Data Provenance All Hands Community Meeting May 21 st, 2015

Meeting Etiquette Please mute your phone when you are not speaking to prevent background noise. – All meetings are recorded. Please do not put your phone on hold. – Hang up and dial back in to prevent hold music. Please announce your name before speaking Use the “Chat” feature to ask questions or share comments. – Send chats to “All Participants” so they can be addressed publicly in the chat, or discussed in the meeting (as appropriate). 2 Click on the “chat” bubble at the top of the meeting window to send a chat.

Agenda TopicTime Allotted Announcements5 minutes Final Wrap up of “Change” and Digital Signatures40 minutes Next Steps/Wrap Up5 minutes

Announcements We will be resuming our 1 hour meeting schedule starting next week May 28 th – We will meet from 2-3 pm ET going forward We are seeking organizations willing to pilot this work – If you are interested in being a pilot please contact: Jamie Parker: Or complete the following pilot form: 4

Change Discussion From HITSC Tiger Team: Choose a definition of “change” to data (for example, transformation with no intent to change the meaning of the data such as content format, terminology, or feature extraction versus substantive changes such as amend, update, append, etc.) and the implications for provenance. If the content changes, the change should be considered a “provenance event” A semantic change would be considered a provenance event. For those items with a digital signature any changes (format and semantic) would be considered a provenance event given a new signature would need to be obtained if changes are made. Once provenance is applied to an artifact and it is signed, any change made to the signed items would count as a provenance change Content changes in format that do not represent a semantic change are not considered a provenance event. Changes to content that affect the semantic meaning of the content (e.g.; changes to a values, additions or subtractions to the set of information, changes in intent that change the business use of the information ) are provenance events. 5

Digital Signature Discussion The Digital Signatures Discussion involved a review of the current capabilities and limitations for digital authentication within CDA. – Digital Signatures will be a strong suggestion in the Functional Requirements document but will not be required (thoughts) 6

Next Steps Join us on our next all hands meeting – May 21st from 2:00 -3:00 pm ET Sign up for pilots – – Or Jamie Parker at: 7

Support Team and Questions Please feel free to reach out to any member of the Data Provenance Support Team: Initiative Coordinator: Johnathan Coleman: OCPO Sponsor: Julie Chua: OST Sponsor: Mera Choi: Subject Matter Experts: Kathleen Connor: and Bob Yencha: Support Team: – Project Management: Jamie Parker: – Standards Development Support: Perri Smith: and Atanu Sen: – Support: Apurva Dharia: Rebecca Angeles: and Zach May: 8

Information Interchange Sub-Work Group –Final 9

Agenda TopicTime Review Tasking5 minutes Working Session35 minutes 10

Framing Question For information exchanged between EHRs, can I trust it, and has it been changed? – Information interchange begins once the exchange artifact is created and it shall not change during transport Consider that, for clinical care, if trending the data, one may need to know the degree to which the information can be trusted. 11

Information Interchange SWG Week of3/43/113/193/264/14/94/164/234/30 Launch SWG: Prepare, organize, plan, review existing materials Define a core set of provenance requirements Identify payloads that we should focus on Identify Candidate Standards to meet the need of requirements Consider implications of security aspects Capture policy considerations and request further guidance Legend: Not Started; In progress; Complete

Information Interchange SWG 13 Goal # GoalArtifact and Description 1 Define a set of basic/core requirements for provenance for information interchange between EHRs: Are there any specific technologies or architecture well suited for us to consider in the implementation guide (e.g.RESTul, Exchange, DIRECT and/or those specified in Meaningful use etc.) What transactions need to be specified in the IG? (For example IHE specification ABC…) Document defining a set of basic/core requirements for provenance for information interchange between EHRs (e.g. REST, Exchange, Direct etc.) and what Transactions needed in the IG 2 What type of payloads should we focus on when looking at information interchange requirements between EHRs (e.g. C-CDA etc.?) – what do we want to start with – pick a payload – this will be dependent on what the Standards group identifies Document, table or list of recommendations for the type of payloads for interchange requirements between EHRs 3 Identify Candidate Standards to meet the requirements of goals 1 and 2 using existing candidate standards list Short list of the proposed candidate standards that can achieve requirements of the first goal 4 Consider the implications of security aspects related to information interchange – Traceability, audit, etc. – what is the impact on the trust decision? (Consider Privacy) List or document of the implications of security aspects 5 If applicable, capture policy considerations related to system behavior and request further guidance from the HITPC. List of questions for HITPC

