The Pathways study is conducted in the province of Québec, Canada. It was developed in close partnership with public health, provincial, and regional organizations (i.e., Food banks Quebec) that contributed to its design and implementation. The Pathways study is funded by the Canadian Institutes of Health Research (grant# omitted for blind review), with additional financial support from the Quebec Ministry of Health and Social Services, the Foundation of Greater Montreal, and Mission Inclusion. We received ethical approval from the Research Ethics Board of the Université de Montréal (n. certificate blinded for peer review).
Design, setting, and target population
Community organizations (through which users were recruited) were situated in rural, semi-urban, and urban areas in four administrative regions in the province: the Island of Montreal, Lanaudière, Mauricie-Centre-du-Québec, and Estrie. All the regions have urban, semi-urban, and rural settings except for Montreal, which is solely urban and semi-urban.
The study population consisted of individuals who used a food donation program offered by a community organization in one of the study regions for the first time in the preceding 6 months. Over the course of the Pathways study, participants were interviewed three times: at baseline (T0), defined as within six months after their first use of a food donation program (September 2018–January 2020); at 12 months follow-up (T1; September 2019–January 2021); and at 24 months follow-up (T2; September 2020–January 2022).
Recruitment
The sample of participants was assembled in two phases. We first recruited community organizations in the four regions from which we then recruited first-time users.
Recruitment Of Community Organizations
Across the four regions, we inventoried a total of 423 community organizations offering FD or FD + CBP. All organizations were contacted to assess eligibility, explain the study, and invite them to help the research team recruit participants. Organizations offering intermittent food donations (e.g. at Christmas) and those providing meals to specific groups (e.g., Breakfast Club) were excluded. Overall, 149 organizations agreed to participate, 17 organizations could not be reached, 160 did not meet inclusion criteria, and 97 refused to participate. We conducted a 30-minute phone interview with one key informant from each organization (usually the director), asking them to describe the activities and services offered by their organization (i.e., whether they provided FD or FD + CBP programs) and to confirm their willingness to help identify and recruit first-time users of their food donation program. Of the 149 organizations, 117 provided the names of potential participants. We were unable to enroll participants from 11 of these organizations. Ultimately, participants were recruited from 106 organizations (71.1% of organizations that agreed to participate). The recruited organizations were categorized by type of program(s) offered: FD (43 organizations: 40.6%) or FD + CBP (63 organizations: 59.4%) and by their setting: urban (48 organizations: 45.3%); semi-urban (29 organizations: 27.4%), and rural (29 organizations: 27.4%). More details on recruitment are provided elsewhere [citation omitted for peer review].
Recruitment Of Participants And Retention
Based on a previous study of new food banks users in the region of Montreal [39], we estimated that, with 1,008 participants at T2, p < 0.05, power 0.80, and intra-class correlation coefficients of 0.05 for HFI, 0.006 for self-rated physical health, 0.01 for self-rated mental health, and 0.106 for social support, we could detect small effect sizes. Based on our retention strategy and previous study results [39], we expected 70% retention at T1 and 80% at T2. With these retention rates, we estimated that 1800 respondents would be needed at T0.
Contact details (names, telephone numbers, or emails) for a total of 1784 individuals willing to participate in the Pathways study were provided by participating food donation organizations or were obtained directly by interviewers on site during food-aid distribution. In total, 1001 participants were enrolled at baseline (56.1%). Of the 783 participants who were not interviewed, 227 (29%) could not be contacted by the research team; 47 (6%) did not meet our inclusion criteria (see Table 1); 125 (16%) were not interested or did not have time to participate; 235 (30%) did not participate for other reasons; and 149 (19%) did not show up to the face-to-face baseline interview appointment. Of the 1001 people enrolled in the study, 181 (18.1%) were recruited in rural settings, 250 (25%) in semi-urban settings, and 572 (57%) in urban settings (Table 2). Considering the type of organizations, 379 (37.9%) enrolled participants were recruited from FD organizations and 622 (62.2%) from FD + CBP organizations.
