisle grocery stores with~170 different items. TCMM provides weekly service to community sites (research and non-research alike) and has several key design features that may facilitate positive outcomes: TCMM focuses attention on culturally connected food preferences based on input from community forums and customer requests.10 13 Additionally, TCMM uses food stocking practices that focus on major food groups (eg, fruits, vegetables, dairy, protein, grains) with limited inclusion of foods for which limited consumption is recommended (eg, salty snacks, chips, desserts, sugar-sweetened beverages) aligned with behavioural nudging and choice architecture techniques that promote nutritious food choices.14 15 To further facilitate nutritious stocking choices, prior to the start of the trial, the investigator team, inclusive of the TCMM investigator, developed nutritious stocking guideline recommendations.16 Low food pricing aims to improve affordability of foods carried on the mobile market. At the time of study funding, it was estimated that prices were on average 10% below prices at local supermarkets; in 2023, an analysis by our team discovered that TCMM pricing was comparable to Aldi (a full-service discount grocery chain) and was 13–38% lower than pricing at Walmart, Target and Cub Foods (a regional full-service grocery chain).17 TCMM’s payment options include cash, credit, debit and Electronic Benefit Transfer./ Nutrition Assistance Program (SNAP), and there are additional incentives for produce purchases through the state-funded and federally-funded Market Bucks programme, which provides coupons for future produce purchases matching current SNAP produce purchases up to US$20.18 19 TCMM aims to provide a safe, welcoming environment with exceptional customer service and opportunities for interacting with other customers/neighbours.10 To maximise participation, TCMM sets recurring weekly market stop times in conjunction with community sites and actively engages with community site partners to address needs and concerns in real time. Lastly, there is a Market Members’ loyalty programme that rewards customers with a nominal value at regular intervals (eg, US$5 every third shopping trip).

This research study does not alter TCMM practices for promoting the service, and all members of the public, participants and non-participants alike, are allowed to use the service at research sites. Participants maintain autonomy over where they procure their groceries, so they can choose to shop at TCMM and/or other food retailers. Likewise, no restrictions are placed on participants’ enrolment in other food access programmes (eg, SNAP, meal delivery, congregate meals, food shelf visits, food prescription programmes).

Outcomes

Primary outcome

Diet quality at follow-up adjusted for baseline: Data collected via three interviewer-administered, 24-hour dietary recalls collected at baseline and follow-up are being used to calculate participant’s Health Eating Index-2020 (HEI-2020) total scores to measure participant’s overall diet quality, the trial’s primary outcome. HEI-2020 scores can range from 0 to 100, and higher HEI scores represent greater consistency with the Dietary Guidelines for Americans. HEI scores will be calculated by time point, and participants will have had to complete at least two of the three dietary recalls per time point to be included in analyses.

Secondary outcomes

Dietary outcomes at follow-up adjusted for baseline: Additional secondary dietary outcomes from the 24-hour dietary recalls to be evaluated include average fruit and vegetables consumed (servings/day), energy intake (kcal/day), added sugar intake (per cent of calories from added sugar) and sodium intake (mg/day). Averages for these variables will be calculated by time point, and participants will have had to complete at least two recalls per time point to be included in analyses.

Food security at follow-up adjusted for baseline: Food security is being measured with the 18-item Household Food Security Screening Module of the United States Department of Agriculture (USDA). This measure was adapted to measure food security in the past 6 months (vs 12 months) to match the trial’s measurement time frame. Scores will be calculated per USDA guidelines for each time point to capture the level of household food security (ie, very low, low, marginal and high food security).20

Improved level of food security or stable high food security at follow-up adjusted for baseline: From the food security data, an additional variable will be created for each participant to capture either improved/stable food security (1) or decreased/stable very low, low or marginal food security (0) from baseline to follow-up. Specifically, improved/stable food security will be quantified as moving from very low, low or marginal food security at baseline to an improved level of food security at follow-up or having stable high food security at both baseline and follow-up. Decreased/stable very low, low or marginal food security will be quantified as decreasing one or more food security levels from baseline to follow-up or remaining stable in a very low, low or marginal food security level from baseline to follow-up.

Average servings of fruits and vegetables procured for the household per week at follow-up adjusted for baseline: The novel Fruit and Vegetable Procurement Tool was developed for this study to measure average weekly servings of fruits and vegetables procured over 4 weeks. The tool contains 4 weekly booklets with instructions and forms that participants used to record the amounts and types of all fruits and vegetables brought into their homes from in-person or online food retailers (inclusive of grocery stores, food pantries, farmers markets, dollar stores and more). As participants use the forms, they are provided reminders to mail their booklets in weekly, using pre-addressed, postage-paid envelopes. Participant fruit and vegetable procurement data is being entered per-protocol to calculate the total edible (one cup equivalent) servings of fruits and vegetables. Average weekly fruit and vegetable servings procured by participants will be calculated for baseline and follow-up. Participants will need to have submitted at least 3 weeks of valid purchasing data at both baseline and follow-up to be included in analyses.

