300 adults will put an awkward food-label question to the test: does ‘ultra-processed’ soy lower blood pressure as well as tofu?

Calling a food ultra-processed can be a useful warning, but it can also put nutritionally different products in the same box. A trial design published on 30 October 2026 sets up a direct test of that problem. The SWITCH trial plans to recruit 300 adults in Canada who have both hypertension and obesity; half of the participants will also have type 2 diabetes.

The researchers will randomly assign people to one of three groups for 12 weeks. One group receives usual care. The other two follow a digital programme that combines elements of the DASH and Portfolio diets, including extra fruit and vegetables and at least four daily servings of soy foods. The crucial difference is the form of that soy.

The high-ultra-processed group will have at least two servings of soy milk and two servings of foods such as soy yoghurt, soy burgers or soy mince each day. The lower-processing group will have at least two servings of minimally processed soy milk, plus a serving of edamame or roasted soy nuts and a serving of tofu or tempeh. The main outcome is the change in systolic blood pressure; non-HDL cholesterol and body weight are key secondary outcomes.

This is both a superiority and a non-inferiority trial. First, the combined diet programme will be tested against usual care to see whether it reduces blood pressure. Second, the researchers will ask whether the high-ultra-processed soy version performs no worse than the lower-processing version by more than a pre-set margin. That is a more precise question than asking whether every ultra-processed food is good or bad.

The design also prevents a common misunderstanding. A soy burger and a sugary drink may both fall into NOVA group 4, yet their protein, fibre, fat, salt and micronutrient profiles can be very different. If the two soy patterns produce similar results, that would support judging foods by more than one processing label. If they differ, the trial may help identify whether processing level adds useful information even within a carefully planned diet.

No health result has been reported. The paper describes what the researchers intend to do, and the registry lists primary completion for June 2027. The programme also changes several things at once — soy foods, fruit and vegetable intake, digital education and support — so even the finished trial will need careful interpretation. For now, it is a useful example of research designed to replace a broad nutrition argument with a measurable comparison, not a reason to change treatment or eat four servings of soy a day on your own.