Portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco
Systematic review summary
This document has been produced by the World Health Organization. It is a summary of findings and some data from the systematic review may therefore not be included. Please refer to the original publication for a complete review of findings.
Key Findings review
- The data included in this review are from predominantly low-deprivation populations in high-income countries and thus the findings may not be applicable to all settings
- In both adults and children, exposure to larger-sized portions, packages, individual units or tableware increased the quantity of food consumed, and in adults, increased the quantity of food selected for consumption
- Shorter, wider glasses or plastic bottles may increase the quantity of non-alcoholic beverage consumed or selected for consumption in comparison with taller, narrower glasses or plastic bottles
- Few studies manipulating the size of cigarettes were identified and no studies investigating the effect of exposure to different sizes or shapes of alcoholic beverage products were indentified
1. Objectives
To evaluate interventions involving different portion, package, individual unit or tableware sizes on ad libitum selection or consumption of food, alcohol or tobacco products in adults and children, and to assess whether study, intervention and participant characteristics modify the effects of such interventions2. How studies were identified
The following databases were searched in January 2015:3. Criteria for including studies in the review
3.1 Study type
Randomized controlled trials, including crossover trials3.2 Study participants
Adults and children3.3 Interventions
Interventions comparing at least two different sizes or physical dimensions (shape, volume, height, width, depth) of either a portion of the same food, beverage, alcohol or tobacco product, its package or individual unit size, or an item of tableware used to consume it(Studies involving non-ad libitum product consumption or selection were excluded, as were studies comparing different types of tableware, e.g., spoons versus chopsticks. Studies involving the indirect manipulation of product sizes/dimensions, e.g., via organizational-level encouragement to reduce portion sizes, were excluded, as were studies using concurrent interventions such as nutritional labelling)
3.4 Primary outcomes
Consumption of a product
4. Main Results
4.1 Included Studies
Seventy-two randomized controlled trials were included in this review4.2 Study settings
4.3 Study settings
How the data were analysedFour main comparisons were made: i) effect of a larger size on consumption; ii) effect of shape on consumption, iii) effect of a larger size on selection without purchase; and iv) effect of shape on selection without purchase. Both random effects and fixed effects meta-analysis were used to produce standardized mean differences (SMD) and 95% confidence intervals (CI). An SMD of 0.2 represented a small effect, 0.5 a moderate effect, and 0.8 a large effect. If a study compared multiple portion sizes, independent pairwise comparisons were made between each incremental size (e.g., 275 g versus 367 g, 367 g versus 458 g, and 458 g versus 550 g). Sensitivity analyses doubling and halving any imputed standard deviations (SD) were conducted. Meta-regression analyses were performed to determine which study-level covariates best explained the between-studies variance in effect size, using the following covariates:
Effect of a larger size on consumption
In meta-analysis of 92 comparisons from 61 trials, exposure to larger-sized portions, packages, individual units or tableware was found to significantly increase consumption (SMD 0.37, 95% CI [0.29 to 0.45], p<0.001; 6711 participants). Sensitivity analyses, in which the imputed SDs for five comparisons were doubled or halved, had no meaningful effect on the results.
Study characteristics
In meta-regression analysis by type of product, the effect of a larger size on consumption was not altered for foods (SMD 0.38, 95% CI [0.29 to 0.46], 86 comparisons/6603 individuals), but for tobacco, exposure to longer versus shorter cigarettes did not statistically significantly increase the quantity smoked (SMD 0.25, 95% CI [-0.14 to 0.65]; 3 trials/108 individuals). Studies using a within-subjects design had a smaller effect size than those using a between-subjects design (SMD 0.23, 95% CI [0.16 to 0.28], 65 comparisons, and SMD 0.64, 95% CI [0.47 to 0.80], 27 comparisons, respectively). Less healthy food and more energy-dense food increased the effect size, with each 10-point increase in the FSA nutrient profile score and each one-point increase in the FSA energy density score corresponding to a 0.06 (95% CI [0.04 to 0.22]) and a 0.04 (95% CI [0.00 to 0.08]) unit increase, respectively. When the intervention food comprised all of the food on offer, the effect size was 0.22 units larger (95% CI [0.02 to 0.41]), and the effect size was also increased when consumption of the intervention food alone was the outcome measure (0.32 units larger, 95% CI [0.16 to 0.48]). No evidence of a difference in effect by the target of manipulation (tableware, package, portion, individual unit, package with individual unit) was found.
Intervention characteristics
In meta-regression analyses, the absolute and relative differences in size between the two portions, packages, individual units or items of tableware were not associated with the effect of larger size on consumption. However, a bimodal distribution in relation to the relative difference in size was noted and investigated, with a greater effect found for items 200% larger than the smaller size (SMD 0.50, 95% CI [0.31 to 0.69], 25 comparisons/1414 participants), in comparison to items 120 to 160% larger (SMD 0.25, 95% CI [0.15 to 0.35], 39 comparisons/2415 individuals).