Additional trial measures

Intervention dose: Data on participants’ purchasing at the mobile market will be collected by TCMM’s Market Members loyalty programme. At the baseline data collection visit, participants are provided information to join the Market Members loyalty programme that provides customers with a US$5 discount every third shopping trip. Participants at intervention sites, along with other non-research participants interested in shopping at the TCMM, are then enrolled in the Market Member’s programme after randomisation. The Market Members programme is linked to the point-of-sale system that captures the frequency of TCMM shopping trips, amount spent and the number and types of items purchased. Using these data, the dose will be calculated as the average number of monthly TCMM shopping trips during the last 2 months of TCMM service of the 6-month exposure period.

Contamination: Additionally, on the follow-up survey, we are collecting data to assess whether and to what extent contamination occurred at control sites with the following question, ‘During the last 6 months, how often did you shop at the Twin Cities Mobile Market?’ with 5-point response options from never to every week.

Socio-demographic characteristics measured as part of the trial include: gender, age in years, race, ethnicity, household size, receipt of SNAP and/or other benefits, income, languages spoken at home and receipt of food from a food shelf, food pantry or soup kitchen in the past 30 days.

Intervention fidelity data are being measured by an electronic survey used by TCMM bus drivers after each intervention site stop. Drivers record the time spent at the stop and any disruptions to staffing, arrival time, stocking, surrounding conditions or the point-of-sale system. These surveys will be regularly monitored for significant deviations. Data will be summarised in the number of deviations documented during the intervention period across several categories including: typical service, atypical service with specification, scheduled no service due to holiday and unforeseen no service due to weather or other cancellation.

Community site characteristic data are being collected for each site recruited for the study. These data include: distance to nearest grocery store of any kind, density of grocery outlets (of any kind) within 1.0 mile; presence and frequency of a food shelf/pantry at the site; presence and frequency of site-provided transit to grocery store; engagement level of community site partner; resident description; type of building (eg, high-rise, family-site, public housing); affordable housing definition; number of units in the building; resident demographic characteristics if able to be provided by the site; presence of congregate meals onsite; presence of regular community events; presence, if any, of supportive services available (eg, human services coordinator, social worker) and whether these services are dedicated to the site or shared across multiple buildings; and other factors.

Timeline

For each study wave, sites were recruited and then participant recruitment followed at each respective site location. Following baseline data collection, we randomised sites to receive the full-service market intervention or serve as a waitlist control. Sites randomised to receive the market began weekly service following ~1 month of strategic community engagement. After 6 months of market operation, follow-up data collection occurs and intervention sites continue to receive the market. Qualitative interviews with a subset of intervention participants occur after follow-up data collection. The waitlist sites for Waves 1, 2 and 3 receive the intervention after follow-up data collection for their respective wave is complete. See the study timeline pictorially in figure 1.

Figure 1

Timeline of study activities for each wave of the study.

Sample size

As reported in our statistical analysis plan located on ClinicalTrials.gov (NCT05672186), power to detect change in our primary outcome, total HEI score, was computed based on a t-test at the individual participant level, but the sample size was inflated for the group randomisation design by a factor of 1+(m−1)×ICC where m is the average cluster size and ICC is the intracluster correlation coefficient that accounts for correlation in outcomes between participants enrolled at the same site. Using an ICC of 0.004 (from a preliminary customer intercept survey study), the HEI SD of 11.2 (from a study with similar participants led by co-investigator LH),90 a planned attrition rate of 15%, and a two-sided type I error rate of 5%, this study was calculated to have 80% power to detect a mean difference of 4.2 HEI points between intervention and control participants with an average cluster size m of 22 and number of clusters k of 12 for a total of 264 participants (132 per group). A mean difference of 4.2 HEI points is clinically relevant as it is associated with reduced all-cause mortality risk.21 Detecting this difference is also plausible as the mobile market has the potential to improve scores on multiple HEI components, because the mobile market is designed to improve access to nutritious foods including fruits, vegetables, low-fat dairy, lean meats, plant-based proteins and whole grains, while minimising foods with less nutritious profiles, such as those high in added sugars. For our secondary outcomes, power to determine effect is detailed on our statistical analysis plan linked on ClinicalTrials.gov.

