Evidence synthesis – Food-based interventions to mitigate household food insecurity in Canada: a systematic review

Abstract

Introduction: Household food insecurity (HFI) is a persistent and important public health and policy concern within Canada that continues to be widespread in the face of economic uncertainties and inflation. The objective of this systematic review was to synthesize the evidence on food-based interventions that could reduce HFI in Canada.

Methods: Studies that assessed a food-based intervention that might reduce food insecurity and measured HFI were included, regardless of whether that was the primary purpose of the study. Four databases were searched up to 19 February 2025. Screening of abstracts and full texts, data extraction, assessments of risks of bias and certainty of the evidence were conducted independently by two reviewers. PROSPERO CRD42021254450.

Results: Exposure to food voucher programs may reduce HFI, but exposure to food box, community gardening, school food, hunting and fishing, and food charity programs may have little to no effect on HFI. The rate of utilization of food banks by food-insecure households may be low and depends upon food insecurity level and population group.

Conclusion: Food charities may be a last resort for those in need of short-term access to emergency food (i.e. populations experiencing homelessness). However, given the pervasive nature of HFI as a marker of deprivation, it is unlikely that food-based responses will have a major impact on overall HFI, which is primarily an economic problem. A more comprehensive public policy approach to mitigate HFI is likely required.

Keywords: household food insecurity, food bank, food charity, food-based intervention, systematic review

Introduction

In 2022, 2.7 million Canadian households experienced food insecurity in the past 12 months, including 1.8 million children aged less than 18 years.^{Footnote 1} This is the highest number recorded in the 17 years since Canada started monitoring household food insecurity (HFI).^{Footnote 1} These statistics refer specifically to households’ inability to acquire or consume an adequate quantity of food or the uncertainty that they will be able to do so because of financial constraints.^{Footnote 1}

HFI is strongly associated with multiple adverse health outcomes for children and adults in Canada, including heightened nutritional vulnerability;^{Footnote 2Footnote 3Footnote 4Footnote 5Footnote 6Footnote 7} increased risk of type 2 diabetes;^{Footnote 8} poorer mental health;^{Footnote 9Footnote 10Footnote 11Footnote 12Footnote 13Footnote 14Footnote 15} higher rates of infectious^{Footnote 16} and noncommunicable^{Footnote 10} diseases, injuries^{Footnote 17} and chronic pain;^{Footnote 18} poorer disease management;^{Footnote 16Footnote 19Footnote 20Footnote 21Footnote 22} higher rates of health care utilization;^{Footnote 23Footnote 24Footnote 25Footnote 26Footnote 27Footnote 28} and premature mortality.^{Footnote 29Footnote 30} These relationships are graded, with more severe food insecurity associated with both a higher number of and worse health outcomes^{Footnote 23Footnote 24Footnote 25Footnote 27} even after controlling for income and other sociodemographic characteristics.

It is worth noting that HFI, as it is defined in this manuscript and by public policy in Canada, is not the absence of food security. According to the Food and Agriculture Organization, “food security exists when all people, at all times, have physical and economic access to sufficient, safe and nutritious food to meet their dietary needs and food preferences for an active and healthy life.”^{Footnote 31,p.1} Chronic HFI is a narrower concept and a strong, validated indicator for social policy and population health.^{Footnote 1} This construct is distinct from broader definitions related to community food security that include constructs of physical, social and economic access to sufficient, safe and nutritious food that meet peoples’ dietary needs and preferences.^{Footnote 32}

Both public policy and food-based interventions have been proposed to reduce HFI. Public policy interventions address the upstream economic determinants of HFI and include indirect supports such as cash transfers, housing assistance and market subsidies. Food-based interventions address shortages of food at the household level. Food-based interventions include food charities that distribute food directly to people, for example, via food banks and soup kitchens, community food programs such as cooking classes, debt counselling for better food budget management, community kitchens and community gardening programs. These programs intend to improve access to food (e.g. by providing free or reduced-cost food), thus reducing the need to expend household resources on food purchases, or they aim to increase program participants’ abilities to manage scarce resources (e.g. by increasing cooking and shopping skills or addressing debt).

There has been little research on the effectiveness of these interventions in reducing HFI in Canada.^{Footnote 33} A synthesis of the available evidence is particularly important given the surge in federal and provincial government funding of these programs throughout the COVID-19 pandemic^{Footnote 34} and a parallel increase in the cost of living since 2022.^{Footnote 35} This policy response is not new, with some food-based interventions such as food banks dating back to the 1980s.^{Footnote 36}

The objective of this systematic review is to synthesize the evidence on the effectiveness of food-based strategies to reduce HFI in Canada.

Methods

Systematic review registration

This systematic review was guided by the Cochrane Handbook for Systematic Reviews of Interventions^{Footnote 37} and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Guideline.^{Footnote 38} The initial overarching research question was, “What interventions are effective in reducing HFI in Canada?”

The systematic review protocol was created a priori and registered in PROSPERO (CRD42021254450).

During the systematic review process and after discussions with HFI experts [VT, LM], it became clear that it would be best to group the interventions into two categories based on the level at which they work, that is, public policy interventions versus food-based interventions. Given that these are distinct intervention categories, this separation facilitated analysis and reporting. Public policy interventions address the underlying determinants of food insecurity and target the economic vulnerability of households. Food-based interventions address food shortages at the household level. The analyses and reporting were conducted separately for the two types of interventions. The systematic review on public policy interventions has been published elsewhere.^{Footnote 39}

This current systematic review summarizes the evidence on food-based interventions to reduce HFI and answers two key questions. The first of these key questions, KQ1, asked, “What is the impact of exposure to a food-based intervention on HFI in Canada?”

Given the low number of studies reporting on the effectiveness of food charity interventions and evidence that food bank utilization rates are low, we included studies of the utilization of food charity interventions by food-insecure households. The second key question, KQ2, reflects the inclusion of these studies: “What is the rate and frequency of utilization of food charity interventions in Canada among food-insecure households?”

