Table 4 Synthesis of data from included studies, categorised by the type of intervention/policy implemented
Study reference | Setting | Intervention dose/policy details | Nutrition education provided also | Process, Impact, Outcome [measures] | Economic measure? | Evaluation findings | Identified barriers | Identified enablers | Specific recommendations (of study authors) | Promising | Comments/Notes |
|---|---|---|---|---|---|---|---|---|---|---|---|
1. Price discount on healthy food sold via retail stores, mobile grocery van, or health service (n = 13 publications, n = 7 studies) | |||||||||||
1a Provided by community retail store (n = 10 publications, n = 5 studies) | |||||||||||
Williams, 2021 [25] | USA: Chickasaw and Choctaw nations, Oklahoma | “THRIVE”: Cluster-controlled trial (in 2 Nations; in each 2 stores received intervention, 2 were controls; 9 months in Nation A; 12 months in Nation B); offered healthy ready-to-eat meals and snacks (high in F&V) at or below prices of competing foods. | N | Weekly sales data (first 6 months of intervention) | N | F&V basket sales higher in intervention stores than controls (significantly higher in one nation, but not the other); total sales remained steady. | A | ABCE | Regular updated promotions needed | Y | |
Blue Bird Jernigan, 2019 [26] | USA: Chickasaw and Choctaw nations, Oklahoma | “THRIVE”: Cluster-controlled trial (n = 1204 in 2 Nations; 9 months in Nation A; 12 months in Nation B); offered ready-to-eat healthy meals and snacks (high in F&V) at or below prices of competing foods. | N | Self-reported dietary intake; recall of promotions and reported purchasing | N | Increased purchasing of fruit, vegetables and other healthy foods; however, F&V intake did not increase in either Nation. | ABE | Y | Discounted ready-to-eat healthy meals/snacks of high relevance to APY communities | ||
Brimblecombe, 2018 [27] | Australia; Northern Territory | “SHOP@RIC” trial: Price discount (20%) on fresh and frozen fruit and vegetables, bottled water and artificially sweetened soft drinks +/− consumer education. Stepped-wedge RCT in 20 very remote communities; n = 148 adults who identified as primary shopper for household. 49 week baseline data-collection, then 24 week intervention, 24 week post intervention follow-up. | Y +/− | 148 adults. Self-reported intake; mediators and moderators | N | Modified perceived affordability of F&V but no substantial consumer behaviour change. | B | Long-term government investment and commitment needed to address underlying constraints, including monetary incentives; need to enhance self-efficacy to cook and try new vegetables. | Maybe, if identified barriers addressed | Discount was not strong enough to overcome constraints in those most disadvantaged. No consideration of social response bias. | |
Magnus, 2018 [28] | Australia: Northern Territory | Estimated cost-effectiveness of 20% price discount on healthy food and beverages (+/− nutrition education); analysis alongside the SHOP@RIC trial (above) in 20 remote communities. | Y +/− | Food sales data; published mortality, disease and RF data; costs and cost-offsets | Y | 20% discount with or without consumer education cost more money without leading to health gain, i.e. it offered poor value for money | B | Maybe, if identified limitations addressed | |||
Brimblecombe, 2017 [29] | Australia: Northern Territory | Price discount (20%) on fresh and frozen fruit and vegetables, bottled water and artificially sweetened soft drinks +/− consumer education (SHOP@RIC trial); stepped-wedge RCT in 20 very remote communities; 49 week baseline data-collection, then 24 week intervention, 24 week post intervention follow-up. | Y +/− | Weekly store sales data | N | Complete implementation of discount promotion and consumer education not achieved in all stores. Positive shift in purchases of F&V and bottled water but not diet drinks. Price discount alone was associated with a 12.7% increase in purchases in grams of fruit and vegetables combined (primary outcome); and a 19.8% increase after discount had ceased (after vs before). Purchases of water and diet and regular soft drinks also increased post-intervention. | BJM | B | Price discount on healthy foods may need to be supported by price increase of unhealthy foods; greater promotion of F&V; cooking and food budgeting programs; improved household food preparation and storage infrastructure; and education to discourage unhealthy choices. | Maybe, if identified limitations addressed | Possible unintended consequences, with cost savings lead to increased consumption of unhealthy products |
