MONTREAL — The Nutrition Facts label required on most food products in the United States and Canada failed to measure up when compared to three other labeling systems in use around the world, according to a study by McGill University researchers published in the December issue of the Annals of the New York Academy of Sciences.

As part of the study, the researchers conducted an Internet-based food-choice experiment and a computational decision-making model.

The Nutrition Facts label, which lists the per cent daily value of several nutrients, took more time to understand than the other labeling systems and led to the least nutritious choices.

The study found the most usable labeling scheme was NuVal, which researchers said quantifies nutritional information, presents it in a way that is quick and easy to use, and resolves nutritional conflicts. NuVal is a shelf sticker used in some American food markets, which indicates the overall nutritional value of each food item with a number from 1-100. NuVal scores are calculated by nutrition experts at several universities, including Yale, Harvard and Northwestern, and emphasize both the positive and negative aspects of each food.

“NuVal labels are fast to use and yield nutritious choices,” the researchers said. “This does not necessarily imply that NuVal labeling should be used everywhere. It is possible that alternate labeling schemes with these same characteristics could do as well or better than NuVal labels.”

Two other labeling methods examined in the study produced mixed results.

The Heart label, a binary labeling scheme designed for the study, is similar to existing binary schemes like Canada Health Check, the Swedish National Food Agency’s Keyhole label, and several others in presenting a symbol that certifies a food item as nutritious. Researchers found that while Heart labels are fast to use, they produce choices that are not especially nutritious.

“With a more realistic higher proportion of foods certified as nutritious, binary labeling schemes could be expected to do even more poorly in nutrition than we find here,” the researchers said. “In more realistic scenarios, binary labeling would produce many more ties between food items with a substantially wider range of nutritive value.”

The fourth system examined was the Traffic Light system used in the United Kingdom. The researchers said the Traffic Light system takes more time to use and yields only moderate increases in nutrition.

“Our computer simulation successfully captures and provides potential explanations for the main findings from the human experiment,” the researchers said. “In the simulation, the nutrition label attributes have the same combined attention weight, regardless of which labeling scheme is used. The greater nutritional success of NuVal, as compared to Traffic Light and %DV, is partly explained by the multiplicity of attributes in Traffic Light and %DV labels (four and nine, respectively). Compared to NuVal, the other three labels also suffer from the fact that they create decisional conflicts. For example, in a particular choice situation, one product may have lower sugar and salt content, but higher fat content and calorie count. It is nontrivial for a human shopper to resolve such a conflict, and the DFT model represents this as a struggle between preferences, as attention shifts back and forth between the conflicting attributes.

“If some other attribute (e.g., taste score or visual appearance) strongly favors a particular choice, it can easily dominate over conflicting nutrition information. In contrast, a single-attribute scheme like NuVal resolves such nutrition conflicts, rather than highlighting them, thus providing more guidance for decision making.”