Eligibility criteria

The population, interventions, comparators and outcomes (PICO) model was used to facilitate search strategy development.

• Population: Households (KQ1) or food-insecure households (KQ2) in Canada.
• Intervention or utilization: Studies that sampled households that were exposed to (KQ1) or that utilized (KQ2) an intervention with the aim of reducing HFI, regardless of whether that was the primary aim of the study.
• Control: Studies with a comparator group (contemporaneous, historical or where participants act as their own control) (KQ1) or no comparator group (KQ2).
• Outcome: Any outcome that aimed to assess HFI or a construct aligned with hunger (KQ1) or the use of food charity interventions (KQ2).

Other eligibility criteria include the following:

• Dates: All studies published from 2000 onwards.
• Languages: English and French.
• Study design: Primary research studies, including controlled trials and observational studies.

For a full list of inclusion and exclusion criteria, refer to Supplementary Material A.

Search strategy

The search strategy was developed by a Health Canada research librarian in collaboration with the authors (see Supplementary Material B for details of the search strategy). The search strategy underwent a Peer Review of Electronic Search Strategies,^{Footnote 40} and was independently reviewed for quality by a second librarian.

We conducted searches in the following four electronic bibliographic databases: EconLit, Embase, Ovid MEDLINE and Scopus. The search was conducted in April 2021 and updated in November 2022, October 2023 and February 2024 and, finally, on 19 February 2025. We conducted a grey literature search using Google Scholar and a targeted website search for key terms in June 2021. In addition, the reference lists of 17 related reviews were hand searched, and experts were consulted to ensure that the database searches did not miss any studies.

Study selection

Search results were imported into the web-based literature review software DistillerSR version 2.37 (DistillerSR Inc., Ottawa, ON, Canada), and duplicates were removed. Two reviewers [LI, TC, AC, EV, SK or CL], working independently, screened all titles and abstracts for potential eligibility using a standardized form developed a priori and piloted by all these reviewers. The reviewers then conducted full-text screenings of the retrieved articles that had passed this first stage of screening. (For a list of excluded studies and the reasons for exclusion at either stage of screening, see Supplementary Material C.)

Disagreements between reviewers were resolved through discussion, with a third reviewer if necessary.

The process was similar for decisions regarding data extraction, risk of bias and Grading of Recommendations, Assessment, Development and Evaluation (GRADE).

Data extraction

A data extraction form for recording relevant information from each included study was developed and piloted by reviewers prior to starting data extraction. For all included studies, the following information was extracted: citation information (authors, title, journal or source, year of publication, language of publication); study information (objectives, study design, time period, description of intervention and the method or tool used to measure HFI); participant characteristics (including any subgroups of interest); and outcomes of interest (HFI severity and utilization rates of food charities).

Data analysis

The final dataset was exported to Microsoft Excel 365 (Microsoft Corp., Redmond, WA, US) for analysis. Narrative synthesis was performed on the dataset. Where three or more studies measured the same outcome, a random-effects meta-analysis using the DerSimonian and Laird method^{Footnote 37} weighting procedure was conducted in the metaprop package in Stata version 18 (StataCorp LLC, College Station, TX, US). Logit transformation was used and the data were then subgrouped by population and level of HFI. Heterogeneity was only calculated when there were four or more observations. Subanalyses by sociodemographic factors were included where possible.

Risk of bias

We used the Risk of Bias in Non-randomised Studies – of Intervention (ROBINS-I)^{Footnote 41} and Risk of Bias in Non-randomized Studies – of Exposures (ROBINS-E)^{Footnote 42} tools and the Cochrane Collaboration’s tool for assessing risk of bias in randomized trials version 2 (RoB 2)^{Footnote 43} to assess risk of bias in the included studies that examined the effectiveness of food-based interventions (KQ1). The JBI critical appraisal tool for cross-sectional studies was used to assess risk of bias in studies of food charity utilization (KQ2) as cross-sectional data were extracted from all of the included studies.^{Footnote 44Footnote 45}

Certainty of the evidence

The GRADE framework was used to rate the certainty and strength of the body of evidence.^{Footnote 46} The purpose of GRADE is to rate the quality of the evidence in relation to research questions in a systematic and transparent manner (see Supplementary Material D for the GRADE decision rules applied in this systematic review). The evidence in the studies that addressed KQ1 was first assumed as being of high certainty, and this was rated down as per recent guidance on non-randomized studies.^{Footnote 47} The studies that addressed KQ2 were also initially assumed to have high-certainty evidence as an adapted GRADE approach was used to assess them.^{Footnote 48} In assessments of rates and frequency of utilization, observational studies can provide robust estimates when they use appropriate measures and enrol representative populations.

The certainty was downgraded for each outcome to “moderate,” “low” or “very low” if there were serious or very serious concerns that reduced certainty in the outcome estimates across the following five domains: risk of bias, inconsistency, indirectness, imprecision or publication bias.^{Footnote 47} The wording of the summary statements regarding the certainty of the evidence is based on published guidance.^{Footnote 49}

Results

A total of 8542 references were screened for eligibility for both key questions. Of these, 21 articles reported on exposure to or utilization of food-based interventions intended to reduce HFI in Canada (Figure 1).

Figure 1: Descriptive text

This figure illustrates the PRISMA flow diagram of searches, which ultimately led to the identification of studies via databases and other sources.

In the identification step of the process, records were identified from:

• Databases (n = 8542)
  • EconLit (n = 133)
  • Embase (n = 2377)
  • Medline (n = 1920)
  • Scopus (n = 4112)
• Grey literature (n = 33)
• Search verification (n = 1)

Of those, n = 4005 duplicate records were removed before screening.

In the screening step of the process, n = 4571 titles and abstracts were screened, leading to the exclusion of n = 3288 records. 1283 full-text articles were therefore screened, which led to n = 1262 reports being excluded for the following reasons:

• Not about utilization or uptake of a food-based intervention (n = 1141)
• No quantitative outcomes (n = 68)
• No odds ratio or frequency data for using a food bank (n = 19)
• Other non-primary articles (n = 11)
• No full text available (n = 10)
• Review (n = 12)
• Not in English or French (n = 1)

This led to n = 21 relevant articles on food-based interventions being included.