Table 1
Recruitment of participants: exclusion criteria
Exclusion criteria | Justification |
---|
Individual participating in a food donation program in another organization in the preceding 12 months. | |
Age over 63 years at baseline | Individuals over this age are eligible for the senior citizens guaranteed income supplements over the course of the study (which would alleviate a primary cause of food insecurity) |
Individuals in situation of homelessness | Homeless people constitute a small minority of users of the organizations of interest in this study. We were not able to assure adequate follow-up of participants with no home address. |
Individuals living with a person already enrolled in the study | |
Individuals who do not speak French nor English. | The diversity of languages spoken in Montreal is such that translating/back-translating questionnaires or using interpreters would have incurred significant costs and introduced measurement bias. In addition, new immigrants to Quebec are enrolled in French integration classes free-of-charge and most quickly become functional in French. |
Table 2
Sample size and retention rate by region and setting*
| | Total | Estrie | Mauricie Centre du Québec | Lanaudière | Montreal |
---|
| | n % | n % | n % | n % | n % |
Rural | T0 | 181 | 74 | 70 | 37 | n.a. |
| T1 retention | 138 (76.2%) | 52 (70.3%) | 56 (80.0%) | 30 (81.1%) | n.a. |
| T2 retention | 120 (66.3%) | 44 (59.5%) | 51 (72.9%) | 25 (67.6%) | n.a. |
Semi-urban | T0 | 250 | 52 | 60 | 39 | 99 |
| T1 retention | 199 (79.6%) | 39 (75.0%) | 49 (81.7%) | 28 (71.8%) | 83 (83.8%) |
| T2 retention | 177 (70.8%) | 33% | 46 (63.5%) | 23 (59%) | 75 (75.8%) |
Urban | T0 | 570 | 155 | 131 | 49 | 235 |
| T1% retention | 408 (71.6%) | 92 (59.4%) | 98 (74.8%) | 31 (63.3%) | 187 (79.6%) |
| T2% retention | 345 (60.5%) | 70 (45.1%) | 85 (64.9%) | 20 (40.8%) | 170 (72.3%) |
Total | T0 | 1001 | 281 | 261 | 125 | 334 |
| T1% retention | 745 (74.4)% | 183 (65.1%) | 203 (77.8%) | 89 (71.2%) | 270 (80.8%) |
| T2% retention | 642 (64.1%) | 147 (52.3%) | 182(69.7%) | 68 (54.4%) | 245 (73.4%) |
T2 retention rates are calculated on the T0 respondents. |
Insert Table 1 and 2 here
Of the 1001 participants at baseline (T0), 745 (74.4%) completed the one-year follow-up interview (T1) and 642 (86.2%) completed the T2 survey (Table 2). The overall two-year retention rate of the study is 64.1%. The retention rate at T2 was slightly lower in urban areas compared to semi-urban and rural areas. Moreover, there are regional differences; the Montreal region had higher retention rates than the other three regions (Table 2).
Data Collection
At baseline, participants were interviewed face-to-face by trained interviewers. The 60-minute bilingual (French/English) questionnaire had been mostly developed, validated, and used in a past study [39]. A similar questionnaire was administered at the two follow-ups, after one (T1) and two years (T2) respectively. Questions were added at T1 and T2 to assess in greater detail pathways of use of food security programs. At the mid-point of the first follow-up data-collection period (precisely on March 13 2020), the first lockdown related to the COVID-19 outbreak was implemented province-wide. Thus, for T1, only 50% of participants were interviewed face-to-face; the remaining half were interviewed remotely (over the phone or on web-based software). All T2 questionnaires were administered remotely. In addition, a section with questions assessing the impact of the COVID-19 pandemic was added to the T2 questionnaire.
Measures
Food security organization type at recruitment was categorized in one of two types: (1) Food donation only (FD), which included organizations providing food baskets to users free-of-charge or at a symbolic price; and (2) Food donation plus capacity building program (FD + CBP), which included those organizations providing food donations as well as food-related capacity-building programs aimed at empowering users and fostering their social integration, while facilitating acquisition of quality food at lower cost (e.g., collective kitchens, collective gardens, food-buying groups, cooking or nutrition workshops).
