The purpose of this page is to capture the goals and requirements to enhance ELR reporting in the context of the COVID pandemic, and in that context provide clear guidance to all parties in the communication flow to enhance their contributions to enhance the ELR transaction to a Public Health Authority.
This is meant to be a working document that we rapidly evolve based on latest understanding. Once it is believed to be sufficiently complete and accurate it will be published elsewhere as a guidance document.
Hans Buitendijk - Cerner, HL7 Orders & Observations Co-Chair, EHRA Standards & Interoperability Chair
David Burgess - LabCorp, HL7 Orders & Observations Co-Chair
Freida Hall - Quest Diagnostics
Jason Hall - CDC
Janet Hamilton - CSTE
Riki Merrick - APHL, HL7 Orders & Observations Co-Chair, IHE Pathology and Laboratory Medicine Co-Chair
Craig Newman - Altarum, HL7 Public Health Co-Chair
Andrea Pitkus- University of Wisconsin-Madison
Dan Rutz - Epic, IHE ???
Kathy Walsh - LabCorp
Michael Waters - FDA
- AOE - Ask at Order Entry
- eCR - Electronic Case Report
- eDOS - Electronic Directory of Services implementation guide or transaction according to that guide
- ELR - Electronic Lab Reporting implementation guide or transaction according to that guide
- IICC - IVD Industry Connectivity Consortium
- LOI - Laboratory Order Interface implementation guide or transaction according to that guide
- LRI - Laboratory Results Interface implementation guide or transaction according to that guide
- O - Optional
- PD1 - Patient Demographics
- PHA - Public Health Authority
- PID - Patient Identification Segment
- RE - Required but may be Empty (Must support)
When ELRs are sent to PHA at times, they may not contain sufficient demographic, and other data as is necessary required by public health by law. While some of that data could be part of a subsequent eCR, if the data is already available at time of order and ELR transmission before an eCR is available (or the next eCR is available) helps accelerate the process.
Although ELR has been required of all laboratories performing testing for reportable conditions, only Eligible Hospital labs have been required to send HL7 v 2.51 messages with LOINC and SNOMED CT codes conforming to each PHA jurisdiction's requirements for the past decade using a vendor system certified to provide functionality according to the HL7 ELR Guide in Meaningful Use regulations. For the remainder of laboratories (i.e. blood banks, reference labs, government/DOD labs, independent labs), they need to meet each jurisdictions ELR requirements based upon the HL7 2.51 ELR IG, but may send messages in HL7 2.31 format, which may or may not contain LOINC or SNOMED CT codes or other fields. If the PHA onboards the lab, they may accept a variety of fields, formats as long as they are receiving key information. Also it's unclear if point of care testing locations like pharmacies, drive up/drive through testing sites or state national guard collection sites for COVID-19 are 1. reporting ELR at all, 2. providing all required information and 3. those reporting on paper requisitions will not be providing LOINC or SNOMED CT codes and are often lacking demographics, specimen source and other required PHA elements.
Through the process from ordering to reporting, there is a need to:
- have the ability to document/collect the data (particularly certain ask-at-order-entry information in context of specific tests), either at the point of order entry by the provider, or at the point of specimen collection by the collector.
- document as part of the order that data (recognizing that in various settings the data is just not available yet, or the urgency does not allow for collection/documentation at that time),
- communicate the data (ensure a consistent format/standard is available) to the performing laboratory directly or through the provider's laboratory to the reference lab performing the test (where the performing lab does not have direct, electronic access to that data in the ordering provider's system)
- retain the data in the laboratory to be able to include it in the ELR
- communicate the ELR with all available data to the PHA (ensure a consistent format/standard is available) (AP. This has been available for over a decade)
There are numerous challenges along in this flow in terms of data not being able to be collected, not being documented, not using standard vocabulary, not including it in the order, not forwarding certain data, dropping standard vocabulary, not retaining data, not communicating data available to the PHA. The challenges this guidance is looking at is ensuring that the relevant data can be communicated from ordering provider through a potential intermediary lab to the performing lab and on to PHA as an ELR in a complete, standard, consistent format and with vocabulary.
Currently the specific challenges in that area are:
- Device identification data is not included in device/test documentation
- The latest LIVD guide does not include an ability for the relevant identifier for each test kit. - QUESTION: it does not have a slot for it (I thought that's what the Equipment (to be renamed to Product) UID is for) - or that information is not currently provided in that slot? HJB: Still TBD in R2 to finalize location, either ObservationDefinition or DeviceDefinition.
