2019-08-21 AUG

From Cloud to Mobile

reviewing DCEG Connect architectures and applications

...

...

Example of a one input field being enough for address validation using GCP Place services:

· Lines of data vs complete data for an individual

· Specific to version 1

· Theoretically, can get details on metadata and lists of Patients, Observations, and Encounters, but couldn’t get API calls to work as described on GitHub https://github.com/synthetichealth/syntheticmass/blob/master/docs/using-fhir.md

· Additional information about Synthetic Mass and the program that generated it are found here https://github.com/synthetichealth

· Very large nested compressed files in multiple formats that take a very long time to extract

· No documentation on organization of compressed folders

· Based on extraction so far JSON and XML files are one file per person per data type (demographic, allergies, etc.)

o This is why demo was either for a sample of 100 patients or single files per patient

· Must then be uploaded to GCP

· Cannot be used with BigQuery without first converting to newline delimited JSON

· Permissions may be affected by account access vs email access of GCP

o May be lag or due to linking nih account when setting up options, since I have access others should experiment.

· No documentation in the FHIR Demo Dataset Google Group, just access to Synthetic Mass data on GCP

· Available folders:

o One set appears to be the 1K Sample FHIR STU3 also available on the download page

o Unclear what r4 refers to