Table 1 Results of the included interventions
Author, reference and country | Design | Study setting | Product and cue category | Information-based cue intervention | Comparison | Duration | Outcome measures | Results | Result | ||
|---|---|---|---|---|---|---|---|---|---|---|---|
Medium of cue | Content of cue | Location of cue | |||||||||
Cues concerning attention | |||||||||||
Achabal 1987 [25] USA | Randomised controlled trial | Supermarket | Fruit and vegetables | Colour signs | Pictures of produce (broccoli, cabbage, carrots, cauliflower, kiwi, tomatoes), plus purchase selection information | Signs placed in a holder strip six feet off the floor, directly about the produce | No signs in store Additional arm: Nutrition signs, which included the same information as the intervention arm, plus calorie content and a panel of key nutrients. | 4 weeks (plus 4 weeks pre-test, and 4 weeks post-test) | Weekly sales data | No variation in sales were observed by the presence of the sign, with no significant effect of the sign (F = 0.684, p = < 0.505). Additional analysis comparing matched store data, found no differences in the number of times the sales were greater in the intervention stores than the control stores. | – |
Payne 2015 [29] USA | Pilot study | Supermarket | Fruit and vegetables | Brightly coloured placards | Graphics of popular fruit and vegetables, plus positive and social norm messages | Placards were placed on the inside front and outside front of the grocery cart | No intervention | 14 days (plus 57 days matched pre-intervention period) | Weekly sales data | There was an increase in average spending per day per person on produce of 16% (< 0.01) in the intervention store and an increase of 4% (p = n.s.) in the control store. | + |
Payne 2016 [30] USA | Pilot study | Supermarket | Fruit and vegetables | Large green arrows | Green arrows included graphics of fruits and vegetables, an emoticon to facilitate social approval and text such as “follow green arrow for health” | Arrows were placed around the store perimeter, on the floor in highly visible places | No intervention | 14 days (plus 14 days pre-intervention period) | Weekly sales data | There was an increase on produce spending per day per person of 8% (p = 0.01). The authors do not state change in spending in the control store. | + |
Hanks 2016 [26] USA | Randomised controlled trial | School cafeterias | Fruit and vegetables | Brightly coloured vinyl banners and/or TV segments | Branded media vegetable characters with human characteristics and super human strength. | Banners were placed around the salad bar and on top of the bar. TVs were placed on small tables near the salad bar Small printed vegetables were also placed on the floor to direct children to the salad bar | No Intervention | 4 weeks (plus 2 weeks pre-intervention) | Food preperation records and count data of children serving themselves | There was an increase from 60 to 185 daily vegetable servings in the schools with the combined vinyl banner and TV segments intervention. There was a reported increase in servings for the individual intervention arms, but data were not significant. Count data showed a significant increase in frequency of children taking vegetables from the salad bar; 12.6 to 24% with vinyl banners, 10.2% to 34.6% with vinyl banners and TV segments. There was no significant change with just the TV segments (13.8% to 18.9%) or in the control schools. | +/− |
Morizet 2012 [36] France | Experimental design (randomised by temporal order and condition) | School cafeterias | Fruit and vegetables | A white paper easel with text information | A basic label, “new carrot/broccoli recipe” and a model label: “new carrot/broccoli recipe, special mix for superheroes” | Labels were presented at the front of the gastronorm tray where vegetables were located. | No label | Two different days (one per vegetable option) | Choice frequency of vegetables | Children selected the familiar carrot or broccoli dish significantly more often when no label was present. For the carrot dish there was a significant difference in the number of children who selected the novel dish with the basic label (p = 0.012), and the model label (p = 0.002) as compared to no label condition. The patterns were similar for broccoli, but not statistically significant. | +/− |
Folta 2006 [31] USA | Randomised controlled trial | School cafeterias | Fruit and vegetables | Audio announcement | Messages promoted beans and featured a magical superhero “bean man” | Audio messages were played during the school morning announcements | The same bean dishes were introduced to the school lunch menus. The children received their normal morning announcements, without the addition of any health messages | Approximately 3 months | Choice of beans | Overall data showed no difference in frequency of bean selection between intervention and control schools. | – |
Cues providing educational information about the product properties | |||||||||||
