OBJECTIVES: The content validity of an instrument depends not only on the concepts embodied in the items but also on how the items are structured to elicit responses from patients. This research explores the “grammar” of individual items and how it varies across a number of instruments in selected disease areas.  The goal is to understand how such considerations affect consistency of content and to classify items by concept. METHODS: The structure of each item is characterized as an item stem with a core concept, with an implicit or explicit context (e.g., a disease), event (e.g. “felt frustrated or impatient”), and stimulus (e.g., “about your symptom”), as well as the recall period and response options. Concepts are classified using the WHO International Classification of Functioning (ICF). Similarities and differences across instruments within disease areas are analyzed. RESULTS: We decomposed over 600 items in at least 23 instruments across 5 disease areas and several generic instruments, capturing and classifying each aspect of the structure of each item.  Most physical function items could be matched with specific 3-4 digit ICF codes; most emotional function items could not be matched as specifically. There was considerable variation across instruments regarding the explicit statement of context as well as the presence of a stimulus.  We observed at least 8 different recall periods ranging from an implied present to “in the past year” to “in 10 years” with distinct patterns by disease area. We observed at least 9 types of response options, but the majority of items used 5-point scales. CONCLUSIONS: There are some commonalities but little standardization in how items are structured, within or across diseases. Classification and comparison of items and evaluating comparative content validity is complicated by the variation in most aspects of how the items are phrased.