Word Document

Owned by Anonymous
Last updated: Aug 06, 2021 by Anonymous
10 min read

Implementation Proposal

 

  1. Standardization of Code Assignment              

There is currently no single standard terminology that fully describes laboratory data sufficiently to support clinical interoperability. Laboratory concepts should therefore be represented by a unique name that represents a comprehensive standardized set of uniform codes and descriptions that fully describe the laboratory test, and the same test should be described by the same set of standardized codes in every instance. The standardized codes assigned to a particular IVD test should be curated by both, the IVD manufacturer of the assay and an authoritative body such as SHIELD.

1.1   Test Order/Reporting Name

The precise name of the test being ordered by the clinician should be represented by a unique LOINC Order Code. The unique LOINC codes should be mapped to a similarly unique SNOMED CT code that shares the exact same meaning. This is so that countries not utilizing LOINC can have clear guidance on which corresponding SNOMED CT code should be used and vice versa.

1.2   Laboratory Test Results (Observation Values)

Laboratory test result values can be quantitative (often with units of measure) or qualitative. Qualitative results can be ordinal, nominal or narrative. Quantitative results should use consistent units of measurements to avoid interpretation errors.

1.21 Component/Analyte

The component or analyte of the test being performed should be represented by the LOINC performable code. The unique LOINC codes should be mapped to a similarly unique SNOMED CT code that shares the exact same meaning.

1.22 Qualitative Results

Ordinal and nominal results should be represented by SNOMED CT Codes. Ordinal and nominal results should be standardized across lab vendor platforms for assays that represent the same test.

1.3   Specimen

The specimen on which the test is performed and its attributes [Specimen Type, Collection Method, Different Focus (Maternal vs Fetal)] should be represented with the SNOMED CT code..

1.4   IVD Instrument and Test Kit Identification

The IVD test kit and instrument should be represented by Universal Device Identifiers. Identification of the instrument and test kit on which the test was performed provides greater granularity on the methodology used to perform the test and is essential for determination of comparability of results. Information on the IVD kit and instrument used to produce results can aid in identifying issues with specific assay platforms, instruments, or lot numbers. Many IVD systems are not equipped to record and communicate the UDI to the LIS and EHR systems. Transmission of the equipment UDI and test kit UDI should be automated and a requirement of IVD systems.

 

  1.  

Lab test order, performable and result code alignment

  1.  

To eliminate incorrect selection of lab test order, performable and resultable code alignment, a prescriptive enumeration of codes will be developed and maintained by the SHIELD program for IVD devices.  This will take the form of the expanded LIVD file in which each commercially available IVD device will be inventoried/listed.  For each IVD platform the following codes will be enumerated for each test the IVD has FDA approval for use.

  1. Test Order/Reporting Name – A LOINC order code (or SNOMED CT Observable entity concept ID  for SNOMED CT Observable entity concept ID for non-LOINC nations) will be enumerated for each assay approved by the FDA for the IVD.  (Note that the LOINC term WILL be modelled, classified and distributed in the US domain as noted below in the terminology section)

  1. Analyte/Observable/Resultable – A LOINC code (or SNOMED CT Observable entity concept ID  for SNOMED CT Observable entity concept ID for non-LOINC nations) will be enumerated for each assay approved by the FDA for the IVD.  (Note that the LOINC term WILL be modelled, classified and distributed in the US domain as noted below in the terminology section)

  1. Qualitative Results – A constrained list of SNOMED CT concepts for qualitative, nominal and/or ordinal results will be enumerated as part of the LIVD file for each assay approved for use on a particular IVD.

  1. Specimen – SNOMED CT concepts will be enumerated for all possible, approved specimen types on which an IVD assay may be performed.

  1. IVD Instrument and Test Kit Identification – Instrument UDI and Kit UDI will be enumerate for each assay for which an IVD device has FDA approval. 

3.1   Technological Gap Analysis and Resolution

EHR and LIS functionality should be standardized to support adequate mapping and transmission of the laboratory element codes.

  1. EHR and LIS Systems should accommodate the designation of multiple LOINC or SNOMED CT codes for each element as needed.

  1. IVD vendors systems should ideally support automated transmission of instrument UDI and Kit UDI transmission.

3.2   The top 1000? laboratory tests in terms of annual volume should be fully mapped with standardized codes by the IVD vendors and SHIELD. Priority should be given to tests with the greatest potential impact to patient safety such as coagulation assays, thyroid function tests and cardiac markers.