Participant characteristics
Effect sizes were larger with increasing age: for each 10-year increase in mean age, a 0.09 unit (95% CI [0.00 to 0.18]) increase in the amount of food or tobacco consumed was observed. This finding was confirmed when comparing studies in children (SMD 0.21, 95% CI [0.10 to 0.31], 22 comparisons/1421 participants) to those in adults (SMD 0.44, 95% CI [0.33 to 0.54], 70 comparisons/5182 participants). No association between sex, BMI, hunger, dietary restraint, or dietary disinhibition and the effect of larger size on consumption was found.
Final model
A final random effects meta-regression model including the covariates study design (within-subjects/between-subjects), FSA nutrient profile score, FSA energy density score, mean age, and whether there were other non-manipulated foods provided (yes/no) explained 91% of the between-studies variation in effect size (R² 90.77%, p=0.001).
Effect of shape on consumption
In one trial involving 50 participants, the effect of shape of water bottle on water consumption following exercise was investigated. Shorter, wider bottles led to increased water consumption in comparison with taller, narrower bottles (SMD 1.17, 95% CI [0.57 to 1.78]).
Effect of larger size on selection without purchase
All identified trials examined the effect of different sizes of portions, packages or tableware on quantities of food selected for consumption. Larger portions, packages, individual units or tableware increased the amount of food selected for consumption (SMD 0.42, 95% CI [0.24 to 0.59], p=0.011; 13 comparisons/1164 participants). Sensitivity analyses in which the imputed SD of one study was altered did not meaningfully affect the results.
Study characteristics
In meta-regression analysis, trials using a within-subjects design had a smaller effect size than those using a between-subjects design (SMD 0.14, 95% CI [-0.07 to 0.34], 4 comparisons, and SMD 0.55, 95% CI [0.39 to 0.68], 9 comparisons, respectively). The effect size was increased when consumption of the intervention food alone was used as the outcome measure (0.41 units larger, 95% CI [0.06 to 0.76]). No evidence of a difference in effect size by the target of manipulation (tableware, portion) was found.
Intervention characteristics
The relative difference in size between the two items was not associated with the effect of larger size on selection without purchase.
Participant characteristics
The effect of a larger size on selection without purchase was found to be present in adults (SMD 0.55, 95% CI [0.35 to 0.75], 9 comparisons/782 participants) but not in children (SMD 0.14, 95% [-0.06 to 0.34], 4 comparisons/382 individuals). No evidence of a difference in effect size by sex was found, and no other study-level participant characteristics could be investigated due to insufficient data.
Final model
A final random effects meta-regression analysis including the covariates study design (within-subjects/between-subjects) and whether the outcome measure included selection of other non-manipulated foods (yes/no) could not be performed due to collinearity. Models including each of these covariates separately both produced an R² of 79%.
Effect of shape on selection without purchase
Three comparisons involving 232 participants investigated the effect of shorter, wider empty glasses or plastic bottles on the selection of fruit juice or water. Shorter, wider glasses or plastic bottles increased the amount of juice selected for consumption (SMD 1.47, 95% CI [0.52 to 2.43]). Due to insufficient data, no meta-regression analyses were performed. In the single study conducted in children, an SMD of 2.31 (95% CI [1.79 to 2.83], 96 participants) was observed. Pooled analysis of the remaining two trials involving 136 adults produced a smaller effect size (SMD 1.03, 95% CI [0.41 to 1.65]).
5. Additional author observations
The overall risk of bias among studies measuring consumption was unclear (58/65 trials) or high (9/65 trials). Similarly, the overall risk of bias among studies measuring selection was unclear (9/13 trials) or high (4/13 trials). Using the GRADE approach, the evidence for an effect of a larger size on consumption was rated as moderate quality overall, and moderate quality for both adults and children. The evidence for an effect of a larger size on selection was rated as moderate quality among adults and overall, and low quality for children. Comparisons examining the effect of longer versus shorter cigarettes on consumption were rated as low quality evidence, and comparisons of shorter, wider glasses or plastic bottles versus taller, narrower glasses or bottles were rated as very low quality evidence. No assessments of the potential effect modification of socioeconomic position could be made, and no trials were conducted in low- or middle-income countries.Based on the evidence reviewed here, larger-size portions, packages, individual units or tableware increase the quantity of food consumed in both adults and children, and increase the quantity of food selected for consumption in adults. Translating the effect size for consumption into calorie intake, a reduction in exposure to larger-sized food portions, packages and tableware across the whole diet could reduce mean UK adult energy intakes by 215 to 279 kcal/day. Few studies assessed cigarette consumption, and no studies manipulating the size or shape of alcoholic beverage products were indentified. Shorter, wider glasses or plastic bottles may be associated with a greater quantity of fruit juice or water consumed or selected for consumption.
While smaller-size portions, packages, individual units, or tableware reduce food consumption in clinical trials, the sustainability of the effect under repeated, prolonged, or free-living conditions remains to be investigated. Studies evaluating the effects of size and shape manipulation on alcoholic beverage consumption are needed. In addition, further trials establishing effect sizes under exposure to different sized portions, packages, individual units, or tableware are warranted, with particular regard to smaller absolute sizes and differences in relative sizes.
*The authors of the systematic review alone are responsible for the views expressed in this section.