Abbreviation: PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Descriptive summary of included studies

The included studies (n = 21) were published between 2000 and 2023, with more than half (n = 13) published since 2014 (see Table 1 for a summary of the study characteristics). Ten studies reported on the impact of exposure to food-based interventions on HFI (KQ1), and 13 on the rates and frequency of utilization of food-based interventions by food-insecure households (KQ2); 2 of these studies reported on both the impact and utilization rate of food-based interventions. The characteristics of the studies are summarized in Table 1. Detailed risk of bias results for all studies are in Supplementary Material E.

Table 1. Characteristics of all included studies, by program type (n = 21)

First author, publication year Study design Dataset	Location Period of data collection Intervention	Population	Comparator	Food insecurity collection method Reference time period	Total sample size, n	Risk of bias
Food voucher programs
Aktary, 2024Footnote 50 Randomized controlled trial Data collected by the authors for the purpose of this study	British Columbia 2019 Farmers’ Market Nutrition Coupon Program	Adults living with low income who have never participated in the Farmers’ Market Nutrition Coupon Program	Adults living with low income who do not receive the Farmers’ Market Nutrition Coupon Program coupons	HFSSM Past 30 days	285	Low
Heasley, 2021Footnote 51 Single-arm pre–post design Data collected by the authors for the purpose of this study	Guelph, Ontario 2019–2020 Food prescription voucher program	Patients with at least 1 diagnosed cardiometabolic condition or micronutrient deficiency (or both) and experiencing food insecurity	Participants pre-intervention	HFSSM Past 4 months	60	High
Food box programs
Miewald, 2012Footnote 52 Cohort (2-arm pre–post design) Data collected by the authors for this study	British Columbia Baseline: 2008; followup: 8 months after baseline Food subscription box	Households with low incomes that received a weekly food box	Households with low incomes that accessed the same social services but did not receive a weekly food box	HFSSM Not specifiedFootnote a	192	High
Gardening programs
Sandha, 2021Footnote 53 Analytic cross-sectional Data collected by the authors for the purpose of this study	Prince Edward Island 2013 At-home gardening	Mothers of children aged 0–6 years who accessed the Family Resource Centre’s gardening program	Mothers of children aged 0–6 years who did not access the Family Resource Centre’s gardening program	HFSSM Reference time period not specifiedFootnote a	282	Some concerns
School food programs
Roustit, 2010Footnote 54 Analytic cross-sectional Health and Social Survey of Québec Children and Youth	Quebec 1999 School food programs (free or reduced-price snacks or meals)	Children and adolescents attending a primary or secondary school with a food supplementation program	Children and adolescents attending a primary or secondary school without a food supplementation program	Three statements from the Radimer/Cornell questionnaire Reference time period not specified	2346	High
Hunting and fishing interventions specific to Indigenous communities
Blanchet, 2021Footnote 55 Analytic cross-sectional Data collected by the authors for the purpose of this study	Syilx People of the Okanagan Nation, British Columbia 2018 Salmon consumption	Adults who self-identified as or were in a kin relationship with a person who self-identified as Syilx of the Okanagan Nation and who eat locally caught salmon	Adults who self-identified as or were in a kin relationship with a person who self-identified as Syilx of the Okanagan Nation and who eat non-locally caught salmon or no salmon	HFSSM adapted to Indigenous populations in Canada Not specifiedFootnote a	265	High
Thompson, 2012Footnote 56 Analytic cross-sectional Data collected by the authors for the purpose of this study	Manitoba 2008–2012 Country Food Programs	Adults in 14 remote communities in northern Manitoba	NA	HFSSM and 3 supplemental questions on gardening, hunting and fishing Not specifiedFootnote a	533	High
Food charity interventions
Loopstra, 2012Footnote 57Footnote b Longitudinal cohort (single arm) Data collected by the authors for the purpose of this study	Toronto, Ontario Baseline: 2005–2007; followup: 2006–2008 Local food banks	Adults in families with low incomes, based on census tracts, who used a food bank in the last 12 months	Adults in families with low incomes, based on census tracts, who did not use a food bank in the last 12 months	HFSSM Past 12 months	371	High
Rizvi, 2021Footnote 58Footnote b Longitudinal cohort (single arm) Data collected by the authors for the purpose of this study	Ottawa, Ontario 2017–2019 Local food banks	People who accessed community food banks, after 6, 12 and 18 months	People who accessed community food banks, at baseline	HFSSM Past 12 months	401	High
Roncarolo, 2016Footnote 59 Longitudinal cohort (2 arms) Data collected by the authors for the purpose of this study	Montréal, Quebec Baseline: 2011–2012; followup: 9 months after baseline Local food banks	Individuals accessing traditional food banks or participating in alternative community interventions (community kitchens, community gardens and buying groups)	Population prior to food bank interventions	HFSSM Past 12 months	824	High
Food charity utilization
Daly, 2023Footnote 60 Cross-sectional Assessing the Impacts of COVID-19 on Mental Health survey	National 2020–2021 Local food charities	Adults in food-insecure households stressed or worried about having enough food to meet the household’s basic needs as a result of the COVID-19 pandemic in the past 2 weeks	NA	One question on food worry and one question from the HFSSM Past 12 months	477	Low
Holmes, 2019Footnote 61 Cross-sectional Data collected by the authors for the purpose of this study	Vancouver, British Columbia 2015 Local food banks	Food-insecure households that accessed a local food bank	NA	HFSSM Not specifiedFootnote a	77	High
Kirkpatrick, 2009Footnote 62 Cross-sectional Data collected by the authors for the purpose of this study	Toronto, Ontario 2005–2007 Local food banks	Food-insecure households	NA	HFSSM Past 12 months	484	High
Loopstra, 2012Footnote 57Footnote b Cross-sectional data taken from a cohort study Data collected by the authors for the purpose of this study	Toronto, Ontario Baseline: 2005–2007; followup: 2006–2008 Local food banks	Adults in families with low incomes, based on census tracts, who used a food bank at least once in the last 12 months	Adults in families with low incomes, based on census tracts, who did not use a food bank in the last 12 months	HFSSM Past 12 months	371	Low
MacBain, 2023Footnote 63 Cross-sectional Data collected by the authors for the purpose of this study	Hamilton, Ontario 2021 Local food charities	Food-insecure households	NA	Hunger Vital Sign tool Past 12 months	173	Low
McIntyre, 2000Footnote 64 Cross-sectional NLSCY Cycle 1	National (excluding the territories) 1994 Local food banks	Households with children aged less than 18 years	NA	Single question on child hunger Any experience (“ever experienced being hungry”)	16 639	Some concerns
McIntyre, 2012Footnote 65 Repeated cross-sectional NLSCY Cycles 2 and 7	National (excluding the territories) 1996–1997; 2006–2007 Local food banks	Households with children aged 2 to 9 years	NA	Single question on child hunger Any experience (“ever experienced being hungry”)	Cycle 2: 8165 Cycle 7: 15 691	Some concerns
Men, 2021Footnote 66 Cross-sectional Canadian Perspectives Survey Series (CPSS)	National (excluding the territories) 2020 Any charitable food intervention	Canadians living in the 10 provinces	NA	Six-item questionnaire adapted from the HFSSM Past 30 days	4410	Low
Parpouchi, 2016Footnote 67 Cross-sectional Data collected by the authors for the purpose of this study	Vancouver, British Columbia 2009–2011 Any charitable food intervention	Adults with a mental illness who were experiencing homelessness	NA	HFSSM adapted to populations experiencing homelessness Past 30 days	497	Low
Rizvi, 2021Footnote 58Footnote b Longitudinal design (cohort) Data collected by the authors for the purpose of this study	Ottawa, Ontario 2017–2019 Local food banks	People who accessed community food banks, after 6, 12 and 18 months	People who accessed community food banks, at baseline	HFSSM Past 12 months	401	Low
Tarasuk, 2009Footnote 68 Cross-sectional Data collected by the authors for the purpose of this study	Toronto, Ontario 2003 Any charitable food intervention	Youth aged 16–24 years without stable or secure housing	NA	Modified version of the HFSSM Past 30 days	261	Low
Tarasuk, 2020Footnote 69 Cross-sectional Canadian Household Panel Survey (CHPS) pilot	Saskatchewan, Ontario, Quebec and New Brunswick 2008 Local food banks	Households (except those on reserves, in religious and other communal colonies or in institutions, and members of the Canadian Forces)	NA	One question from the HFSSM Past 12 months	1593	Low
Vahabi, 2011Footnote 70 Cross-sectional Data collected by the authors for the purpose of this study	Toronto, Ontario 2008 Local food banks	Spanish- or Portuguese-speaking adults aged 20 years or older who had immigrated to Canada in the past 5 years from Central or South America	NA	One question from the HFSSM (translated into Spanish and Portuguese) that asked if participants ever had to eat less because they did not have enough money to buy food Past 12 months	70	Low

