MONTREAL — Artificial intelligence software deployed by the CIUSSS de l’Ouest-de-l’Île-de-Montréal would be able to personalize users’ menus based on their preferences, aversions and specific nutritional needs.
The technology also reduces the risk of error in terms of allergies, cross-contamination and other food incidents, we assure.
The software has been gradually deployed, since June 2023 and after almost ten years of work, in the Lakeshore, Saint Mary’s and Lasalle hospitals, as well as in the Nazaire-Piché and Denis-Benjamin Viger CHSLDs. It could now be deployed in other establishments.
“There is always going to be human intervention,” said Debby Berteau, who is a nutrition research officer. But the goal is to improve the offer while limiting interactions.”
If a user is entitled to a cake for dessert, she cites as an example, the software will take care of calculating the carbohydrate content, “which allows us to offer more variety, (…) we no longer only see the foods that are allowed or forbidden.” This also has the advantage of freeing technicians or nutritionists from these tasks.
It would be difficult to do the same thing manually, said Marie-Hélène Cyr, who is a nutritionist in a consulting role, since “it would take a lot of resources.”
“It would take a dietetic technician who would manually do calculations to know, OK, this user needs 60 grams of carbohydrates at each meal,” she explained. She would have to calculate, and then if there is an aversion or a preference, that changes her calculations. Depending on the offer on the menu, you would have to recalculate every day. We would never have enough manpower to do what the computer is doing for us right now.”
The project “introduces a new approach to nutritional prescription”, we explained by email, an approach which is “based on a description of needs rather than on predefined diagnoses”.
Officials say the software is able to automatically adjust the foods offered, thereby increasing variety and user satisfaction while respecting their dietary restrictions.
They indicate, for example, that artificial intelligence makes it possible to offer 61 additional foods for a potassium-restricted diet. A menu calibrated in potassium, it is specified, offers on average 1800 kcal and 75g of protein, compared to the 1750 kcal and 70g of protein of traditional diets.
The program would also increase the variety of foods offered, since the calibrated menus would include 16% more available foods. For users requiring several restrictions, it is added, the menu now provides 1775 kcal instead of 1400 kcal, an increase of 25%.
Functioning
Dietetic technicians first collect users’ food preferences and aversions, as well as their specific nutritional needs. The software is then configured with food nutritional values, recipes and standardized serving sizes.
Artificial intelligence then automatically adjusts the menus based on the data collected, without requiring manual interventions. Any change in the nutritional values of foods or recipes is automatically taken into account by the system.
Lakeshore General Hospital already had a computer program in place several years ago to supervise and manage certain aspects of users’ diets. It was on this basis that work began in 2014, which then accelerated in 2019 and led to the current tool.
“It requires a lot of data in the background that must be accurate in the computer system,” explained Ms. Cyr. And now we will have to keep them up to date.”
Personalization
Menu customization would now be faster and more accurate, reducing the need for manual corrections. Users would therefore receive foods adapted to their restrictions, with fewer risks to their health.
When dealing with a diabetic user, Ms. Cyr cited the example, “we had to make sure that the total carbohydrates in each meal were not too high.”
“We had to remove the cake for everyone,” she said. No one, who is diabetic, could receive cake at all times. This (almost) never happened. But now, if my target for dinner is 75 grams (of carbohydrates) and the tray total allows it and I have plenty of space left, that evening the person will be able to eat cake for dinner.
And as the clientele served is increasingly older, added Ms. Cyr, undernutrition often becomes a factor that must be taken into account. A contradiction then arose: as much as we wanted to give the user foods rich in protein and energy, we often had to remove these foods due to the diabetic menu that they had to follow.
“Now we can meet both of those needs,” she said. “We can make sure that all the caloric values are met, but also that we respect the total carbohydrates at each meal. So we increase the variety at each meal and people are more inclined to eat.”
The project was first deployed on a single floor of the Lakeshore General Hospital, recalled Ms. Berteau. Users who were hospitalized on this floor were quick to express their dissatisfaction when they were transferred elsewhere and lost access to this personalized menu, she explained.
“The patients did not want to return to the defended exceptions,” said Ms. Berteau. They wanted to keep the new diet which had more variety. They no longer wanted the old regime which they found too restrictive.”
It was still necessary to educate and reassure patients who suddenly saw foods appear before them to which they were not accustomed and which they even believed were forbidden, said Ms. Cyr.
We also had to take the time to show the employees concerned, who were wondering what impact the new system would have on their work, the benefits they would gain from it, added Ms. Berteau, but also the benefits for users.
“There was some reluctance at first because it was a change in practices, but we learned from our previous experiences and worked (with employees) to find solutions,” she said. “Employees contributed solutions, so it made the deployment easier.”
The employees eventually discovered, added Ms. Cyr, that the new system did not take away their work, quite the contrary. Instead, it allowed them to focus on users with more complex needs, and in the end, “we had more users who were better fed.”