Magnus, 2016 [30] | Australia: Northern Territory | Modelling estimated cost effectiveness of six price discount strategies on fruit, vegetables, diet drinks and water | Y +/− | Food sales data; published price elasticity data; Aboriginal population health status indicators | Y | All fiscal strategies modelled had positive impact on diet quality; 5/6 estimated as cost effective (below $50,000/DALY threshold) | Price discounts appear to be potentially cost-effective | Maybe, if similar results after implementation in the real world | Need to consider difficulties around implementation and impact on magnitude | ||
Ferguson, 2017 [31] | Australia: 18 remote communities in the Northern Territory and Western Australia | Four price discount strategies: reduce markup on grocery products; fresh F&V point-of-sale scales; fresh F&V at landed cost; diet soft-drink discount. Implemented since 2010. N = 54 informants | N | Retrospective evaluation of a natural experiment. | N | Discounts applied as intended; varying levels of promotional materials. No effect of price discount (10%). Non-significant reduction of diet soft drinks | B | ABC | Greater dose, duration and promotion of discounts; co-design including customers, store owners and staff; monitoring of implementation | Maybe, if identified limitations addressed | |
Lee, 2016 [19] | Australia: remote South Australia (APY Lands) | Store nutrition policies and interventions to address healthy food availability, accessibility and affordability; time series of cross-sectional studies. N = 7 communities | N | Food price and availability and sales data | N | Decreased price of F&V. Since 1986, cross-subsidisation increased availability and affordability of healthy foods, especially F&V. Increased supply and intake of discretionary foods, too, | L | ACE | Sustained community effort needed to improve availability and affordability of healthy food. | Maybe, if identified limitations addressed | Mai Wiru store policy needs update/revision. Suggested more frequent monitoring of stores and increased engagement of all community members in results |
Blakely, 2011 [32] | New Zealand | “SHOP” RCT: Assessment to determine if effects of price discounts of 12.5% on healthy foods varied by ethnicity, income or educational qualifications. | Y +/− | Purchasing data (barcode scanner) | N | Price discounts had a weaker and null effect among Māori than among European New Zealanders | BI | Better targeting could be warranted | Maybe, if identified limitations addressed | ||
Ni Mhurchu, 2010 [33] | New Zealand | “SHOP” RCT: Price discounts of 12.5% on healthy foods applied automatically at checkout +/− tailored nutrition education; information on price discounts mailed to participants. 12-week baseline; 24 week intervention; 24 week follow-up; n = 1104 adult shoppers at 8 supermarkets (23% Māori). | Y +/− | Pre- and post- data. Purchasing data (barcode scanner) | N | Healthy food purchases improved slightly with price discounts but no significant improvement in nutrient analysis; price discounts had sustained but small effect on F&V purchases; education had no effect on food purchases. | BHI | In-store signage and ‘shelf-talkers’ would be better promotional tools than list of products eligible for discount. | Maybe, if identified limitations addressed | ||
1b Provided via mobile grocery van (n- = 1 paper, n = 1 study) | |||||||||||
Cueva, 2018 [34] | USA (an unnamed Native American community) | Mobile grocery (Mo Gro) offering subsidised healthy food twice a week; 3-month evaluation; n = 92 First Nations households (randomised selection of 20% of households). | Y | Self-reported food purchasing, consumption and perceptions; FS questionnaire | N | Process: Served avg. of 71 customers per visit (twice weekly); Impact: 75% reported change in food purchases, 68% changed dietary patterns; Outcomes: FI declined from 57 to 43% | ABE | Need to include traditional foods | Y | Developed in response to community need assessment | |
1c provided via health service (n = 2 papers; n = 1 study) | |||||||||||
Black, 2014 [35] | Australia: rural NSW communities | Subsidised F&V ($5 for box of $40 value, or $60 if 5 or more children); also in one community vouchers redeemable at F&V shop; 55 low-income families, 121 participating children; duration of program highly variable (several features mimicked a natural experiment). | N | Pre- and post- of varied duration. 24-hour dietary recall and biomarkers | N | 70% of families collected 75% or more of available F&V boxes; improved biomarkers in children but not changes in self-reported intake of F&V | D | Controlled study of subsidised healthy foods is warranted | Maybe, if identified limitations addressed | Occurred in real world setting so difficult to account for all potential confounders. | |