- Deployed lab order implementation guides do not include the guidance on how to convey all the relevant ask-at-order-entry questions and answers recognizably and consistently.
- The latest eDOS guide has the commonly used set of AOEs with standard LOINC codes and where applicable answer sets, but is not complete to cover the latest updates.
- The latest LOI guide addresses the ability to communicate all relevant data, including any AOEs, but is not deployed in active use.
- Upgrading to the latest LOI guide is a major lift for everybody, so although desired for the mid/long-term, it is not a good short-term recommendation.
- Deployed lab reporting implementation guides do not include the guidance on how to convey all the relevant ask-at-order-entry questions and answers recognizably and consistently, nor the relevant device information, nor sufficient clarity on minimum desired demographic data (some fields may need to be marked RE rather than O, or X).
- The latest LRI guide (that now includes ELR) accommodates the ability to send any relevant AOEs, but may need more specificity on other patient demographic data already available in PID or PD1, while it does need guidance on how to convey the relevant device/test kit(s) used to perform the test.
- Upgrading to the latest LRI guide is a major lift for everybody, so although desired for the mid/long-term, it is not a good short-term recommendation.
As indicated before, there are many other challenges to ensure that data if fully, accurately, and consistently documented and communicated, including use of encoded standard vocabulary (at least along with any local coding that remains relevant), but that is not addressed here. Our goal is to provide guidance on how to address the above challenges in the necessary implementation guides that can be deployed as quickly as possible and as consistent as possible across ALL jurisdictions.
We understand that there may be a desire/need to send the original order to PHA directly as well to get earlier insight into certain tests ordered including PHI, particularly during an emergency (as that may invade patient privacy unnecessarily during "normal" situations), but when that interest becomes a requirement, the guidance for the ordering flow is sufficient as the starting point and being extended based on local PHA emergency mandates on how much (or not) PHI is to be removed before forwarding it. That guidance should be as consistent as possible as well, but we will address that when an actual requirement in any jurisdiction is raised.
The following data set is desired to be made available as part of an ELR as much is is available upon placing the order and to the extent that the Laboratory has access to any additions before the ELR is sent.
|1||Patient age||Required-HHS||For reporting from PHA to CDC/HHS, when DOB may not be included|
|2||Patient race||Required-HHS||some jurisdictions used to not allow collection of this information - is that still the case? If so, will this superseed those regulations?|
|3||Patient ethnicity||Required-HHS||some jurisdictions used to not allow collection of this information - is that still the case? If so, will this superseed those regulations?|
|5||Patient residence zip code||Required-HHS|
Some jurisdictions consider a combination of age, sex and zipcode PII, because of the population densitiy - need to accommodate that by allowing creation of regions (and that may be better done in the county field)
|6||Patient residence county||Required-HHS|
Some jurisdictions consider a combination of age, sex and county PII, because of the population densitiy - need to accommodate that by allowing creation of regions
ELR requires use of numeric FIPS 5-2 codes
|7||Patient name (Last name, First name, Middle Initial)||Optional-PHA||It is required for PHA, not desired for HHS|
|8||Patient street address||Optional-PHA||It is required for PHA, not desired for HHS|
|9||Patient phone number with area code||Optional-PHA||It is required for PHA, not desired for HHS|
|10||Patient date of birth||Optional-PHA||It is required for PHA, not desired for HHS|
|1||Ordering provider address||Optional-PHA||It is required for PHA, not desired for HHS|
|2||Ordering provider phone number||Optional-PHA||It is required for PHA, not desired for HHS|
|3||Ordering provider name and NPI (as applicable)||Required-HHS||For some testing there may not be a specific ordering provider (e.g. employment related testing) - it is a CLIA requirement so need to ensure that this is provided|
|4||Ordering provider zip||Required-HHS|
For some low density areas, this may provide too much identification in the HHS feeds
Concern has been raised that this may be hard to decide which to pick, for providers, that have more than one practice location
|5||Date test ordered (date format)||Required-HHS|
|6||Test ordered – use harmonized LOINC codes provided by CDC||Required-HHS||Currently no order code look up available - nor guidance on which LOINC to pick for which order|
Ask at Order Entry
|2||Employed in healthcare?||Y/N/U|
|3||Symptomatic as defined by CDC?||Y/N/U|