Steenhuis 2004 [32] Netherlands | Randomised pre-test, post-test experimental design | Worksite cafeterias | Low fat options | Signs (plus brochures, table tents and optional self-help manuals, newsletters and badges (the additional factors were considered as part of the educational programme)). | The signs consisted of the program logo, the name of the food item, and an indication that it was a healthy choice | Signs were placed in front of targeted food products | No intervention Additional arms: Educational programme, Food supply plus education. | One month (with the option to extend to six months) | Sales proportion of low-fat products per product group A self-administered food frequency questionnaire to measure fat, fruit and vegetable intake. | At one and six months there were no significant changes in reported fat, fruit or vegetable consumption. Sales data showed a significant increase in the proportion of low fat desserts sold in the intervention sites as compared to control sites. No other differences in sales were observed. | +/− |
Vyth 2011 [33] Netherlands | Cluster randomised controlled trial | Worksite cafeterias | Low fat options plus fruit | Placards with a logo. Menus explaining the logo were available. | The logo consisted of a tick, indicating a healthy choice. The logo was part of “The choices intervention” which assigns its logo to foods which meet a determined sodium, added sugar, saturated fat, trans fats, fibre and energy. | The placards were placed next to freshly prepared “choices” sandwiches and soups, and the fruit basket. | No labelling or any other communication regarding the intervention | 3 weeks (plus 3 weeks pre and 3 weeks post intervention periods) | Daily sales data | There was a significant difference in fruit sales between the intervention and control sites during the intervention period. No other differences in sales of targeted products were observed. | +/− |
McClain 2013 [28]USA | Cluster randomised controlled trial with repeated cross sectional assessment | University dining halls | Low fat options | Placards, posters and table tents | Colourful photographs, and messages such as “brain food” | Placards were placed at food stations and on dining tables. Posters were placed around the dining halls. | No intervention | 4 weeks (plus 7 days pre and post intervention) | Harvard food frequency questionnaire data | Students in the control halls consumed significantly more junk food and high fat meat products from baseline to intervention relative to students in the intervention halls. No other differences were observed. | +/− |
Lee-Kwan 2015 [27] USA | Quasi-experimental study | Carryout restaurants | Healthy entrees, sides and beverages | Menu boards and posters Paper menus were also available, which were replicates of the menu boards | Digital colour photographs of selected healthier options plus a green leaf logo to indicate the healthier options | Menu boards were placed next to healthier items. Posters were placed next to the menu boards. | No intervention | 8 weeks (plus 4 week pre-intervention, and a further 16 week intervention of additional interventions) | Weekly sales data | In the intervention group the relative odds of healthy entrees were greater during the intervention period to baseline (OR 1.16, 95%CI: 1.08: 1.26). The changes were not significantly different to control sites.The relative odds of healthy sides and beverages was not significantly different over time in the intervention group. were no significant differences in units sold from baseline in the intervention group or compared to control. | – |
Implicit cues | |||||||||||
Stockli 2016 [35] Switzerland | One-factorial within subjects experimental design study | Vending machines in University and Public Health office buildings | Both healthy and unhealthy snacks | Posters | Study 1: A nature poster, (showing grassland, tress and a blue sky with clouds) an activity poster (showing running legs in sports shoes and asphalt in the background) and a fun fair poster (showed two carousels with a summery blue sky in the background) Study 2: A poster with the Giacometi sculptures, (elongated figures) an activity poster and a fun fair poster (as above) | Posters were placed above vending machines | No posters | 4 weeks | Daily sales data | There was a significant association between poster exposure and snack choice in both studies: Study 1: The percentage of healthy snacks selected was, 34% with the nature poster, 28% with the activity poster, 22% with the control condition and 18% with the funfair poster. Study 2: The percentage of healthy snacks selected was 58% with the Giacometti poster, 44% with the activity poster, 29% with the fun fair poster and 21% with the control condition | + |
Engles 2011 [34] Netherlands | Experimental design study (evenings were randomised between arms) | Bars | Alcohol | Audio: Musical playlists | A playlist of 90 songs (with a 5.5 h duration) with contextual references to alcohol. Songs contained lyrics that referred to alcohol in them, for example “red, red, wine”. | The music was played on a fixed day of the week, at a fixed starting time and were played in a random order by the bartenders | A playlist of songs created by the same artists included in the intervention playlist, but the songs did not refer to alcohol. To ensure a good match, attention was played to timeframe and period of the album, tempo and energetic content of the songs. | The three included bars collected data on 18, (bar A) 12, (bar B) and 16 evenings (bar c). | Sales of alcoholic drinks corresponding to the two hours the music was played | Turnover was significantly higher on nights when music had reference to alcohol played compared to nights when the control playlist was played (p = < 0.05) | + |