The impact of exposure to food-based interventions on HFI in Canada (KQ1)

Food-based interventions included food charity programs (n = 3),^{Footnote 57Footnote 58Footnote 59} food voucher programs (n = 2),^{Footnote 50Footnote 51} hunting and fishing programs specific to Indigenous communities (n = 2),^{Footnote 55Footnote 56} food box programs (n = 1),^{Footnote 52} gardening programs (n = 1)^{Footnote 53} and school food programs (n = 1).^{Footnote 54} Overall, low-certainty evidence suggests that participation in food box, gardening, hunting and fishing, school food and food charity programs may have little to no effect on HFI. Low- to moderate-certainty evidence suggests that food voucher programs may reduce food insecurity. The findings are summarized in Table 2 and described in detail in the following subsections.

Table 2. Summary of findings of studies examining the impact of exposure to food-based interventions on HFI in Canada, by program type (KQ1)

First author, publication year Study design (number of studies)	Population Period of data collection	Outcome (food insecurity level)	Number of participants exposed to intervention, n	Number of participants not exposed to intervention, n	Effect size	Direction of effect	Certainty of evidence
Food voucher programs
Aktary, 2024Footnote 50 Randomized controlled trial	Adult population living with low income in British Columbia 2019	Total HFI (marginal, moderate and severe)	143	142	Post intervention: OR = 0.21 (0.06–0.70); 16-week followup: OR = 0.29 (0.09–0.96)	Favours intervention (participation in Farmers’ Market Nutrition Coupon Program)	ModerateFootnote aFootnote b
Heasley, 2021Footnote 51 Single-arm pre–post design	Food-insecure households in Guelph, Ontario 2019–2020	Total HFI (marginal, moderate and severe)	60	NA	OR = 0.18 (0.07–0.50)	Favours intervention (participation in food prescription voucher program)	LowFootnote aFootnote bFootnote c
Food box programs
Miewald, 2012Footnote 52 Cohort (2-arm pre–post design)	Households with low incomes in British Columbia 2008	Moderate and severe	46	44	No effect	NA	LowFootnote aFootnote bFootnote c
Gardening programs
Sandha, 2021Footnote 53 Analytic cross-sectional	General population in Prince Edward Island 2013	Total HFI (marginal, moderate and severe)	104	175	No effect	NA	LowFootnote aFootnote bFootnote c
School food programs
Roustit, 2010Footnote 54 Analytic cross-sectional	Primary or secondary school-aged children and adolescents in the general population in Quebec 1999	Moderate and severe	678	1524	No effect	NA	LowFootnote aFootnote c
Hunting and fishing interventions specific to Indigenous communitiesFootnote d
Blanchet, 2021Footnote 55 Analytic cross-sectional	Syilx of the Okanagan Nation adults 2018	Total HFI (marginal, moderate and severe)	612	88	No effect	NA	LowFootnote aFootnote cFootnote e
Thompson, 2012Footnote 56 Analytic cross-sectional	Adults in remote communities in northern Manitoba 2008–2012	Moderate and severe
Food charity interventionsFootnote f
Loopstra, 2012Footnote 57 Cohort (2-arm pre–post design) 2005–2007; 2006–2008	Families living with low income in Toronto, Ontario	Moderate and severe	85	286	No effect	NA	LowFootnote aFootnote cFootnote e
Roncarolo, 2016Footnote 59 Longitudinal cohort (2 arms) 2011–2012	People who accessed food banks in Montréal, Quebec	Moderate and severe	372 (traditional food banks)	78 (alternative interventions, i.e. community kitchens, community gardens, buying groups)	OR = 0.44 (0.29–0.67) (for traditional food banks) No effect (for alternative food programs)	Favours intervention (traditional food bank)
Rizvi, 2021Footnote 58Footnote g Longitudinal cohort (single arm) 2017–2019	People who accessed community food banks in Ottawa, Ontario	Moderate and severe	401	NA	No effect	NA