Assumptions/ Out of Scope: 14 Assumptions – keep it simple Black box exchange -EHR to EHR Information interchange begins once the exchange artifact is created Transport Content Neutral (the thing doesn’t get changed and is transported intact) Exchange Artifact shall not change during transport information required for end to end routing must be present in the un-encrypted metadata (to accommodate instances where the content is encrypted or otherwise not accessible- Receiver makes decision to accept message Must know sender in order for receiver to accept it (trust relationship) and trust the transport Out of Scope Intermediaries

Goal 1 Define a set of basic/core requirements for provenance for information interchange between EHRs: – Are there any specific technologies or architecture well suited for us to consider in the implementation guide (e.g.RESTul, Exchange, DIRECT and/or those specified in Meaningful use etc.) – What transactions need to be specified in the IG? (For example IHE specification ABC…) 15

Goal 1: Core requirements – What do we need to know What do I need to knowData ElementDoes it map to a DE in the System Requirement SWG? Who is the sender? May be original organization, original individual or a combination Organization name org ID Individual Name Individual ID Sender Location PARKING LOT On Behalf of (e.g. type) Device (might come up as SR activities) Author (too complex for initial goal) What is being sent (do we need to identify anything about the content)? Content Profile [one or combination: CCDA, Message (x12 or hl7v2) that can be wrapped and sent over content neutral transport] Transaction, Transaction Type (CCDA, v2.x message) Provider Directory Content Profile FHIR Resources Request Response IDGet some form of query ID to respond to request (echo back original data) Time being sentTimestamp Intended RecipientReceiver 16

Goal 2: Payloads Goal 2: What type of payloads should we focus on when looking at information interchange requirements between EHRs – Assumption: Content Neutral Transport MU2 and 3 alignment : – Two focuses (different implementation requirements) » ** START HERE: CCDA R2 – Start here based on requirements of MU (fixed payload) CDP1 is also listed in MU 3 but more constrained because Should start with Document type transaction Per SC: Address Communication/Information Interchange Requirements As a basic requirement, converting between different transport protocols should retain the integrity of the provenance data relating to the payload/content. If conversion is needed consider it internal to the system and part of the system requirements - this conversion happens within the black box of the EHR system » FHIR (as DSTU becomes more mature) cited as a direction not a standard required for implementation (2 different specified content standards within one overall standard) – How to indicate at transport layer what you are representing at the payload layer? What about metadata sent in XDS response? Document type in metadata? A-B (ccda) and A-C (FHIR) –source is same with same content using either protocol get same thing on other side – Result on B and C are same but different mechanism to get there – CCDA R2 – (Document Types – any document type should not present a problem) Appropriate template information based on guidance from structured documents workgroup and at discretion of pilots Goal – standard would support provenance activities 17

Goal 3: Standards to support requirements and payload What can transport C-CDA R2 we are just moving it – (EHR-EHR starting point) Any standards used to support: – Direct– in MU so should support this minimally – CONNECT – Or transport standard as identified by a pilot (i.e. RESTful) IHE ITI41 – Transaction used in cross document sharing (XD*) Consider the implications of security aspects related to information interchange – Traceability, audit, etc. – what is the impact on the trust decision? (Consider Privacy) Other Standards – X12 EDI X as a metadata wrapper can transport payload – and can wrap content – HL7 v2 MDM Informally vet this with task force and community at large Want to be as agnostic as possible and pick one that can support different types of transports (EHR-PMS or EHR to Payer) 18