|4||Symptomatic = Y then Date of Symptom Onset||mm/dd/yy|
|7||Resident in a congregate care setting (including nursing homes, residential care |
for people with intellectual and developmental disabilities, psychiatric treatment
facilities, group homes, board and care homes, homeless shelter, foster care or other setting)
|8||Pregnant?||Pregancy status of the patient||There are 2 LOINCs available |
|1||Date specimen collected (date format)||Required-HHS||ELR allows use of 0000, when date is unknown|
|1||Performing facility name and/or CLIA number, if known||Required-HHS|
|2||Performing facility zip code||Required-HHS|
|3||Device Identifier||Required-HHS||Intent is to know which test has been performed where|
|4||Test result – use appropriate LOINC code, as defined by the Laboratory In Vitro Diagnostics (LIVD) Test Code Mapping for SARS-CoV-2 Tests provided by CDC||Required-HHS||This is refering to the performed test (observation.code in CDA/FHIR)|
|5||Test result (value) – use appropriate SNOMED CT code as defined by the Laboratory In Vitro Diagnostics (LIVD) Test Code Mapping for SARS-CoV-2 Tests provided by CDC||Required-HHS||This is referring to the result value (Observation.value in CDA/FHIR)|
Some results can be numeric, and require units of measure, but no SCT code
some results could be longer text, specifically interpretations
SCT is required for qualitative result values
|6||Test Result date (date format)||Required-HHS|
Updates to Implementation Guides
The following updates need to be made in the base implementation guides for a next version to include the short-term Implementation Guidance provided in the next section.
- Add the device identifier for the test kit in the correct resource (DeviceDefinition or ObservationDefinition?)
- Need a product reference uid to identify the ObservationDefinition OR is it really on the DeviceDefinition?
- If on the ObservationDefintion, need to re-include ObservationDefinition.identifier 0..*
- identifier.type = "product-reference-identifier"
- identifier.system = TBD (for uid)
- If on the DeviceDefinition .....
- Add the assigning authority as well. This is particularly important for OBX-18.3 if the identifier is not an official device identifier.
- Add new AOEs to master list, but nothing needed for the rest
- Add guidance on how to convey device identification information
- Can be sent in OBX 3 as a result for "reagent identifier" build as a test result in the panel reporting COVID results.
- Can also be sent in PFT 15 (I think)
- Can be sent in SPM (field?)
- Orders - Any HL7 v2 ORM or OML message in place between ordering provider and Laboratory as well as Laboratory and Reference Laboratory where the tests are outsourced.
- Use LOI R1 STU 3 Section 1.4.10X, 6.15, Z for specification on how to add AOEs using pre-adopted OBX capabilities.
- Include Profile TBD, but a new profile in MSH-21 so receiving system recognizes it may contain OBX segments that are based on HL7 v2.8.2
- Use eDOS R2 STU 3 Appendix A for an initial list of AOEs PLUS [whatever list Riki is adding if maintained separately for now]
- eDOS would be electronic version of CLIA Specimen Collection Manual
- Use Mapping Spreadsheet below for PID and OBR fields that are RE if not R.
- Results - Any ELR to PHA
- Use LRI R1 STU 3 Section 8.11, 13.3.3, Z for specification on how to add/forward AOEs with the result using OBX segments that are based on HL7 v2.8.2 (AP note: AOEs become/are results and their values in OBX 3/5 messages sent to providers and PHA via ELR. HJB: clarified in both Orders and Results. Thank you!)
- Use Mapping Spreadsheet below for PID, OBR, and OBX fields that are RE if not R.
- Use Device Identifier guidance below
- Strongly Recommended
- Adopt LIVD as quickly as possible to stay informed on latest LOINC mappings and device identification information
- Adopt eDOS as quickly as possible to stay informed on latest AOEs per test.
- Adopt LOI Rn as quickly as possible to enable most consistent and complete communication of Lab orders.
- Adopt LRI Rn as quickly as possible to enable most consistent and complete communication of ELR. (AP Note: LRI is separate form ELR and ELR is required by law, while LRI is not technically required anymore, but was for stage 2 MU of hospital labs only.)
- It is expected that some test kits/reagent packs/cartridges will not have an official Device Identifier, so any unique device identifier will do. For purposes of this discussion that is considered a Device Identifier. Keep in mind for manually performed tests such as rapid kits, point of care testing or non interfaced instruments, these will have UDIs that are manually entered into the LIS.
- As the collection of this identifier may vary or require more/less development/implementation effort, it appears that one must be able to collect just the Device Identifier, just the UDI Carrier (the barcode read in HRF form or AIDC, although the latter was not really referenced in the examples), or a combination of both with possibly other parsed elements of the UDI components of the UDI Carrier.
