Promoting Healthier Food Purchases By Leveraging the Online-Grocery Environment

Rationale: Online-grocery shopping is predicted to be one of the "hottest" food trends of 2014, as national retailers such as Amazon, as well as start-up companies, venture into the e- commerce grocery sector. Importantly, the online-grocery environment could be uniquely manipulated to promote healthier food purchasing and help with weight control. Since consumers tend to choose items listed first on menus and buffet lines, the order of food products displayed on the grocer's website may impact purchasing. Furthermore, it's possible that in an online-grocery environment, nutrition information could be made more salient to consumers. For example, previous research has demonstrated that label color influences perceptions of the healthfulness of foods. The FDA also recently proposed a redesign of foods' nutrition facts panels, which would highlight calorie content in a larger font. Although implementing this label change on all food labels could take years, e-commerce sites could change the format of the nutritional information they display much more quickly.

Objectives: The proposed study intends to nudge consumers to make healthier grocery purchases through three distinct interventions: 1.) Manipulating the order of food items within grocery categories; 2.) Displaying product nutrition information in red or green; and 3.) Presenting calorie information in a larger font size. We propose to examine these concepts in adult consumers using a grocery e-commerce platform servicing socioeconomically and racially diverse communities in the northeastern U.S.

Study Overview

• Status: Unknown
• Conditions: Obesity
• Intervention / Treatment: Behavioral: Nutrition facts automatically or not automatically displayed; Behavioral: Nutrition facts in color; Behavioral: Nutrition Facts Label in Larger Font; Behavioral: Healthier Items First

• Study Type: Observational
• Enrollment (Actual): 400

Contacts and Locations

Study Locations

• United States
  • Vermont
    • Burlington, Vermont, United States, 05405
      • University of Vermont

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child; Adult; Older Adult
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

• Sampling Method: Non-Probability Sample

Study Population

All orders placed on any of the Rosie sites or apps in the study period will be included. No active recruiting of participants will take place, data will only be collected from existing Rosie customers. Participants will not be informed that manipulations are happening on the website, other than notices the company occasionally displays when they are revising features of the website. In order to determine whether any of the proposed "mindless" interventions actually work, it is imperative that consumer attention not be drawn to changes in the site. It is expected that 400-1000 orders per month will be included as part of the study during each two-month intervention period.

Description

Inclusion Criteria:

• Shopping on a Rosie grocery site

Exclusion Criteria:

• No exclusion criteria, all orders will be eligible

Study Plan

How is the study designed?

Design Details

• Observational Models: Other
• Time Perspectives: Prospective

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort

Intervention / Treatment

Grocery Customers; All customers who shop on the Rosie site during the intervention period.

Behavioral: Nutrition facts automatically or not automatically displayed; For two months nutrition information on the Rosie site will be displayed in a "tab" that consumers can click on to display the nutrition facts panels of products. Investigators will use NuVal scores to determine the fifteen "healthiest," and fifteen "least healthy" items in six grocery categories; chips, cookies, cereal & breakfast, yogurt, ice cream, and frozen pizza. We will then compare the number of clicks on nutrition information for healthier versus less healthy items to determine if there's a possibility of willful ignorance coming into play when choosing to purchase less healthy items.; Behavioral: Nutrition facts in color; For two months the font on products' nutrition facts labels will be displayed in red or green. For this proof-of-concept intervention, the color of a label's display will be determined randomly, in order to gather a clearer picture of whether label color could be leveraged to influence product choice. Items in each grocery category will be randomly assigned to receive green or red labels during the intervention period, so 50% of products in each category have red nutrition facts labels, and 50% have green nutrition facts labels. We will then compare proportions of green to red items purchased during the intervention period with the proportion of those same items purchased during the baseline period, when all labels were in black font.; Behavioral: Nutrition Facts Label in Larger Font; For two months the serving size and calorie lines on the nutrition facts labels will be displayed in a larger font for every item on the Rosie site. Average calories in customer purchases will be quantified using purchase data, and the average calories for online-grocery orders during baseline and intervention periods will be compared.; Behavioral: Healthier Items First; For two months, instead of the product popularity default-display option in the online store, thirty healthier items will be displayed on the first page in each grocery category. All grocery categories where there is scoring variation between healthy and less-healthy options (i.e. cereal, chips, bakery, sauces, etc.) will be included in the intervention. The thirty-targeted healthy items will be identified using NuVal scores. NuVal scores will not actually be displayed as part of the shopping site, and will be used exclusively as a guide to systematically determine targeted items in each food category. The proportion of targeted healthy items purchased in each category versus non-targeted items purchased during the intervention will be compared to the proportion of targeted vs. non-targeted items purchased during the baseline period.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Food Purchasing of targeted items	Company purchasing records will be used to track purchasing of items during the baseline period, as well as during the intervention periods, for the healthier items displayed first and nutrition-facts label interventions.	Continuously throughout the intervention expected average of 52 weeks
Click Rate of targeted food items	We will tabulate a "click-rate" for each targeted product during the intervention periods and at baseline to determine whether consumers clicked on the nutrition facts panels more often for healthier versus less-healthy products, and if consumers clicked more often when calories were displayed in larger fonts or were listed in red/green. We will also be able to determine whether customers clicked more often on the targeted healthy items when they were listed first on the website in each category. Clicks for each item will be tabulated weekly and averaged over the intervention or baseline periods.	Continuously throughout the intervention expected average of 52 weeks
Nutrient Analysis of Food Purchased by consumers	To determine whether average number of calories purchased changed when calories and serving sizes were displayed in larger fonts, we will use purchasing data to compare the average number of calories purchased in orders before the font change to the average number of calories purchased during the font change. Calories per item will be determined using the USDA database, which houses nutrition facts for all foods on the Rosie website.	Continuously throughout the intervention expected average of 52 weeks

Collaborators and Investigators

• Sponsor: University of Vermont

Study record dates

Study Major Dates

• Study Start: August 1, 2015
• Primary Completion (Anticipated): August 1, 2019
• Study Completion (Anticipated): August 1, 2019

Study Registration Dates

• First Submitted: June 29, 2015
• First Submitted That Met QC Criteria: June 30, 2015
• First Posted (Estimate): July 3, 2015

Study Record Updates

• Last Update Posted (Actual): May 3, 2018
• Last Update Submitted That Met QC Criteria: April 30, 2018
• Last Verified: April 1, 2018

More Information

Terms related to this study

• Other Study ID Numbers: 029898

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: No
• IPD Plan Description: Because the data is coming from a private business, we do not plan to share the data publicly.