Table 1.

Explanations of variable attributes: data type, scope, and dependency

Variable Attributes	Explanation
Data Type	Standard data formats such as integer, float, boolean, timestamp, or text. "Indeterminate" for non-standard cases (e.g., eligibility based on clinical intuition)
Scope	Defines how data are aggregated across notes
Many Per Note	Variable appears multiple times within a single clinical note (e.g., medication administrations)
One Per Note	Variable is aggregated to a single value within a single clinical note (e.g., highest recorded temperature in a daily progress note)
One Per Patient	Variable is aggregated across multiple notes for a patient (e.g., whether a patient has ever had cancer)
Dependency	Defines whether a variable is independent or dependent
Independent	Extracted directly from clinical text using logical instructions (e.g., date of primary tumor surgery)
Dependent	Computed based on independent or other dependent variables using logical operations, without direct text extraction (e.g., time between definitive treatment and first cancer recurrence)