- Ideally, the LIS should configure the Device Identifier provided in the LIVD file (IICC .csv or FHIR .json file). Then it can be passed along in the test. (AP: What does this mean? Do you mean as a test result and result value? The assay produces the test results)
- Issue: is IVD Test Code always related to exactly one test kit?
- Ed from Abbott is checking on ROMA assay, which utilizes results from two other assays/regent packs to produce the calculated ROMA result
- Outline for LRI
- If the Device Identifier is included on the LIVD file provider to the Lab by the device manufacturer as part of the mapping definition to LOINC, the LIS could include that information in its configuration so that when a test result comes through that involves a known test kit, it can auto-populate that field for ELR. It has been confirmed many LISs do not have that capability today, so may not immediately have that available. The performing laboratory could build it as a OBX 3 to take advantage of existing LIS functionality. Nor was it confirmed that upon receiving the test results from a device for a a particular IVD test code the test kit Device Identifier would be unambiguously known and if not passed on.
- If only device identifier is available (from LIS configured flow or entered manually as the device identifier), use OBX-18
- If a barcode is scanned (UDI Carrier) then EITHER:
- parse the Device Identifier and put it into OBX-18 per above
- Include the UDI Carrier into PRT-10 according to the UDI Pattern R2 guidance in the HL7 v2 section on page 13
- include all the parsed elements as well according to UDI Pattern Release 2 in the HL7 v2 section (page 14) for Device Identifier and Production Identifier
- PRT-4 must be set to xyz
- Enter the barcode value into a designated lab result value field in the LIS to be stored for a patient result value or in another field in the LIS for the test build. When results are reported, the OBX 3 field is populated with the result value of the barcode. UDI barcodes could be scanned with the implementation of reagents in the test build/set up screens for use for each patient instance.
- If only the UDI Carrier is provided, then the receiver (PHA, FDA) needs to parse the Device Identifier to enable comparison with the LIVD files or other analysis, filtering, etc.
- Issue: It was not determined that would be o.k., but we noted that requiring just the Device Identifier may complicate workflows to collect that data and/or impose extra capabilities on the LIS beyond a simple scan.
- Out of Scope
- UDI for equipment/instrumentation platform
- UDIs for control and QC
Use this spreadsheet to collect options and comments around the proposed mapping of the HHS elements to various places in Order and result messages.
Open questions on spreadsheet
- Suggest to organize by data types. IF there is a need to link back to a sequence on an HHS spreadsheet, let's use HHS ID and I'll fix the in-line tables above accordingly to have the link.
- Suggest to replace Must Support with RE for consistency with guides.
- Unclear whether the optionalities by guide are current or recommended - these were the current usages I found, so some will need to be adjusted
- Some of the AOEs may be able to be answered based on knowledge in the system (e.g. is patient hospitalized = Patient class is inpatient in encounter)
- We should do an analysis of which of these HHS questions are available in the EHR-s (and in what format) and then decide on the best way to share that data, e.g., use PID/PD1/PV1/PV2/NK1 (occupation) - I got some feedback, that UNLESS the element is already in ELR R1 (like PV1-2), OBX segments will be easiest in the short term to add to the surveillance system.
- Understanding is some drive up/drive through COVID testing sites are using paper requisitions for ordering, so data elements would be required on paper req to performing lab, who would enter them into their results reporting to PHA ELR message (if provided!).
- Options for differentiating OBX for AOE vs OBX for results: (AP Note: The AOE question and response, would be represented as a result and result value. The LOINC is likely the same for the AOE question and result in most cases. However, they may have an order ID/identifier for the AOE, and a result ID for the result field to distinguish the two uses of the same data item.)
- LOINC code
- ELR R1 expects OBX after SPM for patient age - could put others there
- Use a separate OBR for EPI important information
- OBX-29 in LOI/LRI
- While communication of a .csv or flat file from performing lab to PHA, if the above guidance cannot be implemented in time, is out of scope of HL7 to define format, we could make some suggestions on what to include to match with the ELR transaction that it needs to match up with.
- APHL is creating a flatfile format for this purpose (not csv, but rather tab delimited) - the elments in this flatfile are all mapped to the respective ELR R1 location, so that transformation could occur at the sender or receiver side (conversion tooling for the PHAs is being discussed)
- Suggest to allow for Direct Messaing to convey the file.
- Why not just use ELR in xml format via Direct?
- What shall we point to for LIVD latest guide?
Resources for Background