Fig. 1.

Comparison of restaurant-level policy consistency of menus and orders before and after kids’ meal beverage policy went into effect in California (n 78–110 quick-service restaurants) and Wilmington, Delaware (n 8–9 quick-service and five full-service restaurants).^a,b

^aMissing responses at pre- or post-policy for inside the restaurant and/or in the drive-through window were excluded from analytic sample resulting in variations in sample size in California. n/a signifies that in Wilmington, data from orders were not collected pre-policy and are reported for inside the restaurant only post-policy.

^bMultilevel logistic regression was used to compare pre-post changes in California and for Wilmington, Delaware, adjusting for clustering by chain by and adjusting for restaurant type (quick-service or full-service) and presence of drive-through. **P < 0·01, ****P < 0·0001.

^cMenus consistent with policy if only plain water and/or unflavoured milk with kids’ meal(s) shown on menu boards in California, and if only plain water, any milk and/or any unsweetened juice shown on menus/menu boards in Wilmington, Delaware.

^dOrders consistent with policy if cashier/server(s) initially offer only plain water and/or unflavoured milk in California and only plain water, any milk and/or any unsweetened juice in Wilmington, Delaware.

^ePre-policy data were collected within 1 month prior to the policy going into effect (December 2018–January 2019).

^fPost-policy data were collected 9–12 months after the policy went into effect (August 2019–December 2019).