Table 2. Open in new tab Pathways for...

Onboarding new use cases
Intake and onboarding	Users submit project intake request to be onboarded to HDAP.
Data access	Currently available notes	Onboarded users have immediate access to the centralized and cleaned notes that we currently process (TIU, pathology, radiology) (from step 1).
New types of notes	HDAP already has daily updates of all major free text notes in VA; however new types of notes can also be uploaded, if external, or pulled into our daily ETL processes, if they already exist elsewhere in our systems.
Data annotation	A supported data annotation environment (including eHOST) is available for use across projects with set-up and access supported by our internal HDAP data science team. Scripts and workflows are also available for document prepping, content loading and schema conversion, to send data to and from eHOST schemas. For the SDoH use case, a public website is used to support annotation and provides guidelines and concept definitions. This is the current link: http://ec2-18-206-230-88.compute-1.amazonaws.com/wordpress/?page_id=556
Model training	We provide code and workflows for model training and fine-tuning from data annotation outputs, as well as a range of open-source models that have been brought into HDAP.
Productionizing new use cases	Current downstream locations (HDAP and CDW)	When text processing algorithms are ready for productionization, they can be included in the current daily model running workflows and ETLs, with support from our HDAP data science team.
New downstream operational locations	When necessary, additional ETLs to new locations can be established by the HDAP data science team.

Onboarding new use cases

Intake and onboarding

Users submit project intake request to be onboarded to HDAP.

Data access

Currently available notes

Onboarded users have immediate access to the centralized and cleaned notes that we currently process (TIU, pathology, radiology) (from step 1).

New types of notes

HDAP already has daily updates of all major free text notes in VA; however new types of notes can also be uploaded, if external, or pulled into our daily ETL processes, if they already exist elsewhere in our systems.

Data annotation

A supported data annotation environment (including eHOST) is available for use across projects with set-up and access supported by our internal HDAP data science team. Scripts and workflows are also available for document prepping, content loading and schema conversion, to send data to and from eHOST schemas. For the SDoH use case, a public website is used to support annotation and provides guidelines and concept definitions. This is the current link: http://ec2-18-206-230-88.compute-1.amazonaws.com/wordpress/?page_id=556

Model training

We provide code and workflows for model training and fine-tuning from data annotation outputs, as well as a range of open-source models that have been brought into HDAP.

Productionizing new use cases

Current downstream locations (HDAP and CDW)

When text processing algorithms are ready for productionization, they can be included in the current daily model running workflows and ETLs, with support from our HDAP data science team.

New downstream operational locations

When necessary, additional ETLs to new locations can be established by the HDAP data science team.

Onboarding new use cases
Intake and onboarding	Users submit project intake request to be onboarded to HDAP.
Data access	Currently available notes	Onboarded users have immediate access to the centralized and cleaned notes that we currently process (TIU, pathology, radiology) (from step 1).
New types of notes	HDAP already has daily updates of all major free text notes in VA; however new types of notes can also be uploaded, if external, or pulled into our daily ETL processes, if they already exist elsewhere in our systems.
Data annotation	A supported data annotation environment (including eHOST) is available for use across projects with set-up and access supported by our internal HDAP data science team. Scripts and workflows are also available for document prepping, content loading and schema conversion, to send data to and from eHOST schemas. For the SDoH use case, a public website is used to support annotation and provides guidelines and concept definitions. This is the current link: http://ec2-18-206-230-88.compute-1.amazonaws.com/wordpress/?page_id=556
Model training	We provide code and workflows for model training and fine-tuning from data annotation outputs, as well as a range of open-source models that have been brought into HDAP.
Productionizing new use cases	Current downstream locations (HDAP and CDW)	When text processing algorithms are ready for productionization, they can be included in the current daily model running workflows and ETLs, with support from our HDAP data science team.
New downstream operational locations	When necessary, additional ETLs to new locations can be established by the HDAP data science team.

Table 2.

Open in new tab

Pathways for new use cases.

Onboarding new use cases
Intake and onboarding	Users submit project intake request to be onboarded to HDAP.
Data access	Currently available notes	Onboarded users have immediate access to the centralized and cleaned notes that we currently process (TIU, pathology, radiology) (from step 1).
Data access	New types of notes	HDAP already has daily updates of all major free text notes in VA; however new types of notes can also be uploaded, if external, or pulled into our daily ETL processes, if they already exist elsewhere in our systems.
Data annotation	A supported data annotation environment (including eHOST) is available for use across projects with set-up and access supported by our internal HDAP data science team. Scripts and workflows are also available for document prepping, content loading and schema conversion, to send data to and from eHOST schemas. For the SDoH use case, a public website is used to support annotation and provides guidelines and concept definitions. This is the current link: http://ec2-18-206-230-88.compute-1.amazonaws.com/wordpress/?page_id=556
Model training	We provide code and workflows for model training and fine-tuning from data annotation outputs, as well as a range of open-source models that have been brought into HDAP.
Productionizing new use cases	Current downstream locations (HDAP and CDW)	When text processing algorithms are ready for productionization, they can be included in the current daily model running workflows and ETLs, with support from our HDAP data science team.
Productionizing new use cases	New downstream operational locations	When necessary, additional ETLs to new locations can be established by the HDAP data science team.

Onboarding new use cases
Intake and onboarding	Users submit project intake request to be onboarded to HDAP.
Data access	Currently available notes	Onboarded users have immediate access to the centralized and cleaned notes that we currently process (TIU, pathology, radiology) (from step 1).
Data access	New types of notes	HDAP already has daily updates of all major free text notes in VA; however new types of notes can also be uploaded, if external, or pulled into our daily ETL processes, if they already exist elsewhere in our systems.
Data annotation	A supported data annotation environment (including eHOST) is available for use across projects with set-up and access supported by our internal HDAP data science team. Scripts and workflows are also available for document prepping, content loading and schema conversion, to send data to and from eHOST schemas. For the SDoH use case, a public website is used to support annotation and provides guidelines and concept definitions. This is the current link: http://ec2-18-206-230-88.compute-1.amazonaws.com/wordpress/?page_id=556
Model training	We provide code and workflows for model training and fine-tuning from data annotation outputs, as well as a range of open-source models that have been brought into HDAP.
Productionizing new use cases	Current downstream locations (HDAP and CDW)	When text processing algorithms are ready for productionization, they can be included in the current daily model running workflows and ETLs, with support from our HDAP data science team.
Productionizing new use cases	New downstream operational locations	When necessary, additional ETLs to new locations can be established by the HDAP data science team.

This Feature Is Available To Subscribers Only