Direct answer: A restaurant saying that it offers gluten-free items does not, by itself, show how gluten cross-contact is controlled. A 2026 study in Winnipeg, Canada, found that many restaurants reported gluten-free options and many also reported at least one cross-contact control. The study did not test meals or observe practice, so its results describe reported availability and procedures, not verified meal-level safety [1].
What the 2026 study measured
Defries and colleagues surveyed staff by telephone at 548 restaurants in Winnipeg. They asked whether the restaurant offered gluten-free items, whether those items cost more, and whether the restaurant used methods intended to limit gluten cross-contact [1]. This was an observational, cross-sectional study of restaurant reports in one Canadian city.
Sixty-nine percent of the restaurants reported offering gluten-free items. Among restaurants offering them, 80% reported using methods to limit gluten cross-contact. When the researchers classified reported methods, cleaning was reported by 58%, separate kitchenware by 55%, and a dedicated gluten-free preparation area by 13%. Most restaurants that reported a control described only one of the three categories [1].
These numbers are useful for understanding what restaurants say they do. They do not establish the gluten content of any meal. Staff responses were not independently checked against kitchen observations, records, or laboratory testing [1].
Why the distinction matters
For people with celiac disease, a medically necessary gluten-free diet involves both ingredient verification and control of unintended gluten cross-contact. A menu label can answer part of the ingredient question, but it cannot describe every preparation step.
Experimental and field research also shows why a single generic statement such as “we clean” is not enough to evaluate a process. Tóth and colleagues sampled surfaces in 39 school catering kitchens in Hungary before and after an intervention involving training and food-handling changes. Gluten-positive surface samples decreased after the intervention, but the study was conducted in school kitchens and surface results do not measure what a restaurant diner consumes [2].
Parsons and colleagues experimentally tested several common preparation practices. Their findings varied by food and procedure rather than supporting a rule that all shared equipment is always unsafe or always acceptable [3]. This supports a process-specific approach: identify the actual contact pathway, the control used, and the evidence available for that control.
What diners can ask
- Which ingredients and components of this dish are checked for gluten?
- What parts of preparation are shared with gluten-containing food?
- What specific control is used for those shared steps?
- Is the answer based on a current procedure, staff memory, or a documented check?
- What does the restaurant do when a supplier, ingredient, or preparation method changes?
A clear answer can include limits. For example, a restaurant may accurately explain that it uses separate utensils but does not have a dedicated preparation area. That information is more useful for individual decision-making than an unsupported assurance that the meal is “completely safe.”
What food-service teams can document
- The current recipe and supplier information for each gluten-free menu item.
- The preparation steps where gluten-containing food is present.
- The control used at each relevant step, such as cleaning, separation, or dedicated equipment.
- Who checks that the control is followed and when the procedure was last reviewed.
- What staff should say when a control cannot be confirmed.
GCSI interpretation: the 2026 Canadian study is evidence about reported restaurant practices, not a validation study. GCSI therefore treats a reported control as a useful starting point that should be described more specifically before it is treated as documented or observed evidence.
Remaining uncertainty
The Winnipeg results should not be generalized to all Canadian restaurants or other countries. The survey covered one urban centre, relied on telephone responses, and did not test food. The experimental studies cited here also tested particular foods, surfaces, and procedures. None can guarantee the outcome of a future restaurant meal. The practical aim is narrower: make the preparation process visible enough that diners and food-service teams can identify what is known, what is controlled, and what remains uncertain.