3.2.1         Coagulation Assays

  1. Activated Partial Thromboplastin Time (aPTT)

  1. Thrombin Time (PT) and International Normalized Ratio (INR)

  1. D-Dimer

  1. Fibrinogen

  1. Fibrinogen Degradation Products (FDP)

3.2.2         Cardiac Markers

  1. Troponin I

3.2.3         Thyroid Function Tests

  1. Thyroid-stimulating hormone (TSH)

  1. Free Thyroxine (T4)

  1. Total Triiodothyronine (T3)

  1. Total Thyroxine (T4)

  1. Free Triiodothyronine (T3)

  1. Thyroid binding globulin

  1. Thyroglobulin

 

  1. Laboratory Interoperability Data Repository (LIDR)

As noted in item 2, the curated set of standardized codes in the LIVD format should be aggregated and made available through a centralized repository. Individual IVD vendors may continue to offer the standardized codes for their tests through package inserts or corporate websites, but all the codes from every vendor should be made available through the central repository. The LIDR must have a human searchable interface that would allow users to find the codes needed for mapping to their local test codes and terms by searching for manufacturer and assay keywords. In addition to the standardized codes, the LIDR should also contain information on the harmonization status (i.e. whether or not the assay has undergone calibration to a gold standard reference) of each test entry. The LIDR would serve as the authoritative source for the definitive standardized code representation for each IVD test from each IVD vendor.

LIVD file input will be solicited from the vendors, collected and aggregated in the LIDR using a structured data capture (SDC) tool.  The SDC tool will be available for vendor completion within the LIDR.  Forms can be completed manually or via API.

  1. Laboratory Network Enrollment

A consortium of laboratories from large health care systems and referral laboratories will be recruited to implement the standardized codes developed by the IVD vendors and SHIELD. SHIELD will provide the consortium members with the previously curated standardized codes for the top 1000. The total number of tests coded by the institutions as a percentage of the national annual laboratory test volume could serve as an initial benchmark for the effectiveness of the SHIELD implementation.

  1. Clinical Interoperability

 

Clinical interoperability is the ability of two or more systems to exchange information (laboratory results) and to use equivalent results for trending purposes, clinical decision support and machine learning algorithms. Only equivalent results can be safely trended with locally generated laboratory results. In order for tests to be equivalents and interchangeable, the following conditions must be met:

  1. Same test (defined as same analyte/observable performed on the same specimen type)

  1. Same Instrument Platform

  1. Same Test Kit

 

OR

  1. Same Test

  1. The IVD manufacturer has calibrated their assay to a reference “gold standard” in order to harmonize their results with those of other IVD manufacturers. Only assays that undergone this harmonization process can be considered equivalent regardless of platform and/or test kit.

 

In order for clinical interoperability to be achieved, the information for equivalence determination must be readily available at the time of mapping the external result into the native EHR. The following coded data elements must displayed in the mapping module as an EHR functionality requirement:

  1. Test Analyte/Observable LOINC

  1. Specimen SNOMED CT

  1. Instrument UDI

  1. Test Kit UDI

 

Only external tests that share the exact same attributes should be fully mapped and allowed to trend with the corresponding local test results.

 

For assays that have undergone harmonization with calibration to a gold standard reference, the harmonization status of the assay should be available on the LIDR. Laboratories seeking to map harmonized assays to their local assays would consult the LIDR as a definitive reference.

 

  1. In Vitro Diagnostic (IVD) Data Hub – Contract Research Organization

The consortium institutional laboratories previously recruited in the implementation of SHIELD standardized codes could serve as the basis for an IVD Data Hub. The laboratory results for these institutions would presumably have undergone standardized coding, therefore precluding the need for manual manipulation of the data. The data collected could be utilized for post-market surveillance of IVD vendor assays and conversion from EUA to 510(k) status. The data hub would be supported by fees generated from the collection and analysis for the data for the use of IVD vendors.

The data elements that would need to be collected include:

????

  1. Terminology and Semantic encoding of lab tests – Scott

LOINC, SNOMED CT and RxNorm represent the vast majority of medical terms and concepts necessary to create semantic interoperability for most of the use cases in human health care.  However, the terminologies are not harmonized.  As part of this implementation plan, the LOINC terms will be modeled in accordance with the SNOMED CT concept model as originally contemplated in the Regenstief/SNOMED collaboration agreement.  The vast majority of this work was completed for the top 20,000 LOINC terms and both RII and SNOMED International make these LOINC to SNOMED expression data available to the public.  However, the work has not been amendable to operations due to RII/SNOMED/NLM contractual issues.  The SHIELD implementation plan entails comingling LOINC, SNOMED CT and RxNorm for the US domain.  The joint terminology will be published by the NLM or another licensing body such that all concepts will be represented by their terminology of origin (i.e., LOINC, SNOMED CT or RxNorm) AND these concepts will conform to the SNOMED CT logical model which allows for vastly improved analytics and logical subsumption/aggregation of concepts that is critical for AI, CDSS, public health surveillance and translational/RWE research.