Black, 2013 [36] | Australia: rural NSW communities | Subsidised F&V, as above | N | Health service use; biomarkers and child height and weight | N | Decreased presentations to health clinic | D | Maybe, if identified limitations addressed | |||
2. Subsidies to retail stores, suppliers and producers (n = 8, studies = 2) | |||||||||||
2a Subsidies direct to retail stores (n = 5 papers/reports, n = 1 study) | |||||||||||
CIRNAC, 2020 [37] | Canada: Remote, northern communities | Nutrition North Canada: government subsidy provided directly to (contracted) retailers, suppliers and registered country food processors to reduce costs of nutritious perishable foods for residents of remote northern communities (subsidy rates vary depending on type of transportation used, location of the community, and category of eligible food and non-food items); horizontal evaluation. Program commenced in 2011. | N | Qual: interviews, document review, analysis of program data | N | Low population awareness of program and understanding of how subsidy works; increased access to nutritious perishable food at subsidised rate but subsidy has minimal impact on affordability, especially for people on low income (welfare or minimum wage) and seniors; recommended diet still unaffordable (typical household of four able to afford less than half contents of recommended food basket); some staple items not subsidised e.g. flour and lard; minimal savings often perceived negatively by community. | BCIM | E | Need to work better with communities; increase magnitude/dose; develop indicators that are relevant to CPI; include subsidies for local food production; improve promotion of program. | Maybe, if identified limitations addressed | Subsidies at point-of-sale or directly to vulnerable consumers could be more effective |
Naylor, 2020 [38] | Canada: Remote northern communities | Nutrition North Canada; econometric assessment of pass-through rate through food supply system ($ input v benefit) | N | Published food price data | Y | Subsidy appears to reduce food prices i.e. dollar increase in subsidy is associated with a dollar reduction in final food price; higher pass-through rates in larger communities due to economies of scale and density in air transportation; subsidised food items relatively cheaper than in Ottawa. | E | Increase amount of subsidy provided, and target subsidy to specific food items desired by FI households | Maybe, if identified limitations addressed | Number of eligible communities varied | |
St-Germain, 2019 [39] | Canada: Remote northern communities | Nutrition North Canada; interrupted time series regression analysis (n = 3250 households in 10 communities) | N | Self-reported food insecurity | N | Prevalence of household food insecurity increased from 33.1% in 2010 (year before launch), to 39.4% in 2011 (year of launch) and 46.6% in 2014 (year after full implementation) | C | E | Subsidy on nutritious food only could increase food security for the most economically vulnerable households. More research is needed to investigate food access inequality. | Maybe, if identified limitations addressed | |
Galloway, 2017 [40] | Canada: Remote northern communities | Nutrition North Canada; program evaluation | N | Program data and evaluations; sales and price reports | N | Persistent inequities in food pricing between communities and food items (in absence of price caps) and population groups (e.g. some individuals order directly); subsidies of insufficient magnitude to address inequalities. | BFGJK | DE | Need for increased retailer accountability and regulatory framework. | Maybe, if identified limitations addressed | Retail subsidy not effective where there is not a competitive marketplace |
Auditor General of Canada, 2014 [41] | Canada: Remote northern communities | Nutrition North Canada; program audit. | N | Audit.Qual: stakeholder interviews; policy/document analysis | N | Weight of items subsidised increased by about 25% but did not improve FS; lack of transparency in program management | F | E | Need for greater compliance monitoring and requirement for retailers to provide information needed to assess whether they are passing on full subsidy to consumers | Maybe, if identified limitations addressed | |
