Chat History:
00:16:37 Andrea Pitkus PhD, MLS(ASCP)CM: physician office labs?
00:16:40 Andrea Pitkus PhD, MLS(ASCP)CM: blood banks?
00:17:09 SZARFMAN: Good points Andrea
00:18:34 Dan Rutz: Re: measusrement of effectiveness - if there was a way to measure both accuracy and completeness of LOINC mapping on results, then the change from before implementing LIVD to after could be an interesting set of metrics.
00:18:53 Andrea Pitkus PhD, MLS(ASCP)CM: Great points Dan. Concur
00:19:10 Vaishali: Great suggestion, Dan. We can look into that as part of a broader discussion on metrics (beyond surveys).
00:19:11 Dan Rutz: Also count/distribution (by IP block/location?) of LIVD file downloads, either the spreadsheet or FHIR versions of the content, could help indicate how broadly it's being consumed.
00:19:23 Andrea Pitkus PhD, MLS(ASCP)CM: Interested if the survey might be distributed amongst SHIELD before it's sent for any feedback too.
00:21:29 Daniel Herman: For standardization/harmonization flag, as references change over time and there can be multiple, I wouldn’t expect boolean indicator to be sufficient
00:22:09 John Snyder: The change should be handled by versioning of the LIVD file itself.
00:22:55 John Snyder: Hi Andrea, the purpose would be to determine if the device UUID and calibration information is needed in order to aggregate data under a given LOINC code.
00:25:16 Gregory Pappas: Great job John. I will comment at the end how Greg R and I are working to intergrate this outstanding work into the broader plan.
00:25:59 Molly Polen (she, her, hers): @gregorypappas, could you also give us an update on conversations with potential funders like Rockefeller Foundation and also your internal convos with FDA leadership?
00:26:25 Gregory Pappas: Also great job Hung and Scott. Amazing to see this coming together. Completely agree in the need for staging, phases.
00:26:40 Gregory Pappas: Molly. I will speak at the end and address those issues.
00:27:40 John Snyder: Sorry, but I will have to drop off at 12:30pm and will miss the last half of the meeting.
00:27:49 Daniel Herman: @John, got it, good point. But, time will be complicated if goals include using historical and current results; references change over time; the need for HOW equivalent results are may depend on use cases
00:28:07 SZARFMAN: Are we planning to assess the widespread usability problems as recently demonstrated with the failure to perform of one of the most widely implemented sepsis prediction model in U.S. hospitals?
Wong A, Otles E, Donnelly JP, et al. External Validation of a Widely Implemented Proprietary Sepsis Prediction Model in Hospitalized Patients. JAMA Intern Med. Published online June 21, 2021. doi:10.1001/jamainternmed.2021.2626
Habib AR, Lin AL, Grant RW. The Epic Sepsis Model Falls Short—The Importance of External Validation. JAMA Intern Med. Published online June 21, 2021.
00:29:55 Andrea Pitkus PhD, MLS(ASCP)CM: @daniel. great points. It's true if lab results are mapped to the same LOINC and reference ranges, etc change, they may be comingled in displays etc depending on timeframe. I don't see this factor being able to be addressed by LIVD maps unless there are different LOINCS/assay identifiers,etc for those test results. IT would be a good issue to document on the confluence pages.
00:30:38 Gregory Pappas: Is there time for me to jump in on value propostion? I can write latere. On short. the overall value prop is for patient safety. Second is efficiency, less cost. beyond that it gets to specific stakeholders
00:30:49 Landim Araujo, Victor: Hi @johnsnyder, no problem if you have to drop. The meeting is being recording and will be uploaded to the Coordination Committee Page on Confluence
00:32:40 John Snyder: @Daniel, agree that trending of legacy data where the additional data elements are not available may require the automatic exclusion of certain data pools. Reference ranges is a little more challenging topic. We may be able to investigate using a result + Reference range normalization if the UUID + LOINC allow for data aggregation.
00:37:43 Daniel Herman: @andrea @Daniel. Agree LOINC resolution is not high enough to be used safely for trending. LOINC plus a binary identifier seems too simplistic to flexibly and powerfully handle the complexity of these assays and metadata. Happy to move conversation to confluence somewhere
00:38:16 Samuel McCash: I agree with you Luu.
00:38:26 Samuel McCash: Hung
00:38:57 Richard Moldwin: @Andrea - You are right. In the LIVD spreadsheet, 173 COVID "NAA" tests share the same LOINC code: 94500-6. There is also no common identifier for truly "comparable" tests. This obviously needs to be addressed, as Hung is now discussing.
00:40:30 Daniel Herman: Also worth noting that we can be creating resources and tools to help drive harmonization of the assays themselves
00:42:10 Andrea Pitkus PhD, MLS(ASCP)CM: It would be a huge step to help ($$, assistance) labs to all adopt the same standards (LOINC, SCT< HL7 v 2.51) to get folks all at the same foundational level
00:42:53 Andrea Pitkus PhD, MLS(ASCP)CM: LOINC, standards, and LIVD can't be adopted by those still on paper and faxed based lab reporting.
00:43:22 Samuel McCash: Agree we should be focused on value, but may need some government backing to make a regulation to use a government backed standard.
00:49:55 Andrea Pitkus PhD, MLS(ASCP)CM: Clinical validation of AI/Clinical Decision support is another issue, which is a bit tangential to SHIELD interoperability (a bit further downstream), yet extremely imporant
00:51:29 Joe Sirintrapun, MD: Do have specific goals now? Sounds like clinical interoperability (depending how how you define that) is the journey. Goals can be specific use case targets which we need in strategic alignment
00:52:18 Keith Campbell: Achieving Highly-Reliable Semantic Interoperability is the journey I see…
00:52:59 Keith Campbell: Tie in Highly Reliable Organization principles to the Interoperability Challenge.
00:54:39 Andrew Northup: Question for Strategic Alignment Cmte - are state public health authorities involved? Or are there plans to get them involved?
01:03:52 Gregory Pappas: Thanks everyone for the suppport
Add Comment