Recruitment

Site recruitment: Sites for the trial were recruited in a variety of ways. TCMM staff reached out to potential site locations to query interest. These potential sites were identified by reaching out to sites that had previously contacted the TCMM with interest in service, locating potential sites that were classified in the USDA’s Food Atlas food access map22 as low income and low food access, identifying public and private subsidised housing locations or those that accepted government-funded Section 8 rent vouchers that also were in areas of food access need, leveraging partnerships with Minneapolis Public Health, Minneapolis Public Housing Authority and St. Paul Public Housing Authority, and sending emails and text messages to current customers to query new locations that might benefit from service. All identified potential sites were contacted and screened for eligibility. Additionally, TCMM and research staff visited each site location to meet with residents to ascertain community interest and need for mobile market services. Sites that were eligible and interested then entered a signed agreement with the TCMM to document their interest in receiving mobile market service and willingness for research about the mobile market to happen onsite.

Participant recruitment: Participants from each site were recruited with flyers (posted centrally and delivered door-to-door), tabling in lobby areas and providing information that was published in community newsletters or shared at resident meetings. Those who indicated interest in the study were contacted by phone and screened for eligibility. If participants were found to be eligible after screening and wanted to participate, they scheduled a baseline data collection visit, including consent, at their community site.

Methods: assignment of interventions

Following site recruitment and during enrolment and baseline data collection for each wave, the study statistician, who was blinded to study site names and used computer-generated random numbers, assigned sites to receive the full-service TCMM intervention or serve as a waitlist control. Then, the study statistician was unblinded to site names and shared the site randomisation information with the TCMM co-investigator at The Food Group, who has oversight of the programme. These steps were completed approximately 3 months prior to the start of TCMM service at intervention sites, allowing for TCMM to plan routes and schedules and communicate with any existing TCMM sites about any upcoming schedule changes that might impact them. All other trial investigators and staff, participants and community site staff remained blinded to site randomisation results until after baseline data collection was complete. Investigators and research staff, aside from the study statistician and project manager, are blinded to the links between the site names, assigned site numbers and site randomisation assignment until final outcome analysis is complete.

Methods: data collection, management and analysis

Data collection

Baseline data collection visits occurred at community site locations, most often in community rooms. If needed, screens were used to create private spaces. After participants provided full written consent, participants completed in-person baseline data collection activities. Research staff, trained and certified in conducting 24-hour dietary recall interviews using Nutrition Data System for Research software, conducted recalls with participants using the multiple pass method. Participants were provided with an adapted food amount booklet23 to estimate food and beverage amounts. Participants completed psychosocial surveys via iPads using the secure REDCap (Research Electronic Data Capture) survey platform, with staff providing technological assistance or reading the questions aloud as requested or needed. Research staff also provided participants with instruction on fruit and vegetable procurement data collection activities that would occur following the baseline data collection visit and provided all required materials. After the data collection visit, participants were reminded via phone calls, text messages or emails (based on their individual preferences) weekly to complete and send their fruit and vegetable procurement booklets for 4 weeks. Two additional dietary recalls were conducted within 3 weeks of the baseline visit, allowing for measurement of intake over 2 weekdays and 1 weekend day. At follow-up data collection, the same data collection process is being repeated. Participants are being compensated up to US$200 for completing baseline and follow-up data collection activities (ie, US$10 to US$20 per completed activity).

To promote continued engagement with the study during the intervention period, participants receive four newsletters via mail (approximately every 6 weeks). Each newsletter is brief and includes a reminder to notify study staff if their contact information has changed. To limit attrition due to changing or disconnected phone numbers, contact information for multiple additional people was requested from participants during enrolment. Participants authorised study staff to contact these individuals if staff have difficulty reaching a participant at follow-up using the provided contact information. Prior to contacting alternate contacts, a letter will be delivered to a participant’s place of residence encouraging them to contact the study team. No outcome data will be collected from those who do not schedule a follow-up data collection visit.

Data management

All data linked to participants is coded with a non-identifiable study ID number. Identifiable participant information (names, addresses, etc) collected to maintain contact with participants over the duration of the study is stored in digital files on a secure file server, University of Minnesota Box data servers, which have password protections. These digital files with identifiable information are stored separately from all other study files and are only accessible to the PI and staff whose responsibilities include contacting participants. Physical files with identifiable information are stored in a parallel manner, where identifiable information is stored in separate locked cabinets from all other study files and is only accessible to the PI and staff whose responsibilities include contacting participants. To maintain the separation of identifiable information, participant data collection forms and surveys only contain the study ID number.