2b Subsidies for transport (n = 3 papers/reports, n = 1 study) | |||||||||||
INAAC, 2009b [42] | Canada: Remote northern communities | Food Mail Program (FMP) - subsidised cost of transporting nutritious perishable food to remote communities; program evaluation. | N | Statistical and econometric analyses; Qual – panels, interviews | Y | Reduced prices of food, but still unaffordable for many households. Increased subsidy rates for priority perishable foods (e.g. vegetables, fruit, eggs) in three pilot project communities resulted in significantly higher per capita volume shipments and presumably consumption of perishable items | BCGJ | Need increased transparency and accountability; to engage with Aboriginal organisations to help ensure items are culturally appropriate; support local, sustainable, complementary initiatives e.g. community freezers; increase subsides on ‘staples’ such as bread and milk to increase affordability. | Maybe, if identified limitations addressed | Concern about degree to which subsidies are passed on to communities. Program ran from the 1960s to 2011 in different formats. | |
INAC, 2009a [43] | Canada: Remote northern communities | FMP (as above); program review (separate process to above). | N | Program costs; food prices | Y | FMP successful in lowering the price of food in participating communities; further reductions in shipping rates for “priority perishable foods” (in pilot project) resulted in price reductions of about 15 to 20%. | Need increased transparency and accountability of retailers; investigate redesign of program as a retail subsidy delivering benefits to consumers at point of purchase. | Maybe, if identified limitations addressed | |||
Dargo, 2008 [44] | Canada: Remote northern communities | FMP (as above); independent review. | N | Program data; discussion with stakeholders | N | Poor program evaluation processes; program burdened with “design, logistical, administrative, accountability, negative resident perception and application issues”; Low levels of awareness; many residents concerned subsidy was not being passed on. | FHI | Replace with new program providing better subsidy on core basic items, developed in partnership with Inuit organisations | Maybe, if identified limitations addressed | ||
3. Provision of healthy food (n = 8 reports/papers, n = 7 studies) | |||||||||||
3a Healthy food provided at no cost (n = 5 reports/papers, n = 4 studies) | |||||||||||
Ahmed, 2020 [45] | USA: Rural; Flathead reservation, Montana | Pilot study of “Eat Fresh”; weekly boxes of recommended servings of fresh fruits and vegetables provided for six weeks; n = 19 low-income Native American adults | Y | Pre- and post- intervention. Qual: diet habits and health perception; Quant: diet intake, biomarkers | N | Reported increase in F&V variety; trend of improved diet quality; significant HEI increase post-intervention; BMI and blood pressure increased | CDM | A | Need for multi-strategy, holistic dietary interventions and focus on whole diet; should measure multiple indicators, both qualitative and quantitative; important to collaborate with Community Advisory Board for intervention design and feedback. | N | Small study; objective outcomes worsened |
Briefel, 2021 [also at 3b] [46] | USA: Chickasaw Nation | Monthly food box (shelf-stable nutritious foods) and $15 voucher for F&V for each eligible child; cluster RCT in 40 school districts in Chickasaw Nation over 25 months (n = 2859, 14% Native American) | Food security (survey), food expenditure | Y | Participation rate 61% (boxes had to be ordered online or by phone); did not improve child FS; adult FS improved initially but not at follow up; modest decline in out-of-pocket food expenditure | F | Y | Several confounders including improved economic circumstances of the population and participation in other nutrition assistance programs | |||
Pindus, 2019 [47] | USA: Rural/remote and urban reservations in the Klamath River Basin | Review of the Food Distribution Program on Indian Reservations (FDPIR), providing monthly food packages (perishable and non-perishable) to low-income households living on Indian reservations, on tribal lands, and other designated areas (n = 1053 households). Duration not clear. | Y (variable) | Participation; Qual: FS measures, discussion groups | N | FDPIR was only source of food for 38% of participants; 34% of households had low FS and 22% continued to have very low FS. Food package was inadequate in rural areas; not meeting community needs. | EFH | Establish partnerships, and expand supplemental assistance and/or food access and flexibility | N | Government program - no mention of co-design | |
