W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > May 2019

Re: 7 days left to comment (was Re: representing potential drug-drug interaction knowledge and evidence - Last commenting period for the W3C Community Group report

From: Boyce, Richard David <rdb20@pitt.edu>
Date: Tue, 14 May 2019 17:48:54 +0000
To: "public-semweb-lifesci@w3.org" <public-semweb-lifesci@w3.org>
Message-ID: <7abd6786-b85c-b8b9-30ca-51dc7c9fd3b2@pitt.edu>

Thank you very much for the suggestions. We have modified the note to 1) modified the first DDI example specifically on warfarin and systemic NSAIDs so that the mechanism, serious statement, and clin consequences align; and 2) added your suggested text to the drugs involved paragraph of the example (https://dbmi-icode-01.dbmi.pitt.edu/dikb-evidence/hcls-drug-drug-interaction/index.html#ex-warfarin-nsaids).

These were great suggestions and I am glad you made them. Please let me know if you would like to be listed in the author list of the note.

I will now ask the chairs of the  Semantic Web in Health Care and Life Sciences Community Group to accept the note for publishing.

kind regards,

On 5/7/19 5:47 AM, Macfarlane, Colin R. (ELS-LOW) wrote:
Dear Rich,

Thank you for taking the time to consider my comments and for making changes accordingly.

In your first change you describe the new text which says, "This example states NSAIDs generally and then refers to specific NSAID formulations in Contextual information/modifying factors..." but to my mind that still leaves the <Mechanism of Interaction>, <Clinical Consequences>, and <Seriousness> attributes carrying information that, rather than being "general" is actually specific to systemically available NSAID formulations.

My key point is that <Mechanism of Interaction>, <Clinical Consequences>, and <Seriousness> should provide clinically accurate information for the named <Drugs Involved>. If the PDDI is truly a class effect then it's fine to apply these attributes to the class-level interactants, but if not (as is the case with NSAIDs), then some more granular level of abstraction will be required to represent the interactants. I accept that the new text in Section 3.1 acknowledges this principle but I think there is still scope for implementations of your excellent model to be misleading since, as I've already mentioned, it could be argued that that content of the <Mechanism of Interaction>, <Clinical Consequences>, and <Seriousness> attributes in the standing text is actually specific to "systemic NSAIDs" rather than being applicable to "NSAIDs generally".

I wonder if this issue could be addressed by including a comment in Section 3.1 along the following lines:
"The drugs involved may be described at various levels of abstraction, from drug classes at one level to more closely defined medicinal product concepts at the other. The information carried by the following attributes must remain clinically accurate for the particular level of abstraction chosen for the interactants".

Finally, I note that the recent changes have introduced a consistent misspelling of "ophthalmic".

Kind regards

Colin Macfarlane
Medicines Optimisation Team Lead
ELSEVIER | Clinical Solutions
+44 (0)20 7424 4200 office
125 London Wall,
London, EC2Y 5AS

Richard D Boyce, PhD
Associate Professor of Biomedical Informatics and Clinical and Translational Science in the Clinical and Translational
Science Institute
University of Pittsburgh
Office: 412-648-9219
Twitter: @bhaapgh
Received on Tuesday, 14 May 2019 17:49:19 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 14 May 2019 17:49:20 UTC