Goal 4: Consider the implications of Security and Privacy Consider the implications of security aspects related to information interchange – Traceability, audit, etc. – what is the impact on the trust decision? (Consider Privacy) – Nothing that we are doing from provenance perspective that will change security concepts (strictly at the transport level) – Security might have impact if we are looking at exchanging encrypted payloads and or externally signed payloads – If expectation that receiving system can comply with privacy then ability of metadata evaluation on receipt not consumption must be considered (this might be an exchange issue in general) – might need to have some sort of indication that there is provenance in the content (can recipient support provenance/privacy requirements) This might be a policy questions – how should a receiving system behave if it is unable to comply to the provenance as expected by the sending system (defining system behavior based on conditions is an important consideration) This is an issue of whether the consumer of the data can comply with the senders provenance requirements If the sender holds the responsibility for determining recipient is able to comply with provenance/privacy/obligation is that inherent in trust frame work? (outside of encrypted payload) System Requirements Considerations – Does down stream system have to comply with senders wishes – and if rejected what kind of notification would the sender get? – Some challenges that payloads that have expressed pt. preferences regarding re-distribution (probably out of scope but is something to keep intact for this work)– this might be part of the consumption/System SWG This would happen at packaging or consumption but not as part of the exchange There are some issues with binding provenance to its target – this might be covered in the system requirements – Particularly if the goals is to keep provenance over the long term of the data (data comes in and is de-aggregated and moved into a record….what do we do with the provenance?) – We are not looking at consumption or the creation in the Information Interchange SwG 19

Goal 5: capture policy considerations related to system behavior and request further guidance Goal 5: If applicable, capture policy considerations related to system behavior and request further guidance from the HITPC – Questions For Standards Committee – Is there a need to associate provenance to artifact itself for end to end transport? (layers to artifact metadata) – System Requirements? (Provenance –what is inside and provenance – who sent it) For Standards Committee: Is there a need to accommodate provenance associated with the payload as something necessary for end to end transport of the data (receiver may want to know something about the sender prior to opening the exchange artifact) 20

Information Interchange Appendix 21

Data Elements for Consideration: Interoperability Roadmap (page 80) les/nationwide-interoperability- roadmap-draft-version-1.0.pdf Patient name * Sex * Date of birth * Race Ethnicity Preferred language Smoking status Problems Medications Medication allergies Laboratory test(s) Laboratory value(s)/result(s) Vital signs Care plan field(s), including goals and instructions Procedures Care team members Immunizations Unique device identifier(s) for a patient’s implantable device(s) Notes/narrative 22 Notice for Proposed Rule Making (page 148) inspection.federalregister.gov/ pdf TIN* NPI* Provider type* Patient insurance Patient age Patient sex in accordance with the standard specified in § (n)(1) (HL7Version 3) Patient race and ethnicity in accordance with the standards specified in § (f)(1) (OMB standard) and, at a minimum, (f)(2) (“Race & Ethnicity –CDC” code system in the PHIN VADS) Patient problem list data in accordance with, at a minimum, the version of the standard specified in § (a)(4) (September 2014 Release of the U.S. Edition of SNOMED CT®) Practice site address* Provenance should be captured on all clinical and administrative information * Elements on this list that are appropriate to include in provenance of other elements are those related to the demographics of the author

EHR Transactions Task Force Recommendation To address the priority areas recommended by the Task Force, the HITSC recommends: – The Initiative should begin its focus from the perspective of an EHR, including provenance for information created in the EHR (“source provenance”) and when it is exchanged between two parties. Provenance of the intermediaries is only important if the source data is changed. The notion of “who viewed/used/conveyed without modification along the way” is not important for provenance, as long as the information was not changed. Recommendation follows Scenario 1 of the Use Case: Start Point  End Point – Focus on what happens Inside the EHR When being exchanged between EHRs (assume no change to clinical content during exchange) Per the task force recommendations: assume that what is already in the EHR is good – Our analysis should start from this point and this assumption – The information interchange group can look at the transaction and taking what is available and moving it to another EHR 23 Out of Scope: 3 rd Parties (e.g. HIEs third party assemblers etc.)

