Minutes - Referral WG 2021-12-14

Attendees

• Dr Alain Derom
• Bart Decuypere
• Frederik Lenaerts
• Hans De Keersmaecker
• Jean-Michel Polfiet
• Jos Bellen
• Karlien Erauw
• Kristof Jaubin
• Olivier Lothaire
• Philippe Cauchie (let at 4.30PM)
• Theo Schumacher
• Thibault Mahieu
• Tom Tollenaere
• Richard Francken
• Stefan Waegemans
• Werner De Mulder

Excused/Not present

• Alexis Van Zeveren
• Benny Verhamme
• Frédéric Istace
• Frederik De Kegel
• Joost Van Averbeke
• José Costa Teixeira
• Mieke Buckinx
• Nico Vannieuwenhuyze
• Nick Hermans
• Paul Neyens
• Peter Laridon
• Robert Nicolas
• Tom Fiers
• Toon Schiemsky

Agenda

• Rework on issues resulted from the pilot phase

Minutes

• Up-to-date list of issues can be consulted here
• Issue 33 : DiagnosticReport.code : coding same as composition type

• Set both to "laboratory report"?
• In examples we find e.g. Hematology studies, but our reports are typically more extensive than only hematology
• subject of lab report will be in the subject f.e. hematology will appear as chapter in lab report
• proposal: not to put the type at this level in the lab report, just call it laboratory studies
• Bart looked at LOINC codes (11502-2, not coded as RETAM yet at federal level)
• Philippe proposes to test this first (later today), be careful that an hematology study comes from another service
• let's allow it for now and await the tests

• Issue 23: How to encode comments

• We need to specify how to deal with different kinds of comments. Here an inventory and suggestion of process:

• a. Comment on a result of analysis - may be done .note to the result? But there are different cases possible:

• a.i. result field itself maybe comment or text (e.g. "Positive" or "No longer performed by lab; replaced by analyses XXX"
• a.ii. Comment in commentfield may be combined with number (or comment0 in result field)

• b. Comment regd. reference values (e.g. age-specifics) – in .text to the reference values?
• c. We often created dummy test codes to contain comments. E.g. serum aspect. Is this allowed? There are no LOINC codes for these dummies - will we allow every lab to use their own. Can we use this system? If not, how do we encode such comment?
• d. Comments regarding the laboratory (not the results). E.g. status of accreditation. We currently use dummy tests for that. Acceptable? If not, then how?
• e. Comments regarding tests performed by 3rd parties. We currently do that by means of flags (codes) next to results, and use a dummy test to explain the legend. Allowed? If not then how?
• f. Comments regarding administrative issue, e.g. missing information like gender, DOB. Today we use dummy test codes. Allowed? If not, then how?
• g. There is also a .conclusionCode - to use instead of dummy analyses? Or a combination? Dummy analysis have the advantage they can be linked to a sample (useful e.g. for serum aspect) - .conclusionCodes cannot do this.
• Do we want/need to agree on standards for the above? Or allwo every lab to do their own way?

• there are 3 LOINC codes availabel 55752-0, 94330-8 and 86468-6
• there is also a note field available for comments
• on analysis level there are also comments
• Issue can be moved to "for implementation" status

• Issue 31 given name (patient & practitioner)

• when a person has more than one given namen
• do it in one field in the FHIR resource or do it in seperate fields ?
• proposal to put it in one field as in the EID

• Issue 30: patient status during specimen collection

• Medina never fills this in so field should be optional
• on specimen resource there is a note field of type annotation where you can put any text

• Issue 25: same issue as 33 but on different level (at diagnostic report level while issue 33 is on level)

• move to "done"

• Issue 47 : RequestCode

• where to put the RequestCode : don't put it in the lab result
• the protocol will not always match the request
• the analysis order codes cannot be mapped 1-to-1 to lab codes

• Issue 46 How to flag a receiver patient inversion event?

• consensus: receiver should have a look at the identifier - if the patient is different then they need to remove the data connected to the original patient and have to link it to the data of the patient that is in the diagnostic report
• move to "to be implemented" , there is a test plan available for this

• Issue 45 : narrative

• Medina will put narrative everywhere

• Issue 19 unique identifier for non Belgian patients

• citizens that don't have a INSZ/SSIN or BIS number : how can they be identified ?
• possible solution: unique LAB RIZIV identifier + an identifier composed by the lab (the report id)

• NamingSystem will be vendor specific

• New issue: Create NamingSystem for BIS numbers

Next Meeting: on Tuesday Dec 21 4PM