PoshanSenseThe extra 508 calories: what a small feeding trial really tells us
A landmark 2019 feeding trial helps explain why food context matters. It does not tell you how many calories a packet adds to your day.
A study result is a clue to understand, not a promise for your body.
You might have seen the claim that ultra-processed food makes people eat about 500 extra calories a day. That number comes from a real experiment. It is also very easy to turn into something the experiment never said. Before the number starts policing your biscuit tin, let’s give it its original address.
A small study with unusually close observation
In this 2019 NIH trial, 20 adults stayed at a research centre. Each ate an ultra-processed diet and an unprocessed diet for two weeks, in randomized order. They could eat as much or as little as they wanted. Read Hall and colleagues’ full paper.
Participants consumed an average of 508 more kcal per day on the ultra-processed diet (reported standard error ±106 kcal/day). They gained about 0.9 kg during that period and lost about 0.9 kg during the unprocessed period. The food offered was designed to match several nutritional characteristics; the actual food eaten differed.
A comparison, not a calorie surcharge
What the design can and cannot do
Random allocation strengthens the comparison, but twenty people over four weeks cannot settle long-term health effects or identify one ingredient responsible. Eating rate and diet composition differed. Participants were not shopping or cooking at home. The study had NIH support; one author disclosed food-industry relationships. The full paper lists these details.
Please do not add 508 to your journal
Suppose you log a packaged snack. The number to record is your best estimate for the amount you ate, using the relevant label or food entry. The trial’s average difference is not a hidden tax to add afterwards. Doing that would mix two completely different measurements: the energy in your portion and the difference in total intake between experimental diets.
The same applies in reverse. A home-cooked meal does not automatically earn a 508-calorie discount. Your recipe may include a generous amount of oil; someone else’s may not. Food descriptions are not magical accounting rules. Portions and ingredients still belong in the calculation.
A useful question for your own routine
Our interpretation is to look at the setting around a meal, without borrowing a numerical prediction. Was it a proper lunch, or a succession of handfuls while you answered messages? Did you record the amount before the packet disappeared? These are practical questions about recording and routine, not outcomes established by this study.
Try making one usual meal easier to describe. “Dal, rice and curd” is a start. Adding the type of dal, a measured portion of rice and the amount of curd makes the record more useful. You do not need an elaborate spreadsheet for every lunch. Pick one recurring uncertainty and make it a little less uncertain.
Why a landmark paper still needs company
An influential study is a beginning, not the final episode. Later trials can test different people, settings and menus. A similar result can build confidence; a different one can help reveal which conditions mattered. That is why a good research explainer gives you the study year and design rather than presenting an old result as today’s breaking news.
Our companion article looks at a 2025 trial conducted outside an inpatient ward. Read the two as different experiments, not duplicate confirmations of a universal calorie rule. Neither means every food sold in a packet has the same effects.
Keep the takeaway human-sized
A sensible takeaway is curiosity about your usual meals, not fear of your pantry. A journal can help describe food; it cannot prove what caused a change in your body. If tracking makes eating more stressful, step back and seek appropriate support. This explainer is general education and does not set calorie targets or treatment plans.
Sources & further reading
Cover art is AI-generated conceptual illustration; it does not depict actual study meals. Diagrams are original PoshanSense graphics. Paper figures and publisher images are not reproduced. Examples and practical interpretations are identified in the text.
Published 5 October 2026 · Sources checked 5 October 2026. Suggest a correction.