Mucioki, 2018 [48] | USA: Rural/remote and urban reservations in the Klamath River Basin | Case study. FDPIR (as described above). n = 151 using FDPIR, 275 using other food assistance, 242 not using food assistance | Y (variable) | Qual: perceptions and operations via interviews and focus groups; FS measures | N | Packages do not meet international standards for quality, access, availability, nutrition and cultural appropriateness. Participants desire more fresh fruit, vegetables and traditional foods. Food boxes are essential source of food, but fail to alleviate FI. | EF | Increase amount and frequency of delivery of F&V; support traditional food acquisition; increase eligibility | N | ||
Ichumar, 2018 [49] | Australia: Rural Western Australia | School breakfast program; in two schools with high Aboriginal student populations; duration not clear | Y | Stakeholder interviews, observation, document review | N | Food provided passively to children, not necessarily nutritious; little evidence of health education | F | Schools should explore arrangements with local growers/shop owners with respect to support for the SBP. | N | ||
3b Healthy food vouchers (n = 4 reports/papers, n = 4 studies) | |||||||||||
Briefel, 2021 [also at 3a] [46] | USA | See above ($15 voucher for F&V for each eligible child provided with monthly food box) | As reported above | ||||||||
Jones, 2020 [50] | USA: Navajo Nation | Fruit and vegetable prescription (FVRx): vouchers redeemable for fruit, vegetables and healthy traditional foods from participating retailers; US$1 per household member per day, with a maximum value of $5/day; 243 Navajo children. Ran May 2015 to Sept 2018. | Y | F&V consumption and food security, child height and weight | N | Process: 65% of children retained in program > 6 months; Outcomes: household FS increased from 18 to 35% | CE | – | Y | Multiple confounders - difficult to isolate or attribute outcomes | |
McLaury, 2016 [51] | USA: Rural reservations in Washington State | Cash value vouchers (CVV) for F&V added to WIC food packages (monthly values of $6 for children and $10 for pregnant, breastfeeding, and postpartum women). Duration not clear. | N | Program data (vouchers issued and redeemed) | N | No significant outcomes in American Indian population | F | More research needed to determine causes of low voucher redemption, including socioeconomic and cultural barriers to CVV redemption on reservations. | N | Authors presume barriers such as embarrassment, unfamiliarity with F&V, cost of produce, misunderstandings about how to use vouchers | |
Brown, 2019 [52] | Australia: Remote communities, Cape York, Queensland | F&V voucher; 32 weeks over two phases (1: $10 voucher for minimum $20 spend on F&V; 2: $10 voucher for minimum $15 spend). Pre and post; impact measure was F&V purchasing. | Y (ad hoc) | Pre- and post- Qual interviews; store sales data | N | Trend of reduced F&V sales and overall food and drink; 7% reduction in fruit sales. Average voucher redemption rate was 29%. Highest use of vouchers (44%) in week when project staff promoted program/cooking demonstrations in store. | GM | BC | Target vouchers to women and children; use store loyalty cards instead of paper vouchers; increase flexibility of redemption (greater variety of healthy foods); increase promotion; need more support from store staff. | Maybe, if identified limitations addressed | Not clear how controlled for community numbers/store population. Precursor to study by Ferguson et al. (2017). |
4. Provision of greater income to community members (n = 1 paper, n = 1 study) | |||||||||||
Gordon, 2017 [53] | USA: 14 sites including two tribal nations (Cherokee and Chickasaw) | Piloted “Summer Electronic Benefit Transfers for Children” (SEBTC), cash benefit of $60/child/summer month; n = 42,000 households in 14 sites, 2 tribal nations, duration = one summer period | N | Food frequency questionnaire and food security scale | N | Significantly reduced rates of very low FS (one-third lower for households receiving SEBTC). Children in households receiving benefits consumed more healthy foods including F&V. Impacts in WIC sites were at least twice as large as those in SNAP sites (where benefits could be used to purchase SSBs; WIC-model restricted to healthy food). | E | Model deserves consideration; providing benefits in summer meets gap for children who receive school-based nutrition programs during school terms. | Y | ||
