select search filters
briefings
roundups & rapid reactions
Fiona fox's blog
A study published in JAMA Network Open looks at UPF consumption and behaviour in children.
Prof Stephen Scott, Professor of Child Health and Behaviour, Institute of Child Psychiatry, Psychology and Neuroscience, King’s College London, said:
“I think this paper has significant limitations – the authors have identified an association but they don’t address the possibility that reverse causality could be going on (that children who have behavioural problems eat more UPF). We know that lower socioeconomic status is associated with both UPF intake and behavioural issues in children, but they don’t discuss this as a possible limitation of the association in the discussion section. We don’t know whether the children might have already had behavioural problems at the start – the CHILD study appears to have collected data on behaviour at age 3 but they don’t discuss it in the paper. It seems surprising to me that the referees of this paper did not ask that these issues were addressed by the authors.”
Prof Jules Griffin, Director of the Rowett Institute, University of Aberdeen, said:
“The paper by Kavanagh and colleagues raises concerning questions about how ultraprocessed food (UPF) may impact aspects of behaviour and emotional development in young children, and by implication what the long-term impact may be of diets high in UPF. They initially measure diet at three years of age and then follow-up at five years of age to examine behaviour and emotional development. The study is impressive for its scale, following over two thousand children and collecting data on both behaviour and diet – not easy for children! These types of cohort studies are needed to understand whether an effect may occur in the real world, but care is needed to interpret associations as to what is the cause and effect in such circumstances – both UPF and behaviour may be associated with a third, unidentified factor. The authors address this, to a degree, by examining whether any of the associations could be explained by other common confounding effects, most notably household income, and indeed the associations remain. These cohort studies are challenging to perform with the effect sizes being relatively small and high degrees of uncertainty in some of the results. One of the most interesting analyses the authors perform in the study is examining which components of UPF are most responsible for the associations found. Artificial and sugar-sweetened beverages are identified in the associations between UPF and behavioural outcomes – this is interesting as there is good evidence that sweetened beverages can alter appetite regulation in the brain, so it seems plausible that sweetened beverages may also affect behaviour. While the study raises concerning questions, we need more research to test these associations.”
Prof Franco Sassi, Director of the Centre for Health Economics & Policy Innovation, Imperial Business School, Imperial College London, said:
“This new evidence that UPFs may be linked with adverse behavioural and emotional symptoms by age 5 is alarming, but not conclusive. The finding is based on a rigorous cohort study, but this cannot rule out that the children who consume larger amounts of UPFs and experience behavioural symptoms are different from other children in ways that are not observed and could not be controlled in this research. Further studies will be needed to confirm these findings and identify the underlying mechanism that may link UPFs with behaviours and emotions in young children.”
Dr Fiorella Picchioni, Research Fellow in Food Systems and Social Reproduction, University of Bristol, said:
“Findings from this Canada‑based study indicate that nearly half (45.5%) of the daily energy intake among participating children (age 3) is derived from ultra‑processed foods (UPFs), so understanding how these foods relate to child development is a major public health priority. This study draws on a rich dataset and, although firm conclusions cannot be drawn from statistical associations alone, the findings provide a valuable starting point for more detailed research into how UPFs consumption may be linked to children’s behavioural outcomes. The authors report that modelling the substitution of UPFs with minimally processed foods was associated with improved behavioural and emotional symptoms, reinforcing efforts to enhance the nutritional quality of foods offered in homes, childcare centres, and schools.
“The general body of research on food and health supports strategies that promote minimally processed foods in early childhood, and highlight the need to address the production, accessibility, and heavy marketing of ultra‑processed foods targeted at young children and their families.”
Dr Ramya Srinivasan, Associate Professor in Child and Adolescent Psychiatry, UCL Division of Psychiatry, UCL, and North London Foundation Trust, said:
“UPFs are an area of increasing public and research interest. There has been little research into the effects of UPFs and child emotional and behavioural difficulties however, the results of this study need to be interpreted with caution. Whilst the study finds an association between high UPF diet and higher emotional and behavioural symptoms, it does not show that a UPF diet caused these. In particular, the study does not account for baseline emotional and behavioural symptoms. They do not adjust for prior mental health problems – this is a problem as children with more mental health difficulties at baseline may be more likely to eat more UPFs and therefore have more mental health problems. Similarly, they do not account for parental mental health which is another important confounder (they do adjust for prenatal stress but this is not the same). In addition, they have adjusted for physical activity and BMI at 5 years (when the outcome is measured) which is unusual and they do not explain why they have chosen to do this in the main manuscript. It would be premature to offer dietary advice for behavioural reasons based on this one study.”
Prof Gunter Kuhnle, Professor of Nutrition and Food Science, University of Reading, said:
“The key problem with all observational studies on ultra-processed foods is the difficulty in measuring intake. In nutrition, this is often done by questionnaires – but they usually don’t distinguish between a sandwich made with home-made sourdough bread or a shop-bought one. Unfortunately, this study is not different: it uses a questionnaire that has not been designed to estimate ultra-processed food intake. These questionnaires often combine many similar foods in a single question – sometimes groups of foods that straddle different NOVA groups. This makes the interpretation of results difficult, as they refer more to specific dietary patterns and not necessarily to ultra-processed foods.
“Furthermore, associations between diet and behaviour are notoriously difficult to measure. While diet can affect behaviour, the opposite is also true. Many ultra-processed foods are easy to prepare and often also easy to eat – and children’s behaviour might affect parental choice of foods.
“The effect sizes found in this study were modest, and the results suggest that only two types of ultra-processed foods – ready meals and sugar-sweetened beverages – were associated with worse behaviour. ‘Ready meals’ are a very heterogeneous group of foods and it is difficult to interpret these findings without more information. Sugar-sweetened beverages however have been consistently associated with adverse health effects across study – but not due their status as “ultra-processed foods” but because they are high in sugars and facilitate overconsumption. (In the UK, it is recommended that children 5 years and younger should not consume sugar-sweetened beverages to limit sugar intake.)
“There is a concern that the “ultra-processed food” narrative is distracting from investigating the actual factors underlying those associations. While it is appealing to propose a simple explanation for behavioural problems among children, there is a risk that other factors – especially social factors – are left unaddressed.”
‘Ultraprocessed Food Consumption and Behavioral Outcomes in Canadian Children’ by Meaghan E. Kavanagh et al. will be published in JAMA Network Open at 16:00 UK time on Tuesday 3 March 2026, which is when the embargo will lift.
DOI: 10.1001/jamanetworkopen.2026.0434
Declared interests
Prof Stephen Scott: “No conflicts of interest. I have recently written a book, Eating Together, about the value of family meals.”
Prof Jules Griffin: “I am a consultant for Sitryx, a company specialising in designing drugs to target immunometabolism. The company is looking at drugs for irritable bowel disease and psoriasis.
I have received funding from the European Union to investigate endocrine disrupting chemicals and hold a grant from UK Research and Innovation examining the health benefits of a fish diet.
I hold shares in GlaxoSmithKline and Haleon plc.”
Prof Franco Sassi: “No conflicts of interest to declare.”
Dr Fiorella Picchioni: “No conflicts of interest to declare.”
Dr Ramya Srinivasan: “I do not have any interests/conflicts.”
Prof Gunter Kuhnle: “• Current funding: BBSRC Transforming UK Food System (FoodSEqual);