As briefly described above, data collected from participants are collected primarily via REDCap surveys118 administered using an iPad at data collection visits. The surveys are directly entered into the secure REDCap database through the survey web interface. If there is a technical failure with the Wi-Fi or iPads, paper-pen surveys are completed, and responses are entered into the secure REDCap database by two different study staff and verified for accuracy. Access to the study’s data in REDCap is restricted to only specific members of the study team by username and password and permission of the PI. REDCap data downloads are stored on secure, password-protected University of Minnesota Box data servers. REDCap data are regularly checked for missing values. Dietary recall and procurement data are entered in Nutrition Data System for Research software by trained staff and additional quality assurance procedures are followed, including range checks and common entry error checks. The data are stored on secure university laptops and backed-up to secure University of Minnesota Box data servers. Paper data files with only study, non-identifiable ID numbers are stored in secured, locked file cabinets.

Data analysis

Our plans for data analysis are reported in detail within our statistical analysis plan housed on the trial’s registry at ClinicalTrials.gov. The main analyses will use an intent-to-treat approach, assigning individuals within a site randomised to the intervention group regardless of individual usage of the TCMM.

The primary analysis will compare the HEI score at approximately 6 months post-baseline between the mobile market and waitlist control groups. More specifically, the analytical model for the primary analysis will be a linear generalised estimating equation (GEE) with the outcome variable (HEI at 6 months) and the explanatory variables (an indicator for randomised intervention, baseline HEI score, days since baseline and precision variables noted below). Individuals will be clustered by site, and the GEE will use an independence working correlation structure. The magnitude and p value of the intervention indicator will be used to carry out the hypothesis test for this analysis.

Secondary outcomes include both continuous and dichotomous variables. Continuous diet-related variables (daily servings of fruit and vegetables consumed, daily kcal energy intake estimates, per cent of calories from added sugar and sodium intake) and the food purchasing outcome (average number of fruit and vegetable servings procured per week) will be analysed similarly to the primary outcome. That is, we will use a linear GEE with independence working correlation structure with adjustment for baseline outcome value, days since baseline and precision variables. If any outcomes are zero-inflated, we will consider regression models which accommodate this (eg, negative binomial regression). Dichotomous food security variables will be analysed using a logistic GEE with independence working correlation structure. Since several hypotheses on multiple secondary outcomes are of interest, for secondary analyses we will report CIs rather than null hypothesis significance testing, an approach in line with current recommendations.24

While the main analyses will be conducted with an intent-to-treat approach, we will also perform dose-adjusted analyses, as sensitivity analyses, to investigate the association between frequency and volume of use of the mobile market and the outcomes of interest and whether these usage measures mediate the effects of the intervention.

All models will be adjusted for precision variables. For dietary and food/nutrition security outcome analyses, precision variables are as follows: age; gender; education; race; and ethnicity. These variables were selected as all are known to be associated with these outcomes. Food purchasing outcome analyses precision variables are as follows: age; gender; education; race; ethnicity; and household size. Household size was also included because this food procurement outcome is measured for the household, and thus, it will be important to control for how many people are in the home. Additionally, we will consider adjustment for any imbalances in socio-demographic characteristics between study groups. If variables appear to be scientifically and meaningfully different between groups, we will decide whether to include any of these variables prior to running analyses. We will not use p values to assess for differences by randomised group, in alignment with current statistical recommendations.1

To reduce the influence of outlying and missing data values, we will form analytical samples according to prespecified exclusion criteria in our statistical analysis plan. Additionally, for variables with >10% missing values, we will compare characteristics between individuals with missing and non-missing values. If there appears to be a substantial difference in the characteristics of those with missing and non-missing values, we will perform multiple imputation for analyses where these variables are used as explanatory variables. We will not perform multiple imputation for outcome variables.

Methods: monitoring

Data collection measurement and the intervention of this trial pose minimal risks to participants. As such, data safety monitoring is led by the PI and data safety parameters are formally reviewed by an internal Data Safety Monitoring Officer (DSMO). The DSMO is an National Institutes of Health (NIH)-funded researcher at the same institution as the study team but independent of the study team in accordance with our data safety monitoring plan as approved by the Institutional Review Board (IRB) and funder, the National Institutes of Nursing Research (NINR). Aspects of the study that are reported to and monitored by the PI and DSMO for trial performance and safety include: participant accrual (monthly), retention (monthly during each wave), intervention fidelity at intervention sites (monthly), adverse events (within 24–72 hours of study staff learning of an event) and safety reports (at a minimum of every 6 months). In the event of a study-related adverse event or death of a participant (study-related or not study-related), the event is reviewed with the DSMO in real-time and a report is also submitted to the IRB within 24–72 hours. Additionally, NINR is notified of any study-related serious adverse events within 7 days; all other events are included in the annual report to the NINR.