 

  1. Standardization of Qualitative Laboratory Result Reporting

Currently there is no standard for reporting of qualitative results for IVD vendor assays. This leads to variability in terms of how results for similar assays are reported and leads to inefficiencies and impacts the aggregation of laboratory data for secondary use. Qualitative results including ordinal, nominal and narrative results should be standardized for similar assays. SHIELD along with IVD vendor stakeholders should harmonize the reporting for similar tests moving forward in order to advance semantic and clinical interoperability.

10.   Laboratory Test Instance Data

There are currently two different levels at which laboratory equipment can be identified. The “kind” (eg, IVD instrument model, a test reagent kit) and the instance (eg, IVD Instrument serial number, test reagent kit lot number, calibrator kit number). Although the LIVD format is currently intended to address only the laboratory equipment “kind” data, there is value in identifying laboratory instance data for both clinical care and secondary use.

11.   Harmonization of Laboratory Results

Currently there are only ~100 assays out of thousands of clinical laboratory tests that have undergone calibration to a gold standard in order to be harmonized across vendor platforms. Harmonization allows these test results to be considered equivalent for the purposes of patient care. Harmonized test also allow for the promulgation of national guidelines for care of disease states since laboratory cutoffs used in the guidelines would not be dependent on a particular methodology or instrument platform. With the current low number of harmonized tests available, the most practical strategy might be to make the harmonization status of the assays available through the LIDR. One of the goals of SHIELD moving forward should be the promotion of the harmonization of laboratory results across IVD vendor platforms in order to promote true interoperability.

12.   Traceability/Provenance of Data

 



 

Implementation Plan:

 

  1. As discussed above, tooling will be developed to support the expansion of the LIVD file to include all necessary LOINC/SNOMED coding for test orderables, performables and resultables by device UDI/test kit UDI combination.  Tooling will support collection, curation and maintenance of the LIVD file.  Tooling for the LIVD file will include the following capabilities:

a.       Storage, versioning, curation, and maintenance of all LIVD file data

b.       IVD vendor device characteristic input by form and API for submission of IVD/test kit information

c.       Consumer/IVD user look up tooling for reference and population of local EHR/LIS dictionaries.

d.       API access for IVD users for device and test kit coding alignment

  1. Terminology tooling to comingle and publish the unified LOINC, SNOMED and RxNorm terminology will be developed for ongoing maintenance and distribution of the unified controlled medical terminology to support primary and secondary uses of encoded laboratory, and other, health data.  Tooling will coordinate and manage licensing of LOINC, SNOMED and RxNorm in a unified format such that licensing terms for each individual terminology are maintained while realizing the needed benefits for a unified distribution of these controlled medical terminologies.  Tooling will further enable rapid authoring of new terms and logical organization of terms to support new IVDs and/or test kits that come to market.

  1. Initial implementation of terminology alignment and LIVD file alignment will focus on the top 1000 IVD lab tests with primary focus on tests with high-volume and high patient safety impacts.  Terminology will be developed, published and incorporated into the LIVD file with the additional data elements already discussed.

a.       Pilot phase

                                                               i.      Initial pilot testing of encoded, aligned data for the top 1000 labs will be performed at the five initial academic medical centers engaged initially with SHIELD.

  1. Lab data will be encoded in the LIS/EHR at each site

a.       EHR/LIS vendors at each site must be engaged to facilitate the incorporation of LIVD file information into the LAW data exchange and the LIS to EHR/EHR to EHR data exchange use cases.

b.       Public health reporting will also be tested at this point

  1. Lab data will be exchanged between each institution using a series of real-world test scenarios commonly encountered in US health care but highly dependent upon interoperable laboratory data for proper patient care.

  1. Secondary uses of these pilot data will also be tested for efficacy in public health surveillance, QA, analytics and observational research.

  1. Learnings (successes and failures) will be documented and addressed.

                                                             ii.      After a successful pilot, the program will be expanded to include X% of national and regional laboratories