DiseaseEpisodes
Episodes are clinically and analytically relevant episodes of disease and treatment. Examples include first disease occurrence, remission, progression, stable disease, treatment course, treatment cycle, treatment regimen. Disease and treatment episodes are represented in the Episode table (link to the table). If known, connection between disease/treatment episodes and lower level events (conditions, procedures, drugs) can be stored in the Episode_Event table (link to the table). More details are available here: https://github.com/OHDSI/OncologyWG/wiki/CancerModelsRepresentation.
Episodes can be derived algorithmically pre- or post-ETL or extracted from the source data directly (e.g. Tumor Registry). Currently available is post-ETL derivation of first disease occurrence and first treatment and algorithmic derivation of chemotherapy regimens from individual medication records. They are described in this section. We seek data scientists and developers to develop advanced tools for derivation of disease and treatment episodes and outcomes from lower level events.
Derivation of the first disease occurrence depends on the data source. The algorithm described below is for post-ETL derivation from the US Tumor Registries given that the first occurrence is explicitly marked as such.
-
Add an Episode record from Condition_Occurrence based on the First Occurrence modifier
a) Link patient’s Condition_Occurrence record and to the Measurement record with
Measurement_Concept_ID = 32528 (‘First disease occurrence’) and
Measurement_Type_Concept_ID = (‘Tumor registry’)
b) Add an Episode record and populate it from the following fields from the Condition_Occurrence record identified in Step 1a:
i. condition_person_ID -> episode_person_ID
ii. condition_start_datetime -> episode_start_datetime
iii. condition_concept_id -> episode_object_concept_id
iv. condition_type_concept_id -> episode_type_concept_id
v. 32528 (‘Disease First Occurrence’) -> episode_concept_id
vi. 1 -> episode_number
-
Add an Episode_Event record to connect Episode and Condition_Occurrence records from Step 1 and populate it as follows:
i. episode_id -> episode_id
ii. condition_occurrence_ID -> episode_event_event_id
iii. and ‘1147127’ that is a concept for 'condition_occurrence.condition_occurrence_id' -> episode_event_field_concept_id
-
Add a set of “clean” Episode modifiers into Measurement table based on the Condition_Occurrence modifiers. There should be only one diagnostic modifier of a kind associated with a disease episode. If there are multiple modifiers (e.g. repeated measurements of the same modifier) associated with a condition occurrence, only one modifiers should be selected. It could be based on the most recent date, data source, or other information.
a) Link patient’s Episode_Event, Episode, Condition_Occurrence and Measurement records
b) Add new Measurement records with only one modifier of a kind from the existing condition Measurement records and copy all the respective fields except for:
i. episode_id -> modifier_of_event_id
ii. and ‘TBD’ that is a concept for 'episode.episode_id' -> modifier_field_concept_id
