Table 1 Included studies characteristics

Walsh and Kiviniemi, 2014 [17]

RCT

117

Laboratory

USA

28.3 ± 0.0343

Subjects’ affective associations with fruit and vegetables were experimentally altered using an implicit priming paradigm. They were then asked to choose between fruit or a granola bar for a snack.

Privitera and Zuraikat, 2014 [18]

RCT

56

Laboratory

USA

79.3 ± 17.7 [outlier]

Snacks were placed closer or farther away from subjects and their snack choices recorded and compared between control and experimental groups.

Gittelsohn et al., 2013 [19]

Cross-sectional (pre-post comparison)

145

Market

USA (Navajo)

7.84 ± 12.7

Availability of healthy options was increased over a 14-month period. Outcomes were measured in change of purchases of healthy items.

Kiesel and Villas-Boas, 2013 [20]

Cross-sectional (pre-post comparison)

4000

Market

USA

12.8 ± 17.2

Nutritional labelling was implemented over a four-week period. Outcomes were measured by purchases of healthy items.

van Kleef et al., 2012 [21]

RCT

67

Laboratory

USA

41.4 ± 11.8

Size of bowl used to serve snacks was varied and consumption measured and compared between control and experimental group.

Marchiori et al., 2012 [22]

RCT

88

Laboratory

Belgium

129 ± 0.703 [outlier]

Participants were given snacks in differently sized containers and their consumption measured.

Dumanovsky et al., 2011 [23]

Cross-sectional (pre-post comparison)

7311

Restaurant (fast-food)

USA

2.20 ± 5.02

Measurements of average energy content of purchases were made before and 21 months after addition of calorie labels to menu.

Finkelstein et al., 2011 [24]

Cross-sectional (pre-post comparison)

9823

Restaurant (fast-food)

USA

−1.31 ± 3.37

Calorie labelling was implemented in one county and compared to a county where it was not, using a five-month baseline period and six-month post-intervention period for results calculation.

Hoefkens et al., 2011 [25]

Cohort

657

Canteen

Belgium

2.20 ± 10.4

Nutritional information was posted in a workplace canteen and employees surveyed regarding their lunchtime choices pre- and post-intervention.

Ogawa et al., 2011 [26]

RCT

1684

Market

Japan

9.10 ± 1.26

Point-of-purchase nutritional information was added. Over a 60 day period, sales of healthy items were measured and compared between a control store and an experimental store.

Pulos and Leng, 2010 [27]

Cross-sectional (pre-post comparison)

16000

Restaurant (other)

USA

2.00 ± 1.09

Nutritional labelling was added in a restaurant. Entrée sales from 30 days before and 30 days after implementation were used for comparison.

Roberto et al., 2010 [28]

RCT

303

Restaurant (other)

USA

11.3 ± 5.34

Experimental subjects were given a label with calorie information while the control group was given a normal menu. Calorie intake was compared between the groups.

Shimizu et al., 2010 [29]

RCT

122

Laboratory

USA

21.7 ± 10.5

Presentational cues were used to indicate either a “meal” or a “snack” condition, and intake compared between control and experimental groups.

Wisdom et al., 2010 [30] - A

RCT

290

Restaurant (fast-food)

USA

5.80 ± 4.31

Calories were added to fast-food restaurant menus and average calorie intake per purchase compared between control and experimental groups.

Wisdom et al., 2010 [30] - B

RCT

342

Restaurant (fast-food)

USA

9.00 ± 7.64

In the experimental group, selection of unhealthy items was made less convenient by adding a step to the ordering process and moving them to the second page of the menu.

Chu et al., 2009 [31]

Cross-sectional (pre-post comparison)

13951

Canteen

USA

1.98 ± 0.14

Calorie information was posted in a workplace canteen. Calorie content of meals was calculated from sales data before and after calorie posting.

Gerend, 2009 [32]

RCT

288

Laboratory

USA

5.54 ± 5.52

In the experimental group, subjects received a menu with calorie information, while the control group received a regular menu. Calories per meal requested were compared between groups.

Kelly et al., 2009 [33]

RCT

43

Laboratory

UK (Northern Ireland)

11.7 ± 7.97

Portion size was randomly varied over four days in the lab and consumption measured and compared between conditions.

Stroebele et al., 2009 [34]

RCT

59

Laboratory

USA

44.8 ± 2.68

Subjects randomly received snack packs of different sizes for a week. Consumption measured and compared between groups.

Ueland et al., 2009 [35]

RCT

33

Laboratory

USA

7.17 ± 17.3

Subjects were told that they had received a standard portion or a larger portion, though in fact the amounts were the same. Consumption measured and compared between conditions.

