Observations

The MHDR is about Observations!

The observation_fact table is where all the actual data (“facts”) about the patient population are stored.  This is the largest table in the i2b2 system, and the most vital to the function of the system.

The standardized i2b2 data structure allows any type of information to be stored in the observation_fact table.

observation_fact
Variable	Type
ENCOUNTER_NUM	int
PATIENT_NUM	int
CONCEPT_CD	varchar
PROVIDER_ID	varchar
START_DATE	datetime
MODIFIER_CD	varchar
INSTANCE_NUM	int
VALTYPE_CD	varchar
TVAL_CHAR	varchar
NVAL_NUM	decimal
VALUEFLAG_CD	varchar
QUANTITY_NUM	decimal
UNITS_CD	varchar
END_DATE	datetime
LOCATION_CD	varchar
OBSERVATION_BLOB	text
CONFIDENCE_NUM	decimal
UPDATE_DATE	datetime
DOWNLOAD_DATE	datetime
IMPORT_DATE	datetime
SOURCESYSTEM_CD	varchar
UPLOAD_ID	int
authority	varchar

The data in the MHDR is updated on a monthly basis.  Information on the types and frequency of data can be found here.


Changes in data collection methods or priorities within the hospital over time will impact the amount of data available for a particular concept from year to year. These fluctuations in data density significantly impact outcomes measures, resulting in apparent “trends” that may be artifacts of the data.

The interactive chart above contains patient counts, concept counts, and concepts per patient counts for Adult Primary Care patients for concepts utilized in the domains of public health modeled within the MHDR. Researchers are encouraged to explore these trends in data quantity when evaluating measures generated from the MHDR HOME Cell.