Abstract P323: Effectiveness of Point-of-Purchase Labeling on Dietary Behaviors and Nutrient Contents of Foods: A Systemic Review and Meta-Analysis
Background: While point-of-purchase food labeling in restaurants is increasingly being planned or implemented to promote healthy choices, its effects on consumers’ behaviors or industry’s product formulations are not well established.
Method: Using MOOSE and PRISMA guidelines, we systematically searched multiple online databases to identify interventional trials or prospective observational studies of point-of-purchase labeling in restaurants and changes in intakes of foods/nutrients, industry food reformulation, and CVD risk factors (adiposity, blood pressure, blood lipids). Studies were excluded if cross-sectional, not aiming to improve diet, or evaluating supermarkets. Data were extracted independently and in duplicate by 2 investigators. Pooled estimates were calculated using inverse variance-weighted meta-analysis. Pre-specified potential sources of heterogeneity (study location, targeted consumers, type of restaurant, presence of legislation, label placement, label format and delivered information) were assessed by using meta-regression. Funnel plots and Begg’s and Egger’s tests assessed potential for publication bias.
Results: Of 1,899 abstracts, 19 met inclusion criteria: 15 from North America and 4 from Europe. These included 8 RCTs, 6 non-randomized trials, and 5 prospective cohorts; conducted in laboratory (N=8 studies) or restaurant (N=11) settings. Labeling provided information on contents of energy or specific nutrients (N=13), traffic light codes or other grading systems (N=6), or health claims/symbols (N=3); and were placed on menus (N=10) or next to the food at the point-of-purchase (N=9). As percent change from baseline/control, pooled analyses found no significant association of point-of-purchase labeling with intake of total energy (N=23, -3.4%, 95%CI -8.2%, 1.3%), total fat (N=8, -4.5%, -14.7%, 5.6%), saturated fat (N=4, -6.4%, -29.5%,16.7%), carbohydrate (N=4, 0.6%, -3.3%, 4.5%), protein (N=3, -1.1%, -9.9%, 7.7%), sodium (N=3, -6.9%, -25.2%,11.5%), or vegetables (N=3, -1.0%, -17.7%,15.6%). When industry reformulations were evaluated, labeling was associated with lower sodium content (N=5, -4.7%, -8.8%, -0.6%), but not total energy (N=5), saturated fat (N=3), cholesterol (N=3), or fiber (N=3) contents. Too few studies assessed labeling and CVD risk factors. In meta-regression, no significant effect modification of findings was evident. Little evidence was seen for publication bias.
Conclusion: These interventional and prospective findings suggest limited effects of point-of-purchase labeling in restaurants on consumers’ dietary behaviors, and potential effects on food industry reformulations of sodium content but not other factors. These findings highlight the need for additional investigation of labeling as well as prioritization of more robust evidence-based policies.
Author Disclosures: S. Shangguan: None. J. Smith: None. W. Ma: None. L. Tanz: None. A. Afshin: None. D. Mozaffarian: B. Research Grant; Modest; R01.
- © 2015 by American Heart Association, Inc.