5. Government strategy/policy (not otherwise described above) (n = 5 reports/papers; n = 4 studies) | |||||||||||
George, 2021 [54] | USA: Navajo Nation and bordering towns | Tax Waiver (healthy foods): Hypothecated tax on unhealthy foods in Navajo Nation (2%; revenue directed to local community wellness projects) and waiver of sales tax on healthy items (including water, fresh fruits and vegetables and nuts). | N | Store surveys in 2013 and 2019 (matched sample of 71 stores: 51 in Navajo Nation, 20 in border towns) | N | Since 2013 (after adjusting for inflation), average cost per item of fresh fruit decreased by 13% in Navajo stores and increased in border stores, resulting in comparable prices in Navajo and border stores in 2019. Pricing trends among vegetables and other healthy foods were inconsistent. | B | Maybe, if identified limitations addressed | Only measured changes in pricing and food availability (and in-store promotion). | ||
ANAO, 2014 [55] | Australia: National, focus on remote communities | National Food Security Strategy: Australian Government implementation of food security initiatives under ‘Close the Gap’ for remote Indigenous communities including targeting affordability. | Y | Performance audit | N | Pilot sites identified. However, there was no evidence that initiatives to decrease the prices of healthy foods had been implemented. | G | Strategy required a funded action plan, and implementation. | Maybe, if identified limitations addressed | Throughout Australia “basic, healthy foods” do not incur 10% GST. Australian Government currently developing another Remote Indigenous community food supply strategy. | |
Bray, 2014 [56] | Australia: Northern Territory | Income Management: Northern Territory (NT) New Income Management policy, operationalised through use of EFTPOS card (“BasicsCard”) able to be used only in approved stores and not to purchase prohibited goods (e.g. alcohol) or withdraw cash. Intervention commenced in 2007. Policy evaluation. | N | Longitudinal survey; store transactions | N | No evidence of changes in spending patterns, including food and alcohol sales, other than a slight possible improvement in the incidence of running out of money for food by those on Voluntary Income Management, but no change for those on compulsory income management. | HKM | Program should be voluntary | N | ||
Brimblecombe, 2010 [27] | Australia: 10 remote communities in the Northern Territory with stores managed by Arnhem Land Progress Association. | Income management policy: 50% of income support and family assistance payments (and 100% of lump sum payments) to Indigenous people living in remote areas of the NT to be used only for items considered essential by the government, such as food, clothes, rent, etc. Analysis of sales data from 10 community stores over 3 years (October 2006 to September 2009). | N | Interrupted time series analysis: store sales before (18-month period) and after introduction (4-6 month period) of income management, 3 months coinciding with a government stimulus payment, and remaining income-management period | N | Income management had no effect on fruit and vegetable sales or turnover; significant increase in sales (total store, total food and beverage, fruit and vegetable and soft drink) during period of government stimulus payment. | HKM | Unintended consequences with increased sales of sugar sweetened beverages. | N | Limited products studied | |
Thompson, 2012 [57] | Canada: Rural Northern Manitoba | Country food programs that support people living off the land to feed the local community; participatory process over four years; analysis of 7 communities with “best practice” in food programming and 7 with limited uptake. Random selection of 553 households. Duration of program unclear. | N | Food costs and food security surveys (553 households in 14 rural communities) | N | Country food programs were related to better food security; food programs that enable sharing of traditional foods improved food security more than other variables, such as access to stores. | A | Improving food access requires community control over funding and decision-making without undue restrictions on country foods. | Y, where sustainable traditional foods available | May not apply in all First Nations communities, especially where traditional food systems are under threat by climate change, population pressure etc. | |