Food voucher programs

In a randomized controlled trial, adults in households with low incomes received coupons valued at CAD21 each week for 10 to 15 weeks to purchase healthy foods from local farmers’ markets in British Columbia.^{Footnote 50} Aktary et al. reported that the odds of experiencing short-term HFI were 79% lower at the end of intervention (p = 0.01) and 71% lower at 16 weeks postintervention (p = 0.04) in the group receiving the coupon compared to the control group.^{Footnote 50}

In a pre- and postintervention single-arm study conducted in Guelph, Ontario, community health centre patients with a diagnosed cardiometabolic condition, a micronutrient deficiency or both and who were experiencing food insecurity were prescribed 12 weekly vouchers to use at community food markets.^{Footnote 51} Heasley et al. reported that the mean adult and child food insecurity scores decreased significantly from baseline to followup (p < 0.001 for adults; p = 0.01 for children).^{Footnote 51}

Food box programs

In a cohort study conducted in Vancouver, British Columbia, Miewald et al. assessed the effects of monthly distributions of food boxes of fruits and vegetables over an 8-month period to people in households with low incomes or in areas with poor food access or to older adults assessed.^{Footnote 52} There was no statistically significant change in HFI among those who enrolled in the food box program versus those who did not.^{Footnote 52}

Hunting and fishing interventions specific to Indigenous communities

Two analytic cross-sectional studies assessed interventions specific to Indigenous communities. Blanchet et al. found no association between the type of salmon eaten (locally caught salmon versus imported or no salmon) and HFI among Syilx of the Okanagan Nation adults in British Columbia.^{Footnote 55}

Thompson et al. assessed households’ access to hunted and fished foods in 14 remote Indigenous communities in northern Manitoba and found no association between access to these foods and HFI.^{Footnote 56} In the same study, the presence of a country food program (OR = 20.6; 95% CI: 2.4–176.1), road access to urban areas (OR = 7.6; 95% CI: 1.2–51.5) and access to a public transit system (OR = 3.9; 95% CI: 1.5–9.9) were all associated with lower rates of HFI, whereas living in a geographically compact community was not associated with lower rates of HFI.^{Footnote 56}

Gardening programs

In an analytic cross-sectional study conducted in Prince Edward Island, Sandha et al. compared the levels of HFI of mothers with children aged 0 to 6 years who accessed Family Resource Centre gardening programs with those of mothers with children the same age who did not access these programs. The authors reported that there was no relationship between HFI and access to the gardening progam.^{Footnote 53}

School food programs

In a cross-sectional study, Roustit et al. assessed elementary and secondary school food supplementation programs in Quebec and found no relationship between exposure to these programs and HFI.^{Footnote 54} Students in schools without (10.4%) and with (12.7%) these supplementation programs were food insecure (p = 0.22).^{Footnote 54}

Food charity interventions

Three studies assessing four food charity interventions were included.^{Footnote 57Footnote 58Footnote 59} The interventions included both traditional food bank models (where food bank clients receive a hamper of food regularly, usually once per month) and alternative community-based food charities (where participants can, for example, “shop” for groceries at a food bank or take part in community kitchen programs).

In a study that evaluated changes to HFI over a 9-month period, Roncarolo et al. found that severe HFI significantly decreased among participants in Montréal, Quebec, who accessed a traditional food bank (from 89.6% to 61.1%; OR = 0.27; 95% CI = 0.14–0.54 for severely food-insecure households), but did not significantly decrease among those who participated in alternative interventions (community kitchens, community gardens and buying groups).^{Footnote 59}

In a longitudinal study, Rizvi et al. found that most of the people accessing food banks in Ottawa, Ontario, were food insecure at baseline and remained food insecure at the 18-month followup, although there was a small downward trend in the proportion of people who were severely food insecure (from 38.5% to 24.6%; no measure of significance reported).^{Footnote 58}

In a study of families with low incomes accessing food banks in Toronto, Ontario, Loopstra and Tarasuk reported that 13.0% were no longer severely food insecure, 40.7% remained severely food insecure and 9.3% had become severely food insecure at followup 1 year post-baseline, for a net change of 3.7%.^{Footnote 57}

Rates of food charity utilization among food-insecure households (KQ2)

Eleven studies reported on rate (n = 11)^{Footnote 57Footnote 60Footnote 62Footnote 63Footnote 64Footnote 65Footnote 66Footnote 67Footnote 68Footnote 69Footnote 70} of utilization of food charity interventions among food-insecure households.

The rate of food-insecure households that used a food charity varied by severity of HFI and population group (Table 3; Figure 2). Those with severe HFI generally accessed food charities more often than those with moderate levels of HFI (Figure 2). The rate of food charity utilization was highest among those who were precariously housed or experiencing homelessness (Table 3; Supplementary Material F).