Geographical setting comprises urban, semi-urban, and rural areas, defined according to the definitions of Statistics Canada’s Census Metropolitan Areas (CMAs) and Census Agglomerations (CAs) [40] in conjunction with Regional County Municipalities’ development plans used by administrative regions in Quebec to structure development and to allocate resources to municipalities. Urban settings were defined as urban centers with a core population ≥ 50,000 for CMAs and ≥ 10,000 for CAs. They are characterized by higher population and building density as well as more mixed land uses (residential, commercial, services, industries). Considering CMAs and CAs, semi-urban settings are generally located on the periphery of urban centers; they have relatively dense built environments and mixed land uses. Outside of CMAs and CAs, semi-urban settings include municipalities with ≥ 5000 inhabitants with an important land-use mix, characterized by low diversity; are often close to rural areas; and are not contiguous to urban centers. Rural settings are characterized by low population density and a lack of diversity in land use. They are composed of municipalities inside or outside CMAs and CAs. There are no rural areas within the Montreal CMA.
Primary And Secondary Outcomes
Pathways of use of food donation programs and household food security are the main outcome variables in the Pathways study. Other variables of interest investigated in the study are diet, food-related competencies, food acquisition patterns, perceived food environment, physical and mental health, social support and isolation, negative life events, socio-demographic characteristics and housing conditions. The impact of the COVID-19 pandemic and lockdown was investigated through specific questions during the second follow-up (T2).
Pathways of use of food donation program was assessed at T1 and T2 by asking participants about their use of food donation programs in each month of the preceding year with the following questions: “Since our last interview on [DATE] have you used food donation programs? If yes, did you use it in [list all months]?” The use of capacity-building programs was also investigated with similar questions.
Household food security was measured using the Household Food Security Survey Module (HFSSM), a well-validated, 18-item scale that measures inadequate or insecure access to food due to financial constraints [41]. Developed by the U.S. Department of Agriculture (USDA) [42], it was subsequently approved by Health Canada as the measurement tool for HFI in Canada [43]. The questions differentiate the experiences of adults from those of children, recognizing that in households with children, adults might compromise their own food intake to reallocate scarce resources for children. Based on their responses to the HFSSM, households were categorized into one of the four food-security categories: food secure (households had access to enough food at all times throughout the previous year), marginally food secure (some concern or problem of food access), moderately food insecure (compromises in the quality and/or quantity of food consumed), and severely food insecure (extensive compromises, including reduced food intake) [1, 43].
Diet was measured using an adapted version of the Short Diet Questionnaire (SDQ), a 36-item food-frequency questionnaire developed and validated for use in Quebec [44]. The frequency of restaurant meals and meal frequency were also evaluated through ad hoc multiple-choice questions.
Food-related competencies were assessed with eight questions related to food planning and preparation based on questions from the Canadian Community Health Survey. Examples of questions are “Do you try to avoid foods that are high in sugar, salt, and fat?” or “In the past 12 months, did you shop with a grocery list?” [45].
Self-rated physical and mental health was assessed with the SF-12v2 [46], which generates two component summary scores, the Physical Component Summary (PCS) and the Mental Component Summary (MCS). We further assessed psychological distress with the Kessler six-item scale [47]. A score of 13 or higher on the Kessler six-item scale indicates psychological distress.
Perceived social support was measured with a modified version of the Multidimensional Scale of Perceived Social Support (MSPSS) [48], which consists of 12 items evaluating perceived social support from friends, family, and a special other.
Social isolation was measured with five questions from the Social Isolation subscale of the Nottingham Health Profile scale [49].
Food acquisition patterns. The type of food store used, its location, frequency of food purchasing, frequency of use of the food donation program, transportation mode and travel time to the most commonly used grocery store were investigated. Summertime fruit and vegetables market use and gardening, either at home or in a community garden, were also assessed.
Perceived food environment was evaluated with the Perceived Food Environment Questionnaire, which assesses the quality, variety, quantity, and affordability of healthy food in the reference grocery store, as well as the accessibility of restaurants and of a wide variety of food near the home [50]. Difficulties in shopping (physical and financial difficulties) were also evaluated with questions previously tested [51].
Socio-demographic characteristics (i.e. employment, income, source of income) and housing conditions (i.e,. dwelling type and characteristics) of participants were assessed with questions from the Canadian Community Health Survey and the Canadian census.[45]
Negative life events in the year preceding the survey were used to evaluate events that might have brought people to ask for food aid; these were measured with questions from the Holmes and Rahe stress scale.[52].
Lastly, questions were added to the T2 questionnaire to investigate the impact of the COVID-19 pandemic, to document changes in the household’s financial situation, changes in employment (job losses, job changes, or temporary layoffs), and to determine if the respondent or any close relative had been affected by COVID-19.