Create XSLT for non-relevant data

[pat-johnson-dev]

Please complete following implementation steps from RFC for protecting PHI from CCD data ingestion:

• TechBD will analyze CCD file structures to identify screening data tags (e.g., , ).
• Create an XSLT file with transformation rules to exclude non-relevant tags.
• Test the XSLT file with multiple CCD examples to ensure correct functionality.
• Publish the XSLT file on TechBD’s public GitHub repository.

See this link for details.

[cibijacob]

I spoke with @shah and have a general understanding of the requirements. I have started reviewing CDA.xsl to get an understanding of the current CDA specification as it pertains to NYHER Screening data in XML files that QEs will send. I should have the first version of the draft XSLT spec by Friday or Monday.

[cibijacob]

1. The CCD file structures were analyzed to identify the following patient and screening data tags:

• Patient (recordTarget/patientRole)
• Organization (author/assignedAuthor/representedOrganization)
• Encounter (componentOf/encompassingEncounter)
• Consent (authorization/consent)
• Observation (component/structuredBody/component)

2. A sample XSLT file was developed to extract the specified sections (Patient, Organization, Encounter, and Observation).
3. A channel was configured in Mirth Connect to test the functionality of the XSLT file.

[cibijacob]

1. Updated the XSLT file to extract the Consent section and header details.
2. The XSLT file is tested with multiple CCD files as well as CCD files with sections other than the specified sections (Patient, Organization, Encounter, Consent and Observation).
3. The XSLT file cda-phi-filter.xslt is saved in the GitHub repository under the folder support/specifications/ccda.

[shah]

@cibijacob please check all of these tags to ensure that only the subtags and attributes used by NYHER are allowed while non-NYHER screening subtags are filtered:

• Encounter
• Observation

As you know many different types of encounters can be covered under encouter tag and same for observation (for example, labs, and device data can come in observation that should be filtered).

Please explain in this ticket how you're properly filtering only NYHER content.

[cibijacob]

Hi @shah,

The Observation (component/structuredBody/component) can again to be filtered out to extract the Social History (Code : 29762-2) which includes the questionnaire, along with the Patient, Organization and Consent details.
And the following sub sections can be excluded.
ALLERGIES AND ADVERSE REACTIONS : 48765-2
MEDICATIONS : 10160-0
PROBLEMS (Active Problems) : 11450-4
PROCEDURES : 47519-4
RESULTS : 30954-2
VITAL SIGNS : 8716-3
ENCOUNTER DETAILS : 46240-8
PROGRESS NOTES : 10164-2
PLAN OF TREATMENT : 18776-5
VISIT DIAGNOSES : 51848-0
CARE TEAMS : 85847-2
FAMILY HISTORY : 10157-6
IMMUNIZATIONS : 11369-6
MEDICAL EQUIPMENT : 46264-8
INSURANCE PROVIDERS : 48768-6