Table 3. Summary of findings of the rate of food charity utilization by food-insecure households according to HFI level, Canada (KQ2) (n = 11)

First author, publication year Study design	Period of data collection	Sample size who used a food charity, n	Sample size who did not use a food charity, n	Meta-analysis subtotal, %Footnote a	Interpretation of effect	GRADE
Marginal, moderate and severely food insecurity
Tarasuk, 2020Footnote 69 Cross-sectional	2008	276	1342	21	A small proportion of households that were food insecure used a food charity	ModerateFootnote bFootnote c
Loopstra, 2012Footnote 57 Cross-sectional	2005–2007, 2006–2008
Men, 2021Footnote 66 Cross-sectional	2020
Daly, 2023Footnote 60 Cross-sectional	2020–2021
MacBain, 2023Footnote 63 Cross-sectional	2021
Moderately food-insecure households
Loopstra, 2012Footnote 57 Cross-sectional	2005–2007, 2006–2008	63	237	21	A small proportion of households that were food insecure used a food charity	LowFootnote cFootnote dFootnote e
Kirkpatrick, 2009Footnote 62 Cross-sectional	2005–2007
Severely food-insecure households and child hunger
McIntyre, 20126Footnote 65 Repeated cross-sectional	1996–1997, 2006–2007	230	427	35	A small proportion of households that were food insecure used a food charity	ModerateFootnote cFootnote d
Loopstra, 2012Footnote 57 Cross-sectional	2005–2007, 2006–2008
Kirkpatrick, 2009Footnote 62 Cross-sectional	2005–2007
McIntyre, 2000Footnote 64 Cross-sectional	1994
Moderately and severely food-insecure household within immigrant populations
Vahabi, 2011Footnote 70 Cross-sectional	2008	34	5	90	A high proportion of this immigrant population used a food charity	LowFootnote bFootnote cFootnote e
Marginally, moderately and severely food-insecure populations experiencing homelessness
Parpouchi, 2016Footnote 67 Cross-sectional	2009–2011	431	101	83	A very high proportion of people experiencing homelessness used a food charity	HighFootnote c
Tarasuk, 2009Footnote 68 Cross-sectional	2003

Figure 2: Descriptive text

Food Insecurity Studies: Effect Sizes and Weights

Study	Effect Size (95% CI)	Weight (%)
Marginal, moderate, and severe food insecurity
Tarasuk (2019)	0.20 (0.15, 0.27)	19.36
Loopstra (2012)	0.28 (0.23, 0.34)	19.92
Men (2021)	0.07 (0.05, 0.10)	21.02
Daly (2023)	0.13 (0.10, 0.16)	20.82
MacBain (2023)	0.39 (0.32, 0.46)	18.88
Subtotal (I2 = 96.35%, p = 0.00)	0.21 (0.12, 0.30)	100.00
Moderate food insecurity
Loopstra (2012)	0.25 (0.18, 0.33)	34.70
Kirkpatrick (2009)	0.19 (0.14, 0.25)	65.30
Subtotal	0.21 (0.16, 0.25)	100.00
Severe food insecurity
McIntyre (2012) Cycle 2	0.34 (0.26, 0.44)	16.09
Loopstra (2012)	0.39 (0.31, 0.49)	16.44
McIntyre (2012) Cycle 7	0.32 (0.24, 0.42)	16.32
Kirkpatrick (2009)	0.40 (0.32, 0.49)	19.30
McIntyre (2000)	0.31 (0.25, 0.38)	31.85
Subtotal (I2 = 6.50%, p = 0.37)	0.35 (0.31, 0.39)	100.00
Moderate and severe food insecurity (immigrant population)
Vahabi (2011)	0.90 (0.76, 0.96)	100.00
Marginal, moderate, and severe food insecurity (homeless population)
Tarasuk (2009)	0.88 (0.84, 0.92)	63.81
Parpouchi (2016)	0.74 (0.69, 0.79)	36.19
Subtotal	0.83 (0.80, 0.86)	100.00

Abbreviations: CI, confidence interval; ES, effect size; HFI, household food insecurity; I², measure of heterogeneity.

Frequency of food charity utilization by food-insecure households (KQ2)

Six studies^{Footnote 57Footnote 58Footnote 60Footnote 62Footnote 66Footnote 68} assessed the frequency of utilization of food charity interventions by food-insecure individuals. Although there was no consistent measurement of the frequency of utilization of food charities or food banks, it was clear that the frequency of food bank use among food-insecure households was low (Table 4). Youth aged 16 to 24 years experiencing homelessness appear to use charitable meal programs (soup kitchens, outreach vans, drop-in centres, shelters) more frequently than households that are food insecure.

Table 4. Summary of findings of frequency of food charity utilization, Canada (KQ2) (n = 6)

First author, publication year Study design Population	Food insecurity levels	Size of food-insecure population, n	Percentage of the food-insecure population, %	Frequency	GRADE
Holmes, 2019Footnote 61 Cross-sectional Food-insecure households	Severe	43	70	Low (< 1 visits/month)	LowFootnote bFootnote cFootnote d
			68	Medium (1–2 visits/month)
			64	High (> 2 visits/month)
Kirkpatrick, 2009Footnote 62 Cross-sectional Food-insecure households	Moderate	182	18.7	Used a food bank at least once in the previous 12 months
	Severe	134	40.3
	Moderate	182	5	Used a food bank at least once in 10 or more of the previous 12 months
	Severe	134	6.7
Loopstra, 2012Footnote 57 Cross-sectional Food-insecure households	Marginal	47	10	Used a food bank at least once in the previous 12 months
	Moderate	118	27
	Severe	112	39
Men, 2021Footnote 66 Cross-sectional Food-insecure households	Total HFI (marginal + moderate + severe)	540	7.4	Received charitable food assistance at least once in the past 30 days
			4.3	Received charitable food assistance more than once in the past 30 days
Rizvi, 2021Footnote 58 Cross-sectionalFootnote a Food-insecure households	Marginal	226	24.3	Used a food bank once in the last 3 months
	Moderate	444	22.6
	Severe	408	22.0
	Marginal	226	22.6	Used a food bank twice in the last 3 months
	Moderate	444	17.7
	Severe	408	50.4
	Marginal	226	45.6	Used a food bank 3 or more times in the last 3 months
	Moderate	444	50.4
	Severe	408	47.2
Tarasuk, 2009Footnote 68 Cross-sectional Female youth experiencing homelessness	Moderate and severe	112	27	Used a charitable meal program 1–2 days in the last 7 days	ModerateFootnote e
			38	Used a charitable meal program 3–5 days in the last 7 days
			21	Used a charitable meal program 6–7 days in the last 7 days
Tarasuk, 2009Footnote 68 Cross-sectional Male youth experiencing homelessness		149	22	Used a charitable meal program 1–2 days in the last 7 days
			57	Used a charitable meal program 3–5 days in the last 7 days
			11	Used a charitable meal program 6–7 days in the last 7 days