Scope Address Communication/Information Interchange requirements: – The integrity of the provenance data for clinical content should remain intact during transport. For the purposes of this use case, start with the assumption that at the point for information interchange, the “source provenance” is good, complete, trusted Coupling sender and receiver to content? Access to payload is not the question is there a dependency on having access to get to that point 24

Goal 1: Define a set of basic/core requirements for provenance for information interchange between EHRs Methodology: – Start with MU Specified Transports Focus at higher level of Transport Protocol – for any of the identified protocols we will do a “deep dive” based on need – Start at the abstract: For example lets determine between the exchange parties what do we need to know? – Who is the sender? – Who is the intended recipient? – What is being sent? » This helps us determine what needs to be exchanged and vet this against the technologies available 25

System Requirements Appendix 26

RECAP: Minimum Set of Requirements to Review Identifying provenance requirements of an EHR system – what are the events we expect them to manage Import- New Artifact Arrived (decomposing/disassembling content prior to accepting/putting in EHR record and then maintain) – Decompose (include verification by human to make reliability judgment) – Disassemble to incorporate into EHR Use or View- show all detailed data Create Update Maintain (not necessarily a provenance event as we have already created and updated which are provenance event) – Compose Content (as done in EHR system) – Assemble Composed Content (as done in EHR system) Export – Artifact ready to go (Transmit perhaps Information Interchange) NOTES: – Assembling = done by software – Compose = done by human and software – Policy committee – viewing and accounting of disclosers - if no change to clinical data 27 Out of Scope: 3 rd Parties (e.g. HIEs third party assemblers et) = as identified by the SC Task Force

EHR Transactions Task Force Recommendation To address the priority areas recommended by the Task Force, the HITSC recommends: – The Initiative should begin its focus from the perspective of an EHR, including provenance for information created in the EHR (“source provenance”) and when it is exchanged between two parties. Provenance of the intermediaries is only important if the source data is changed. The notion of “who viewed/used/conveyed without modification along the way” is not important for provenance, as long as the information was not changed. Recommendation follows Scenario 1 of the Use Case: Start Point  End Point – Focus on what happens Inside the EHR When being exchanged between EHRs Per the task force recommendations: assume that what is already in the EHR is good – Our analysis should start from this point and this assumption Functions of the EHR can include: – Creating new data (adding new clinical content) – Creating new artifacts (e.g. assembler functions) which are prepared for transmittal – The information interchange group can look at the transaction and taking what is available and moving it to another EHR 28 Out of Scope: 3 rd Parties (e.g. HIEs third party assemblers et)

Data Elements for Consideration: Interoperability Roadmap (page 80) les/nationwide-interoperability- roadmap-draft-version-1.0.pdf Patient name Sex Date of birth Race Ethnicity Preferred language Smoking status Problems Medications Medication allergies Laboratory test(s) Laboratory value(s)/result(s) Vital signs Care plan field(s), including goals and instructions Procedures Care team members Immunizations Unique device identifier(s) for a patient’s implantable device(s) Notes/narrative 29 Notice for Proposed Rule Making (page 148) inspection.federalregister.gov/ pdf TIN NPI Provider type Patient insurance Patient age Patient sex in accordance with the standard specified in § (n)(1) (HL7Version 3) Patient race and ethnicity in accordance with the standards specified in § (f)(1) (OMB standard) and, at a minimum, (f)(2) (“Race & Ethnicity –CDC” code system in the PHIN VADS) Patient problem list data in accordance with, at a minimum, the version of the standard specified in § (a)(4) (September 2014 Release of the U.S. Editionof SNOMED CT®) Practice site address.

Start Point – End Point Scenario 0Consented%20Use%20Case_ pdf/ /DPROV%20Use%20Cas e%20_%20Final%20Consented%20Use%20Case_ pdf 0Consented%20Use%20Case_ pdf/ /DPROV%20Use%20Cas e%20_%20Final%20Consented%20Use%20Case_ pdf 10A.1 User Story User Story 1: A patient arrives at the ophthalmologist’s office for her annual eye exam. The ophthalmologist conducts an eye exam and captures all of the data from that visit in his EHR. The ophthalmologist electronically sends the information back to the patient’s PCP (where all data in the report sent was created by the ophthalmologist). User Story 2: A patient has a PHR that allows them to record their daily dietary intake. The patient accesses the PHR and requests that their dietary intake for the past month be transmitted to their PCP prior to their visit next week. The patients uses a PHR to transmit the dietary record to the PCP. The PCP understands from the document’s provenance that the data was generated by the patient and that it is authentic, reliable, and trustworthy. (this is outside of the EHR to EHR) 30