In addition to monitoring by the DSMO, research staff, under the supervision of the PI, are auditing all study-related documents, data and materials of randomly selected participants (1 out of every 50 participants recruited) for compliance with all IRB requirements/protocols including documentation of written informed consent, data collection completion and data entry verifications. The trial is also subject to routine or for-cause auditing by the Quality Assurance Program of the Human Research Protection Program at the University of Minnesota; we will comply with any requests for audit.

Ethics and dissemination

Research ethics approval and consent procedures

Trial protocols and any amendments have been approved by the IRB of the University of Minnesota. Substantial protocol amendments are communicated with investigators, the funding agency (NINR) and via an update at ClinicalTrials.gov. Participants were provided with a consent form in advance of their baseline data collection visit, and trained study staff reviewed consent forms with participants prior to participants providing written consent at the baseline visit. Additionally, in the process of informed consent, participants were informed that effort will be made to limit the use and access of participant personal information. Specifically, participant personal information is only accessible to staff needing the information to complete their job duties (research staff who contact participants; organisations like the IRB and other representatives of this institution, including those that have responsibilities for monitoring or ensuring compliance). Participants were also informed during the consent process that the research team will use and may share anonymised data for future research.

Declaration of interests

The PI has received funding from NINR to conduct this trial. The PI and all authors have no other financial or competing interests. There are no limitations on investigators’ access to the final data set.

Ancillary and post-trial care

The trial has been deemed no more than minimal risk, so it is unlikely that ancillary or post-trial care would be needed. Participants were informed during the consent process that if participating in research results in an injury, treatment will be available, including first aid, emergency treatment and follow-up care as needed. Care for such injuries will be billed to participants in the ordinary manner, to them or their insurance company.

Dissemination plans

The full IRB protocol is included with this online supplemental file 1 and is available via ClinicalTrials.gov, as is our full statistical analysis plan. The investigators plan to disseminate trial findings to academic and community audiences via publications and presentations. Additionally, we plan to prepare an anonymised data set that can be shared with the research community and community partners. We will also disseminate aggregate trial findings to our participants. We will not use professional writers and will extend authorship to any members of the study team (academic and community) who meet authorship guidelines as outlined by the International Committee of Medical Journal Editors.

Supplemental material

Trial registry

The trial is registered at ClinicalTrials.gov (NCT05672186). We also provide the information that aligns with WHO registration details in online supplemental file 2, table 1 as well.

Supplemental material

Ethics statements

Patient consent for publication

Acknowledgments

We would like to acknowledge the following study staff, Julia Steiner, and student study staff, Emiliano Chavez-Dreier, Blaine Damte, Sebrum Herron, Mikayla King, Morgan McNamara, Katelyn Nettesheim and Briana VanCura, for their assistance with recruitment, data collection and data entry. We also want to express gratitude for the Twin Cities Mobile Market and building site staff as well as all our participants who make this work possible.

References

Footnotes

  • X @DrJWolfson

  • Contributors MLH: Principal investigator and guarantor of the study, secured grant funding, provided oversight of the trial protocol development, provides oversight of the trial implementation and drafted and revised the manuscript. Accepts full responsibility for the trial and this manuscript. KSG: Project manager for the study, assisted with development of the protocols, leads trial implementation and reviewed and revised the manuscript. SW: Co-investigator on the study, supported protocol development and directly supports mobile market implementation, and reviewed and revised the manuscript. JF: Co-investigator on the study, assisted with securing grant funding, supported protocol development and supports trial implementation, and reviewed and revised the manuscript. ML: Co-investigator on the study, assisted with securing grant funding, supported protocol development and trial implementation, and reviewed and revised the manuscript. JW: Co-investigator on the study, assisted with securing grant funding, supported protocol development and supports trial implementation, randomised participants per protocol and reviewed and revised the manuscript. LH: Co-investigator on the study, assisted with securing grant funding, supported protocol development and supports trial implementation, and reviewed and revised the manuscript.

  • Funding Research reported in this publication was supported by the National Institute of Nursing Research of the National Institutes of Health under award number R01NR020539 (principal investigator: MLH). This study used REDCap (Research Electronic Data Capture) for data collection which was supported by Clinical and Translations Science Institute grant (UL1TR002494) from the National Institutes of Health’s National Center for Advancing Translations Sciences. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

  • Competing interests The PI has received funding from NINR to conduct this trial. The PI and all authors have no other financial or competing interests. There are no limitations on investigators’ access to the final data set.

  • Provenance and peer review Not commissioned; peer reviewed for ethical and funding approval prior to submission.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.