In review of the FHIR R5 NutritionIntake resource, a few concerns include:

• Definition of NutritionIntake.code is "Code representing an overall type of nutrition intake" - the example value set suggests this element addresses the "diet" applicable for the NutritionIntake event (e.g., tube feeding, enteral feeding) but it seems to address the methodology for the intake and clearer definition will help. A separate element would be required to differentiate a specific "diet" (e.g., exclusive breast milk, diabetic diet).
• ConsumedItem.type and ConsumedItem.nutritionProduct each has a cardinality of 1..1. That means that every entry must address a specific product and the item description (type).  The term "type" seems to be incorrect for the concept defined by the example value set of consumable items.  Type would imply a different meaning.  Also, the decision as to whether full reference to a NutritionProduct should depend on the use case. That means that both ConsumedItem.type AND ConsumedItem.nutritionProduct should have cardinality of *..1 rather and 1..1. Current modeling may be useful for a dietician, but use to document a nursing Intake template will often just indicate the consumed item without the need for a full reference to the details of the nutrition product. A cardinality of *..1 allows flexibility for the use case without overloading all use cases with superfluous information.
• The value set Binding: Types of Edible Substances (Example) should be included in THO and perhaps we could suggest a better name consistent with the SNOMED hierarchy such as Edible Substances