Discussion

The objective of this systematic review was to synthesize the evidence on food-based interventions to mitigate HFI, recognizing the importance of this indicator for social policy and population health. This systematic review found limited evidence (of low certainty) for interventions assessing the effectiveness of food-based strategies on HFI, except for food voucher programs, for which there was some moderate-certainty evidence favouring intervention. Of note, no evidence was found for some strategies, for example, breakfast clubs or Meals on Wheels programs. Each category of intervention is discussed separately in the following subsections.

Food charity interventions

Low-certainty evidence suggests that food charity interventions may result in little to no difference in HFI. One possible explanation is that HFI is not a measure of access to food, but of absolute economic deprivation. A household that cannot afford to buy food likely cannot afford to buy necessary medications, pay the rent or pay for electricity, water and other essentials. Given that food is a basic human need, HFI is a clear marker of a household’s economic resources. Thus, HFI is not about an inability to access food, but rather a measure of economic insecurity. Receiving food charity may have a minimal effect on these economic resources or circumstances. This highlights the need for a more comprehensive economic response to the issue of HFI.

Another possible explanation for this lack of change in HFI is that food charities are limited in the amount of assistance they can provide to any one household. Food charities are neither a consistent nor a reliable coping strategy over the long term and are unlikely to change levels of HFI because households require food every day. Sustaining this level of access to food banks is likely not an economically viable option^{Footnote 71} as food charities lack the resources to meet clients’ food needs indefinitely.

This systematic review revealed that less than 40% of severely food-insecure households use food banks, and that most of the households that use them access them very infrequently. The low utilization rate may reflect food-insecure households’ recognition that food charities can only provide them with limited assistance. This may be due to inconvenient opening hours or because of a lack of the quantity, quality or type of food they need or prefer, among other reasons. This systematic review also found that increased severity of HFI was associated with higher utilization of food banks. This is in line with previous research that showed that the lack of utilization of these programs likely results from intrahousehold dynamics (such as household economics or coping strategies such as borrowing food and money from friends and family) that cause people to only access food charities when they reach the end of their capabilities.^{Footnote 57Footnote 65}

Recognition of the stigma associated with using food banks, among other reasons, has resulted in the development of alternative food-charity interventions,^{Footnote 59} but evidence that these programs reduce HFI is lacking. The findings of this systematic review also bring into question the role of food banks specifically and food charities more broadly in addressing food insecurity. While beyond the scope of this review, the persistence of food charities as a dominant response to HFI would ideally be researched by experts in the field of social science.

We found the proportion of people utilizing food charity programs to be much higher in populations experiencing homelessness, which may highlight the lack of other coping strategies available to them. It is important to note that even with high rates of utilization of food charities, this population continues to face severe and chronic food insecurity. An HFI-reduction strategy that is more comprehensive than solely focusing on provision of food is required.^{Footnote 67} This notion is corroborated by a recent systematic review that outlined the experience of HFI among people experiencing homelessness in high-income countries.^{Footnote 72} Easton et al. found that people experiencing homelessness are in a system that maintains food insecurity through oppression (i.e. structural inequities and institutionalism in finding adequate housing), an inability to fulfill basic needs, a lack of facilities for meal preparation and barriers to food assistance such as not having an address or means of identification.^{Footnote 72} Looking beyond food-based interventions is important to resolve the larger issue of the extreme deprivation of populations experiencing homelessness.

The main reasons for the low certainty of the evidence of the effect of food charities on HFI are the limitations in the study designs (inability to control for concurrent interventions outside the parameters of the study), the small study sample sizes and the significant dropout rates. Large longitudinal prospective cohort studies, such as the Pathways study in Quebec,^{Footnote 73} will help identify the long-term effects of food charity programs on food insecurity. Without a stronger evidence base for these types of interventions, their effect on HFI in Canada remains unconfirmed. The Office of the Auditor General (OAG) of Canada reached a similar conclusion in their 2021 review of Canada’s pandemic-related expenditures of the food assistance programs intended to mitigate HFI.^{Footnote 74} The review concluded that lack of data and performance measurement meant that the departments and agencies the OAG audited did not know whether the initiatives had achieved all their outcomes for reducing HFI.^{Footnote 74} While a lack of evidence does not indicate program ineffectiveness, a growing body of evidence shows that implementing more holistic public policy interventions that address income supports and root economics is effective in reducing HFI.^{Footnote 39} The evidence thus points towards a more holistic public policy approach that tackles the root economics of the problem in Canada.

Food box programs

The evidence in this systematic review suggests that food box programs that regularly supply fresh foods may result in no effect on HFI. The long-term impacts of such programs on HFI may be limited to the lifetime of the interventions, which require sustained and considerable economic input to regularly supply program participants with sufficient fresh food.^{Footnote 71} This is likely because these interventions, as with all the food-based interventions assessed in this systematic review, do not address the underlying causes of HFI as a marker of economic deprivation.

Food voucher programs

This systematic review found that food voucher programs may decrease food insecurity in households with low incomes for the duration of the intervention. This result aligns with those of two other systematic reviews that assessed interventions intended to reduce HFI in Canada and the United States.^{Footnote 75Footnote 76} These systematic reviews found that food vouchers were associated with a statistically significant decrease in HFI in both general and clinical populations.^{Footnote 75Footnote 76} Supplying households with vouchers may be a way to provide households with the economic resources to purchase culturally appropriate foods and thus increase their purchasing power.

