Provenance

NCQA Comment on Signature: For USCDI v5

1. Signature:

NCQA supports the advancement of the Provenance signature element for eventual USCDI inclusion. Signature is an important element to address data integrity use cases and support trust in the electronic health information (EHI) exchanged, which is essential to meeting the goals of interoperable data exchange. Specifically, provenance signature can support data audit processes and data verification. Signature is established in CDA and is supported by the US Core FHIR IG Provenance profile.

HL7 FHIR Record Lifecycle Event Implementation Guide

Please add/update reference to HL7 FHIR Record Lifecycle Event Implementation Guide, published December 2023:  http://hl7.org/fhir/uv/ehrs-rle/Informative1/

Provenance Use Cases

Provenance was a field added to USCDI with a lot of hope. Unfortunately the role out of the fields have not added a lot of value as they remain lacking the specificity to truly understand the provenance of a given field/item. I think the problem may be that the important components of provenance may differ with different types of health care data. For example author may be critical for a note but may not make sense for a vital sign or an immunization. Provenance for a problem list item that has been updated over time may be much tougher to understand than it would be for an immunization that was given once. 
 

I would recommend shifting the provenance paradigm and rather than focusing on a singular model for all data types, focus more deeply on a given data type and solve it for that type and then advance to the next priority. USCDI should move the data provenance to the data fields it applies to and then solve for that use case. Perhaps first start with clinical notes and immunizations. They have different provenance needs but both are solvable. Then once that is demonstrated can continue to other pami data or vital signs for example. 

PACIO Comments on Provenance

The PACIO Project workgroup offers the following comments on the USCDI V4: 

• Data Class: Provenance (Draft V4) 

• Data Elements: Author (Level 2), Signature (Comment), Author Roles (Comment), and Purpose of Capture (Comment) 

• Recommendation: Include data element “Author” in USCDI V4 and advance data elements, “Signature,” “Author Roles,” and “Purpose of Capture” to level 1. 

• Rationale: The PACIO (Post-Acute Care Interoperability) Project, established February 2019, is a collaborative effort between industry, government, and other stakeholders, with the goal of establishing a framework for the development of FHIR implementation guides to facilitate health information exchange. The PACIO Community recommends the following data elements that are currently at other levels should be upgraded. 

  • Author (currently on level 2) - Author time and organization is not going to have a lot of meaning without the author particularly in non-institutional data sources. As we move to more patient centered care, there will be other contributors of data including patients and non-clinician Caregivers 

  • Signature (Currently Comment) – Some documents and other information, such as end of life, like POLST, orders need to be signed in order to trusted and used. Without a signature there is no way to validate the veracity of data that may not be coming from a direct trusted source. 

  • Author Role(s) (Currently Comment) – The role in which data is captured is important to more fully understand data. It is important to know not only what organization or author created the data but the capacity in which they were operating under in order to properly understand the data.

  • Purpose of Capture (Currently Comment) – Information is gathered from numerous sources for their own specific purposes. The level of detail, completeness, and quality of the information is going to be highly dependent on the interests of those capturing the information. This is important to understand more about the data and how it can be further used for things such as population and public health. 

PACIO Comments on Provenance

The PACIO (Post-Acute Care Interoperability) Project, established February 2019, is a collaborative effort between industry, government, and other stakeholders, with the goal of establishing a framework for the development of FHIR implementation guides to facilitate health information exchange.

• The PACIO Community recommends Elements that are currently at other levels should be upgraded.
  • Author (currently on level 2) - Author time and organization is not going to have a lot of meaning without the author particularly in non-institutional data sources. As we move to more patient centered care, there will be other contributors of data including patients and non-clinician Caregivers
  • Signature (Currently Comment) – Some documents and other information, such as end of life, like POLST, orders need to be signed in order to trusted and used. Without a signature there is no way to validate the veracity of data that may not be coming from a direct trusted source.
  • Author Role(s) (Currently Comment) – The role in which data is captured is important to more fully understand data. It is important to know not only what organization or author created the data but the capacity in which they were operating under in order to properly understand the data.
  • Purpose of Capture (Currently Comment) – Information is gathered from numerous sources for their own specific purposes. The level of detail, completeness, and quality of the information is going to be highly dependent on the interests of those capturing the information. This is important to understand more about the data and how it can be further used for things such as population and public health.

