2018-06-18 - OBSERVABLE Meeting

1

Welcome & apologies

@Daniel Karlsson

2

Conflicts of interest

@Daniel Karlsson

• None stated

3

Previous minutes

@Daniel Karlsson

4

Free and Total Substances

@Daniel Karlsson

@Former user (Deleted)

@Yongsheng Gao

@Farzaneh Ashrafi

• See item discussion page.

• The current plan is to represent the base substance, the bound and the free substance as three siblings with a "substance X and derivatives" common subsumer.

• This would support the Observables use cases of allowing querying for substances without regard for "bound-ness", and to distinguish between bound, free, and total.

• Three patterns for Observables would be:

  • Free X Observable === Observable & component = Free X

  • Bound X Observable === Observable & component = Bound X (or X bound to Y)

  • Total X Observable === Observable & component X (base substance concept)

4

LOINC collaboration

@Suzanne Santamaria

@Farzaneh Ashrafi

@Daniel Karlsson

@Jim Campbell

• Technology preview and application of co0ncept model

• Likelihood

  • Two distinct cases were identified:

    1. Specifying certainty (or uncertainty) about the result of an observation. Examples: Fetal sex, Trisomy 21 present by "Dosage of chromosome specific cf DNA" method.

    2. Prediction of a possible future event based on varying amount of information about the patient, from population-only based assessment (i.e. prevalence) to assessment based on several variables. Examples include 1p36 deletion risk based on population risk, Trisomy 21 risk based on maternal age, to the CHADS-VASC instrument.

Example values:

5

Susceptibility

@Suzanne Santamaria

@Farzaneh Ashrafi

• Mapping of LOINC component for Susceptibility LOINC terms

  • Two alternatives for representing LOINC component / SNOMED Towards:

    1. Use substances: have to add combined substances for combined antibiotics, e.g. Amoxicillin+Clavulanic acid. Cannot represent strenghts. Cannot represent only vs. contains semantics(?).

    2. Use products: Depending on drug project, may represent only/contains, can represent strength, can represent combinations. This was the alternative proposed by the project group.

  • Strength-specified antibiotics are used in some susceptibility tests, particularly for mycobacteria, and the model would need to represent strengths somehow.

• There will likely be a new "diagnostic product" hierarchy parallel to the medicinal product hierarchy, reusing attributes and patterns from that hierarchy. The priority of this work is currently unknown

7

Next meeting

@Daniel Karlsson

• Next meeting is on on 2018-07-16

• As the LOINC project is on hold, we will shift to monthly Observables calls

8

Vancouver meeting

@Daniel Karlsson

• Request for a face-to-face?

• @Daniel Karlsson will request 1/2 day meeting

9

AOB

@Daniel Karlsson