Given the low number of studies that have assessed food voucher programs, more studies are needed to assess their effectiveness as well as the dose–response relationship between duration, frequency of exposure, long-term sustainability and dollar values of these interventions and food insecurity status.

Interventions specific to Indigenous communities

This systematic review found low-certainty evidence for the effect of hunting and fishing programs on HFI in Indigenous communities. Food insecurity in Indigenous populations ranges from 48% among on-reserve First Nations communities to 57% among Inuit in Nunavut.^{Footnote 77} Given these disproportionately high rates of HFI, it is critical to find culturally sensitive and workable solutions that valorize Indigenous ways of knowing. Previous reviews have explored programs and policy interventions related to Indigenous populations,^{Footnote 39Footnote 78} while Drysdale et al. synthesized the interventions intended to reduce HFI in remote regions in Canada as well as Australia and the United States.^{Footnote 79}

Gardening programs

Although many studies have assessed the potential benefits of gardening interventions, very few measured the effect of these programs on HFI using a validated scale. In a 2022 systematic review that assessed the impacts of community gardens on health in Canada, Japan, the Netherlands, South Africa, the United Kingdom and the United States, Hume et al. concluded that although HFI was not a directly measured outcome, community gardens likely do not affect HFI.^{Footnote 80} Future studies in this area should include a validated measure of HFI to assess the dimension of community garden and urban agriculture utilization.

School food programs

Although school food programs are often touted as an important mechanism for reducing hunger and HFI among youth,^{Footnote 81} the evidence for their effect on HFI is sparse as most studies on school food programs retrieved for this systematic review were screened out because they did not report on HFI. Two recent reviews by the Alberta Health Services and by Nova Scotia Health independently found that school food programs do not alleviate HFI.^{Footnote 82Footnote 83} The Nova Scotia public health authority concluded that school food programs “are not an appropriate or sustainable solution to HFI as they do not address its root causes—primarily, inadequate income.”^{Footnote 83,p.5} These programs may, however, be useful in achieving other outcomes such as improved dietary behaviour and critical food literacy skills (learning, culture and social norms).^{Footnote 84}

Limitations of the included studies and recommended future research

Limited data

Very few studies of the studies screened used a validated scale to assess the impact of a food-based intervention on HFI or related constructs of hunger, even though the stated purpose was to reduce HFI. Future studies as well as government programming that aims to increase the effectiveness of food-based interventions should include a measure of HFI in their evaluations.

Intervention variation

The design and administration of interventions varied considerably, making it difficult to measure their impact on HFI. Studies that assess the effectiveness of different implementations of similar interventions would be beneficial. Authors who report on food charity studies should carefully detail the intervention, the implementation of the interventions and any possible variations. The limited number of studies precluded examination of interjurisdictional variation. However, as there are increasingly large interprovincial differences in food-insecurity prevalence and severity,^{Footnote 1} it might be important for future studies to take into account the policy context of these food-based interventions.

Type of included studies

A major limitation of this systematic review was the high risk of bias of these largely observational studies and the difficulty in implementing an experimental trial to assess the efficacy of food-based interventions on HFI. The high risk of bias was due to the presence of confounding factors (5 of 9 studies) and to the high levels of missing data because of high dropout rates and loss to followup (4 of 9 studies). Assessing the effectiveness of these interventions (assessed through actual utilization rates) also remains a challenge.

Potential missing studies

This systematic review only synthesized the evidence that is publicly available and may have missed evidence on other types of food-based interventions that have not been published (e.g. breakfast clubs). In addition, it is also important to report null or negative studies in the literature, to give a fuller picture of the situation.

Conclusion

Food-based interventions date back more than 40 years,^{Footnote 85} yet we only retrieved 10 studies that assessed the effectiveness of these interventions on HFI or a hunger-related construct. When categorizing food-based interventions (i.e. school food programs, food charities), we found very few or in some cases no studies of program models that are particularly prominent in Canada, such as children’s breakfast clubs or Meals on Wheels.

The certainty in the evidence that food-based interventions have an effect on HFI is low, indicating that these interventions may not in fact affect HFI. Given the pervasiveness of HFI, and the fact that it is a marker of economic deprivation, it is unlikely that a food-based response will have much of an impact on overall HFI, which is primarily an economic problem. Emerging evidence suggests that more comprehensive public policy approaches are required to mitigate HFI.^{Footnote 39}

Acknowledgements

The authors wish to thank Bernard Choi, Janet Potvin and Genevieve Gariépy, of the Public Health Agency of Canada at the time of their involvement, for helping with the screening process; Mallory Drysdale and Prinon Rahman for helping to assess the risk of bias; and Kate Morissette for helping to undertake the GRADE process. We would like to thank the Health Canada Library for their assistance in designing and implementing the search strategy and the Public Health Agency of Canada Library for the updated search.

Funding

None. This research received no grants from any funding agency or from the commercial or not-for-profit sectors.

Conflicts of interest

The authors have no conflicts of interest to declare.

Authors’ contributions and statement

• LI: Conceptualization, investigation, formal analysis, methodology, project administration, supervision, validation, writing—original draft, writing—review and editing.
• CL: Formal analysis, investigation, data extraction, project administration, software, writing—original draft, writing—review and editing.
• TC: Formal analysis, investigation, writing—original draft, writing—review and editing.
• EV: Investigation, writing—original draft, writing—review and editing.
• AC: Investigation, writing—original draft, writing—review and editing.
• SK: Investigation, writing—review and editing.
• LM: Conceptualization, validation, supervision, writing—original draft, writing—review and editing.
• VT: Conceptualization, validation, supervision, writing—original draft, writing—review and editing.
• AJG: Conceptualization, methodology, supervision, writing—review and editing.

The content and views expressed in this article are those of the authors and do not necessarily reflect those of the Government of Canada.

Previous | Table of Contents | Next