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: Fri, 3 May 2019 09:44:42 +0000
To: "public-semweb-lifesci@w3.org" <public-semweb-lifesci@w3.org>
Message-ID: <9568c7ea-8965-6f01-5e0f-abd13ba316fc@pitt.edu>
Hi Colin,

We have drafted some changes that I think address your concerns. These are as follows:

1) Added the following statement to section 3.1  (https://w3id.org/hclscg/pddi#ex-warfarin-nsaids) :

'This example states NSAIDs generally and then refers to specific NSAID formulations in Contextual information/modifying factors (since the interaction is only relevant for bioavailable forms). A reasonable alternative approach would have been to create two PDDI knowledge artifacts - one that refers to "warfarin - systemic NSAIDs", and the other referring to "warfarin - non-systemic NSAIDs."'

2) Changed the 'topical diclofenac' value set to "topical or opthalmic diclofenac" and edited the evidence support to mention the evidence you provided in your comment. Made reference to this modified value set consistent throughout the Contextual information/modifying factors for the  warfarin-NSAIDs example, and wherever mentioned in the rest of the note

3) modified R6 (OCS) to include this statement: "The description MAY include more than one operational classification statements as long as each statement depends on a single set of contextual information/modifying factors. If there is more than one operational classification statement, each operational classification statement MUST be mentioned along with the specific contextual information/modifying factor that applies to the situation. " -- This makes it so that the issue of an inconsistent use of OCS should not occur because there is can only be either one OCS for the whole PDDI knowledge artifact or a specific OCS for each contextual information/modifying factor, but both are not allowed.

We would like to proceed to publish but I await your feedback in case you feel that these changes do not adequately address the concerns you raised.

kind regards,

On 4/2/19 2:54 PM, Macfarlane, Colin R. (ELS-LOW) wrote:
Dear Rich,

Thank you for your helpful responses to my comments. As I mentioned before, my concerns are largely peripheral to the model itself although it's great to now have a visual representation of the model included in the report.

Much as I respect the expertise of your group, I cannot accept that the potential bioavailability of ophthalmic formulations of diclofenac is the same as systemically administered formulations (and therefore, if I've understood you correctly, suitable to be grouped together as "NOT topical diclofenac"). The UK Summary of Product Characteristics for Voltarol Ophtha Multidose Eye Drops states, "No measurable levels of diclofenac could be found in humans after ocular application of diclofenac sodium eye drops". The FDA-approved SPL for Diclofenac Sodium Ophthalmic Solution 0.1% states, "Results from a bioavailability study established that plasma levels of diclofenac following ocular instillation of two drops of Diclofenac sodium ophthalmic solution, 0.1% to each eye were below the limit of quantification (10 ng/mL) over a 4-hour period. This study suggests that limited, if any, systemic absorption occurs with Diclofenac sodium ophthalmic solution". Gold Standard's Clinical Pharmacology text summarises the pharmacokinetics of ophthalmic diclofenac by saying, "Systemic absorption from the ophthalmic formulation is not clinically significant". Perhaps I've misunderstood your assertion but anyway, you are quite right to point out that the report presents exemplars.

I agree that references to drug classes need to be accompanied by definitions that allow users or readers of the system to unambiguously understand the membership of those classes.

I take your point about presenting a more comprehensive level of drug information but to do so with clarity will require a complex implementation. Given your scenario of warfarin + topical diclofenac the user (be that a system implementer or a clinician) is presented with:

  *   Drugs involved: warfarin + NSAIDs

  *   Mechanism of interaction: Antiplatelet effect of NSAIDs increases bleeding risk

  *   Clinical consequence: Increased risk of bleeding including GI bleeding, intracranial haemorrhage, cerebral haemorrhage

  *   Seriousness: Serious

  *   OCS: No special precautions

  *   Evidence: Topical diclofenac has low systemic absorption

So it needs to be made clear to the user that most of this information does not apply. The key piece of information here is the Evidence because without that the OCS appears to be inconsistent with everything else presented. But, according to your model, Evidence isn't a mandatory data item. I think this approach represents a risky strategy. The lower-risk approach would be to ensure the drugs involved are precisely defined (e.g. warfarin_systemic + diclofenac_topical) and then present correspondingly appropriate mechanism, consequence, etc information. If someone (for whatever reason) chooses not to bother with the Evidence, they have still got a clinically accurate set of data.

Kind regards
Colin Macfarlane

Colin Macfarlane
Medicines Optimisation Team Lead
ELSEVIER | Clinical Solutions

Richard D Boyce, PhD
Associate Professor of Biomedical Informatics and Clinical and Translational Science in the Clinical and Translational
Science Institute
Director of the Informatics Core for the Center of Excellence for Natural Product- Drug Interaction Research (NaPDI)
Faculty, Center for Pharmaceutical Policy and Prescribing
Faculty, Geriatric Pharmaceutical Outcomes and Gero-Informatics Research and Training Program
University of Pittsburgh
Office: 412-648-9219
Twitter: @bhaapgh
Received on Friday, 3 May 2019 09:45:09 UTC

This archive was generated by hypermail 2.3.1 : Friday, 3 May 2019 09:45:10 UTC