kumc-bmi/naaccr-tumor-data

discrete table including names, MRNs, and accession numbers

Closed this issue · 2 comments

dckc commented

one site wants a table that includes this PHI
fields.csv doesn't have any PHI fields
tumor_item_type.csv has MRN and accession number but
name is in the Patient-Confidential section, which is completely left out.

dckc commented

some progress: b9b197f I added PATIENT_ID_NUMBER_N20, PATIENT_SYSTEM_ID_HOSP_N21 for patient mapping.

dckc commented

As of d20cb0d, you can load-layouts and then pick out the relevant substring of the raw record from observation_blob. For example, to get accessionNumberHosp:

select layoutVersion, layout.naaccrItemNum, layout.name, shortLabel, startPos, length
     , substr(tumor.observation_blob, startPos, length) value
from tumor
join layout on layout.layoutVersion = tumor.naaccr_record_Version_n50
where layout.name = 'accessionNumberHosp'