Viskaal-van Dongen et al., 2009 [36]

RCT

51

Laboratory

Netherlands

36.8 ± 2.13

Foods with similar caloric content and either hidden or visible fat were presented and consumption measured and compared.

Bodor et al., 2008 [37]

Cross-sectional (internal comparison)

102

Neighbour-hood

USA

35.7 ± 1.73

Household survey regarding healthy food consumption. Groups were compared based on their proximity to a local grocery store.

Raynor and Wing, 2007 [38] - A

RCT

28

Home

USA

44.7 ± 2.68

Portion size adjusted and consumption measured and compared between groups.

Raynor and Wing, 2007 [38] - B

RCT

28

Home

USA

6.43 ± 32.2

Size of individual foot unit adjusted and consumption measured and compared between conditions.

Rolls et al., 2007 [39] (1)

RCT

23

Home

USA

14.6 ± 2.49

Portion size adjusted and consumption measured and compared between groups

Rolls et al., 2007 [40] (2)

RCT

119

Laboratory

USA

2.62 ± 1.85

Size of plate used was changed between control and experimental conditions, while the amount of food was held constant. Consumption was measured and compared between groups.

Antonuk and Block, 2006 [41]

RCT

67

Laboratory

USA

37.0 ± 18.2

Subjects were given snack food with randomly varying nutritional labelling. Their consumption was measured and compared.

Wansink et al., 2006 [42] - A

RCT

85

Laboratory

USA

12.6 ± 11.9

Larger serving utensils were used in the experimental group and consumption measured and compared.

Wansink et al., 2006 [42] - B

RCT

85

Laboratory

USA

23.6 ± 15.3

Larger bowls but identical serving sizes were used in the experimental group and consumption measured and compared.

Hetherington et al., 2006 [43] - A

RCT

37

Laboratory

UK (England)

11.2 ± 2.82

Subjects ate a meal either in the company of friends or of strangers. Consumption in the two conditions was compared.

Hetherington et al., 2006 [43] - B

RCT

37

Laboratory

UK (England)

15.4 ± 2.82

Subjects ate a meal either while watching TV or while alone. Consumption in the two conditions was compared.

Huang et al., 2006 [44]

RCT

456

Online

Australia

0.620 ± 0.165

Availability of healthier options in online supermarket was adjusted. Change in sales of healthy items was compared.

Norton et al., 2006 [12]

RCT

30

Laboratory

UK (England)

13.7 ± 4.02

Sandwiches were provided either with a variety of fillings or with a homogenous filling, though energy content was constant. Calorie intake was measured and compared.

Wansink and Kim, 2005 [45]

RCT

72

Movie Theatre

USA

31.2 ± 10.2

Popcorn in a movie theatre was sold in containers of various sizes. Consumption was measured and compared.

Wansink et al., 2005 [46]

RCT

54

Canteen

USA

42.2 ± 18.8

Experimental subjects were unknowingly given self-refilling bowls to make second helpings more convenient. Consumption was measured and compared.

Devitt and Mattes, 2004 [47]

RCT

20

Laboratory

USA

6.04 ± 5.70

Food unit size was randomly adjusted on four separate days. Consumption was measured and compared.

Diliberti et al., 2004 [48]

RCT

180

Canteen

USA

20.2 ± 0.567

Portion size of canteen entrées randomly varied in a workplace canteen. Sales data used to calculate energy content of meals.

Levitsky and Youn, 2004 [49]

RCT

13

Canteen

USA

23.6 ± 2.42

Portion size of canteen entrées randomly adjusted. Consumption measured and compared between conditions.

Rolls et al., 2004 [50] (1)

RCT

60

Laboratory

USA

12.1 ± 1.70

Subjects were given snacks of differing portion sizes. Consumption between control and experimental groups compared.

Rolls et al., 2004 [51] (2)

Cohort

75

Laboratory

USA

44.7 ± 1.90

Over four weeks, subjects were given sandwiches of varying sizes at a once weekly lab lunch. Consumption between conditions compared.

Steenhuis et al., 2004 [52] - A

Cross-sectional (pre-post comparison)

290

Canteen

Netherlands

−4.51 ± 8.92

In a workplace canteen, availability of healthy options was increased or held constant at seventeen worksites that were randomly assigned. Sales were used to calculate calorie content of meals and compared between conditions.

Steenhuis et al., 2004 [52] - B

Cross-sectional (pre-post comparison)

215

Canteen

Netherlands

−8.05 ± 9.98

At 17 randomly assigned worksites, healthy items were given additional labelling or left in original state. Sales were used to calculate calorie content of meals and compared between conditions.