How does Libre Assist work?

Food decisions happen at the speed of life, and they can be even more complex if you’re living with diabetes. But now there’s support in the form of Libre Assist1, a new feature in Abbott’s Libre app2 that helps you predict3 and understand how a meal might impact your glucose — before you take a bite.

How exactly does it do that? First, Libre Assist gets its information about the meal straight from the source, either a picture you take or a text description you provide. Then, after analyzing each part of the meal, Libre Assist gives you an easy-to-understand, color-coded prediction:

• Green: Minor impact on your glucose
• Yellow: Moderate impact on your glucose
• Orange: Major impact on your glucose

Knowing that your response to food is unique — the same meal might affect your glucose levels differently than it would someone else — Libre Assist also delivers a few personalized tips4 to help lower the glucose impact, if the prediction shows a moderate or major impact.

Finally, if you decide to eat the meal and you’re wearing one of Abbott’s FreeStyle Libre continuous glucose monitoring (CGM) systems sensors, after three hours, Libre Assist will take your actual glucose data into account and confirm what really happened to your glucose levels.

To Sunny Nagra, it represents a major leap forward from traditional food-logging apps, which tend to only give you insights after you’ve already made your food choices and eaten. Nagra is a product expert in Abbott’s diabetes care business. He helped develop the early prototype for Libre Assist and worked with his colleagues in product innovation, engineering, medical and regulatory to bring it to the Libre app. At the same time, the team had the data validated by certified diabetes care and education specialists (CDCES)5.

“We took a lot of care in trying to make the Libre Assist experience easy,” Nagra says. “It’s giving guidance with a simple representation of potential glucose impact in that score of green, yellow or orange.”

Once you set up Libre Assist, it can be as easy as snapping a pic of your meal to get that prediction and guidance. So, with Nagra’s help, let’s go deeper into how this app feature helps you make more informed food choices.

Who is Libre Assist for?

Libre Assist¹ was designed to help people living with diabetes make mealtime decisions.

It is a feature within the Libre app², so you’ll need to download and set up the Libre app in order to use it. It can be downloaded from the Apple or Google Play stores.

To access Libre Assist, open the Libre app and tap the “Insights” tab on the bottom navigation.

How do I analyze a meal?

You can get started in three ways:

• Take a photo of a meal with your phone
• Upload a photo you already took
• Type in a text description of what you’re planning to eat

“If you take a photo, Libre Assist’s AI platform analyzes the photo,” Nagra explains. “It’ll look at the components of that meal that is in the photo. For example, if you have a photo of a breakfast, it’ll recognize the toast, the eggs, the sausage, the hash browns in there.”

Your meal is given a name (so you can view your history later), and then Libre Assist assigns green-yellow-orange scores to each ingredient of the meal.

You also see an overall glucose impact prediction for the meal as a whole — again, green for minor impact, yellow for moderate and orange for major.

Can I edit a meal?

Yes. As Nagra explains: “You know what you’re about to eat, so if anything is wrong you can just use natural language to correct what the meal is. And once you correct it, Libre Assist will run the analysis again and update the prediction.”

What do I do after the prediction?

If the prediction is a moderate or major impact, there are a few additional actions you can choose to take before eating your meal.

For starters, Libre Assist¹ pairs these predictions with helpful tips⁴ on how to lessen the glucose impact.

The advice might be something like eating the fiber component of the meal before the starch. Or asking for whole grain bread for a sandwich.

You can also look at the individual ingredient scores to see if the majority of them come from one or two meal components. That makes it easier to consider an ingredient swap, whenever that’s possible.

If Libre Assist predicts a minor impact, there’s no action provided and nothing standing in the way of you enjoying your meal.

Can I match up the prediction with what actually happened after I ate?

If you’re wearing one of Abbott’s Libre CGM systems sensors, Libre Assist will leverage your actual glucose data after you have eaten the meal — typically within three hours after eating. Libre Assist then assigns the meal a final glucose impact that takes into account where your glucose levels started and where you ended up.

This helps confirm the meal’s actual impact, so you know a little bit better what to expect the next time you eat it.

How does Libre Assist perform the analysis?

Libre Assist’s AI platform analyzes and interprets the contents of the meal you are considering. It’s important to note that this platform was validated by certified diabetes care and education specialists (CDCES), to make sure there was agreement on the impact predictions Libre Assist is giving. Abbott’s decades of leadership in diabetes care also informed the platform’s development.

“We went the extra mile to make sure the app is not putting any judgments on food or telling you not to eat something,” Nagra says. “It’s giving tips so that you feel more confident the next time you eat that meal.”

Do you have to pay to use Libre Assist?

No, it is included in the Libre app, with no purchases or fees required.

The product images are for illustrative purposes only. Not actual patient data.

The views expressed should not be used for medical diagnosis or treatment or as a substitute for professional medical advice.

References

¹ Libre Assist is a feature within Libre app that uses generative artificial intelligence to provide information on how foods could impact your glucose levels. Generative artificial intelligence may not always be accurate, and it should not be used to make treatment decisions.

² The FreeStyle Libre systems apps are only compatible with certain mobile devices and operating systems. Please check our website for more information about device compatibility before using the apps. Use of the FreeStyle Libre systems apps may require registration with LibreView.

³ Predicted glucose impact is based on user-provided food data and may differ from actual impact, which depends on sensor readings and factors like activity, stress, medication, and alcohol. For personalized advice, consult your healthcare provider.

⁴ Personalized food suggestions are based on food preference information inputted by the user.

⁵ Data on File, Abbott Diabetes Care, Inc.