31 Scenarios from Use Case Sequence Diagram Assembler/Composer

Data Elements in the Use Case Start Point- 32 RoleData CategoryData ElementComments Start PointWhoSending System Sending System Organization Author Custodian Role WhenSend Date Send Time WhereAddress State Zip Type (What)Software Device WhyClinical Context Purpose Integrity/ Authenticity Digital Signature Additional Patient Record Target Assigned Author Informant Service Event Performer Authenticator Legal Authenticator Notes from our call today: Since EHR will be the point of origination we may not need a start point. The start point of our use case would be the originator (not focusing on compiler or composer). It was also suggested that we rethink roles because the Start point in an EHR and the start point of the exchange are different. We may need to come up with 2 different names for the “start point” roles Potential Removal or rename Start Point of Exchange? (see notes below) 0Case%20_%20Final%20Consented%20Use%20Case_ pdf/ /DPROV%20Use%20Case%20_% 20Final%20Consented%20Use%20Case_ pdf

TransmitterWhoTransmitter OrganizationThis might be looked at by the Information Interchange SWG Transmitter System WhenTransmission Time Sent Transmission Date Sent WhereTransmitter Location Transmitter System Location Type (What)Transmission Device Transmission Software Transmission Hardware Transmission Method WhyPurpose of Transmission RoutingTransmitter Sender Address Receiver Address Integrity/ AuthenticityDigital Signature WhoTransmitter Organization Transmitter System AdditionalPatient Record Target 33 Data Elements in the Use Case: Transmitter Transmitter based on diagrams and community call was proposed for removal but might be a good candidate for review in the Information Interchange SWG

OriginatorWhoOriginator Organization Originator Author Originator Enterer Originator Attester Originator Verifier Originator System WhenOriginator Time Created WhereOriginator Locations Originator System Location Type (What)Originator Event AdditionalPatient Record Target Author Assigned Author Authoring System Authoring Organization Informant Service Event Performer Participant Custodian Authenticator Legal Authenticator 34 Data Elements in the Use Case Originator Keep and rename to follow diagram to “Initiating System?”)

AssemblerWhoAssembler System Assembler Organization Intended Recipient WhenAssembly Date Assembly Time WhereAddress State Zip Type (What)Software Device WhyAssembly Purpose Integrity/ Authenticity Assembly Participants Attestation/Nonrepudiation of data AdditionalPatient Record Target Author Assigned Author Authoring System Authoring Organization Informant Service Event Performer Participant Custodian Authenticator Legal Authenticator 35 Data Elements in the Use Case – Assembler Assembler proposed for removal based on diagram?

Data Elements in the Use Case – Composer ComposerWhoComposer System Composer Organization WhenComposition Date Composition Time WhereAddress State Zip Type (What)Software Device WhyComposing Purpose Integrity/ AuthenticityComposing Participants Selector AdditionalPatient Record Target Author Assigned Author Authoring System Authoring Organization Informant Service Event Performer Participant Custodian Authenticator Legal Authenticator 36 Composer based on diagrams –proposed for removal

Start Point – End Point Scenario 0Consented%20Use%20Case_ pdf/ /DPROV%20Use%20Cas e%20_%20Final%20Consented%20Use%20Case_ pdf 0Consented%20Use%20Case_ pdf/ /DPROV%20Use%20Cas e%20_%20Final%20Consented%20Use%20Case_ pdf 10A.1 User Story User Story 1: A patient arrives at the ophthalmologist’s office for her annual eye exam. The ophthalmologist conducts an eye exam and captures all of the data from that visit in his EHR. The ophthalmologist electronically sends the information back to the patient’s PCP (where all data in the report sent was created by the ophthalmologist). User Story 2: A patient has a PHR that allows them to record their daily dietary intake. The patient accesses the PHR and requests that their dietary intake for the past month be transmitted to their PCP prior to their visit next week. The patients uses a PHR to transmit the dietary record to the PCP. The PCP understands from the document’s provenance that the data was generated by the patient and that it is authentic, reliable, and trustworthy. (this is outside of the EHR to EHR) 37