wiki:i2b2 Data Import PATS Mapping v1

Version 8 (modified by Richard Bramley, 12 years ago) ( diff )

--

Back

The PATS mappings are split into text and numeric fields.

Text Fields

Text fields are mapped to the i2b2 Observation Fact table like this

i2b2 FieldSource
Encounter_Numi2b2 Patient ID
Patient_Numi2b2 PatientID
Concept_CdEnum lookup value (see below)
Provider_Id"@"
Start_DateThe source table observation date field (see below)
Modifier_Cd"@"
ValType_Cd"T" (text)
Units_Cd"@"
Location_Cd"@"
Update_DateThe current date using the GETDATE() Function
SourceSystem_Cd"BRICCS"
Upload_Id"1"

SOMETIME (RB): Convert Update_Date to use a date parameter. It's just a little quicker.

SOMETIME (RB): Should the SourceSystem_Cd not be "PATS"? Or maybe "MINAP" or whatever.

SOMETIME (RB): The Upload_Id is presumably always one because everything is reloaded each time. Might be nice to change this to something more meaningful.

Enum Lookup Value

The Concept Code to use is found in the table DATA_DICTIONARY_ENUM_VALUES in the DWBRICCS database on UHLDWH server. This table contains records for concept code keyed on source table name, column name and column value. NULL column values may be mapped to the field '(empty)'.

ACTION (RB): Check if all possible values of column value contain a mapping in the DATA_DICTIONARY_ENUM_VALUES table.

Source Table Observation Date Field

For each source table in the there is a single field that is used as the start date for all columns from that field. The observation date field to use for each table is stored in the pats_tables table in the DWBRICS on the UHLDWH server.

Numeric Fields

Numeric fields are mapped to the i2b2 Observation Fact table like this

i2b2 FieldSource
Encounter_Numi2b2 Patient ID
Patient_Numi2b2 PatientID
Concept_Cd"PTS:{TABLE_NAME}_{COLUMN_ID}"
Provider_Id"@"
Start_DateThe source table observation date field (see above)
Modifier_Cd"@"
ValType_Cd"N" (numeric)
TVal_Char"E" (see below)
Units_CdValue calculated by ufn_get_units function (see below)
Location_Cd"@"
Update_DateThe current date using the GETDATE() Function
SourceSystem_Cd"BRICCS"
Upload_Id"1"

TVal_Char

The TVal_Char describes the meaning of the value contained in the NVal_Num field.

For all values currently the value of TVal_Char is set to "E", which means equals. That is, the value in the NVal_Num field is the result of the test. All possible values of TVal_Char are:

ValueMeaning
EEquals
NENot Equal
LLess Than
LELess Than or Equals
GGreater Than
GEGreater Than or Equals

From a casual look at the data, I think that a value of less than X, or greater than Y is sometimes used for Pathology data.

ufn_get_units Function

The function ufn_get_units in the DWBRICCS database is called with parameters of table_name and field_name to determine the units.

ACTION (RB): Work out what this function actually does and then move the logic into a join if possible. This should be clearer and also quicker.

Note: See TracWiki for help on using the wiki.