AI predicts healthiness of eat-out menus in Britain
Study conducted by the University of Cambridge
Cambridge researchers have leveraged Artificial Intelligence (AI) to forecast the nutritional quality of menus in cafés, takeaways, and restaurants in Britain.
This groundbreaking study, featured in Health & Place, reveals disparities in food environments across British local authorities, especially affecting residents in economically deprived areas.
The investigation, led by Yuru Huang from the University of Cambridge, underscores the prevalence of less nutritious ‘out-of-home’ dining options, which often contain higher calorie, fat, and salt content than homemade meals.
Concerns mount over its impact on public health with an increasing reliance on such menus, particularly in deprived regions.
Healthiness score
Drawing data from 55,000 food outlets, Huang and collaborators devised a healthiness score for each menu, encompassing various factors such as special offers, dessert options, and vegetable diversity.
Utilising a sophisticated deep-learning model, they extended their analysis to 180,000 out-of-home food establishments nationwide.
Researchers identified noteworthy trends by categorising outlets and their menus into cafés, fast-food joints, pubs, and restaurants.
Surprisingly, the predictive power of an outlet’s name emerged as a significant indicator of menu healthiness, albeit supplemented by hygiene ratings.
Despite the limitations of their approach, particularly in capturing nuanced menu characteristics, the study’s findings shed light on geographical disparities.
Notably, areas with higher deprivation levels exhibited lower menu healthiness across all food outlet categories and a proliferation of such establishments.
The University of Cambridge researchers stress the urgency of targeted interventions in unveiling this ‘double burden’ scenario, in which deprived areas harbour a higher density of less healthy food outlets and poorer menu choices.
Addressing these disparities could mitigate the health consequences disproportionately impacting vulnerable communities.
While acknowledging the study’s scope, Huang emphasises the need for nuanced interventions beyond mere menu healthiness assessments.
Strategies encompassing portion control, healthier cooking methods, and community engagement initiatives could offer tangible solutions in combating the adverse effects of unhealthy food environments.
This pioneering research, supported by the Medical Research Council and Gates Cambridge, highlights AI’s critical role in public health and underscores the imperative for localised interventions to foster healthier eating habits across diverse socioeconomic landscapes.
Featured image: The study’s findings shed light on geographical disparities. Credit: Arnold Pinto
Last Updated on 9 months by News Desk 1