HL7 Patient Empowerment PCD WG USCDI v3 feedback re: Provenance

USCDI consists of categories of data elements representing information captured from different systems as well as by individuals. However, attribution of information created or supplied by an individual, as a patient or other role, is not reflected within the Provenance data class in the latest USCDI draft version 3 consisting of the following two (2) data elements: Author Time Stamp and Author Organization. We recommend at minimum the Author data element classified as Level 2 be added to USCDI v3. 

In addition to USCDI v3 Provenance not specifying Author, the data class is not fully inclusive of information created outside an organization (or changes to information such as updates). To better understand the origin and changes to data that can occur with or without exchange, and improve Provenance with context that identifies the type of actor or system creating or updating data, we recommend two (2) additional data elements: 

1. Author Role to clarify the type of actor, such as patient, especially where information is contributed by individuals. Author should also accommodate device-generated data such as wearables. 
2. Updates to understand whether data has changed either from the point of creation or in exchange

We also suggest reframing Author Organization to be location-agnostic or account for environments outside an organization which may include a personal device or system.

COVID-19 pushed care delivery outside organizations with a boom of consumer-facing technologies to enable remote care and support emergent public health priorities. The pandemic also highlighted the need to understand factors outside of healthcare settings that have a large effect on health status (and outcomes). However, the current USCDI does not explicitly represent health and other data that might be contributed by individuals outside health care organizations, in spite of introducing new data classes and elements often dependent on individual/patient input. Several current or potential future data classes contain elements often asserted or informed by the patient including (but not limited to):

• Demographics (include identity, race and ethnicity, sexual orientation and gender identity, and other data)
• Health Insurance Information
• Goals (Patient and SDOH)
• Problems (includes SDOH Problems/Health Concerns and dates from diagnosis and resolution)
• Health Status (Includes Health Concerns, Functional Status, Disability Status, Mental Function, and Smoking Status)
• Clinical Tests (currently only reflects those performed in lab though perhaps could be expanded to include self-administered at-home tests)
• Observations
• Assessments and Plan of Treatment (includes SDOH assessment)
• Advance Directives (not in USCDI v3)
• Vital signs (e.g. where capture from remote sensing devices)

We see patient contributed data which is person-generated or self-reported as equivalent to clinical and administrative data recorded by health care professionals or systems. Therefore supporting provenance that captures the most granular metadata associated with data creation in USCDI is foundational for all information contributed by patients to be recognized and trusted. 

As ONC continues to prioritize health equity and patient engagement, the importance of data provenance will only increase as USCDI evolves and information used to identify, assess, treat, coordinate, and measure care increasingly relies on patient generated or contributed information. Updating the USCDI to ensure instances where data originate or are reported by individuals are appropriately attributed is a first step in ensuring PCD is recognized, represented, and leveraged to promote equity, participator care, and support broader national health care and interoperability goals.

Thank you for considering HL7's Patient Contributed Data (PCD) Committee (Sub-Workgroup of the HL7 Patient Empowerment Workgroup) comments in finalizing the USCDI v3.

Clinovations GovHealth USCDI Draft V2 Comments: Provenance

Considerations for provenance attributed to discrete structured data

The Author Time Stamp and Author Organization data elements are important for payloads exchanged, but they are also critical for the discrete, structured data that reside in exchanged payloads. We raise the consideration of how to evolve the USCDI to support attribution to a specific provider or organization through provenance retained for specific data that may be reconciled.  Without the capability to see the original author time stamp and organization for problems, medications, allergies, and other structured data, providers are likely to continue experiencing hesitation to retain provenance or insight into whether external data is reconciled. Today, practices vary across health IT developers with regard to reconciled data. Some solutions attribute reconciled data with the individual performing the reconciliation, while others retain the provenance / attribution of the external data.

In addition, when generating a C-CDA or using FHIR to exchange USCDI data within an enterprise, multi-specialty organization to send or make available for a 3rd party, will provenance be at the document level for all problems and medications while different specialties may be managing different conditions or medications?