D8.2 Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in Quality Assurance of Clinical Data

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Grant Agreement No. 875299 Project acronym: UNICOM Project full title: Up-scaling the global univocal identification of medicines in the context of Digital Single Market strategy Call identifier: H2020-SC1-DTH-2019 Deliverable D8.2: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in Quality Assurance of Clinical Data Version: 1.0 Status: Final Dissemination Level1: PU Due date of deliverable: 30.03.2024 Actual submission date: 05.04.2024 Work Package: WP8: IDMP and Clinical Care Lead partner for this deliverable: I-HD Partner(s) contributing: DWIZ, HL7, GNOMON, ZINDEX, BIDMC, SNOMED, IPU Deliverable type2: R Delivery date: 05/04/2024 Main author(s): Name Vander Stichele Robert Name Dipak Kalra 1 Dissemination level: PU: Public; CO: Confidential, only for members of the consortium (including the Commission Services); EU-RES: Classified Information: RESTREINT UE (Commission Decision 2005/444/EC); EU-CON: Classified Information: CONFIDENTIEL UE (Commission Decision 2005/444/EC); EU-SEC Classified Information: SECRET UE (Commission Decision 2005/444/EC) 2 Type of the deliverable: R: Document, report; DEM: Demonstrator, pilot, prototype; DEC: Websites, patent fillings, videos, etc.; OTHER; ETHICS: Ethics requirement; ORDP: Open Research Data Pilot D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 2 of 62 Other author(s): Robert Vander Stichele I-HD Leonora Grandia ZINDEX Dipak Kalra I-HD Jane Millar SNOMED Int. Alexander Berler GNOMON Caitriona Wray IEDOH Yuri Quintana BIDMC Robert Stegwee HL7 Nicole Veggiotti DWIZ Catherine Chronaki HL7 Alan Reilly IPU Geert Thienpont I-HD Christophe Maes I-HD Jens De Clercq I-HD Argiris Gkogkidis GNOMON Yuri Quintana BIDMC D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 3 of 62 Revision history Version Date Changes made Author(s) 1.1 05.01.2024 Initial outline RVS 1.2 02/02/2024 Second with plan of action RVS 1.3 14/02/2024 Second draft with full table of contents RVS 2.0 27/02/2024 Version for internal review RVS/DK 3.0 27/03/2024 Version for submission RVS/DK/ALL 4.0 02.04.2024 Final Submission Statement of originality This deliverable contains original unpublished work except where clearly indicated otherwise. Acknowledgement of previously published material and of the work of others has been made through appropriate citation, quotation or both. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 4 of 62 Deliverable abstract In this deliverable we discuss the final results of task T8.1 on IDMP and Clinical care. We explore what IDMP can mean for quality of care in pharmacotherapy, and how it can be integrated in clinical decision support systems. First, we discuss the grouping of active substances and medicinal products into pharmaco-therapeutic groups, with the ATC classification, SNOMEDCT and RxNorm, Classifications of international Medicinal product Dictionaries, the Standardised Drug Groups of the Uppsala Monitoring Centre, and a classification for medical education and patient information. Then, we describe a methodology for linking International Drug Classifications to National Medicinal Products, from national medicinal product packs to global pharmaceutical product, to the ATC level V class, and to Clinical Classifications. In a next chapter, we describe the initiatives to produce meaningful sets of medicinal products, standardised to IDMP, such as the UNICOM Pilot Product List, the UFIS Database, the T6.1 UNICOM FHIR Server, and the Minimal Data Set, containing all medicinal products for 4 substances from 10 countries. The objective of this description is to illustrate the importance of having IDMP-compliant data for demonstration purposes. Finally, the use of this Minimal Data Set in the Patient Facing APP and in decision support systems is discussed. This work concludes by stating that the implementation of the ISO/CEN standards will bring precision and robustness to the application of decision rules, expressed in general drug class statements, in the national Medicinal Product Dictionaries, allowing decision support systems to cross the borders of the member states in Europe in a feasible way. This document contains material, which is the copyright of the members of the UNICOM consortium listed above and may not be reproduced or copied without their permission. The commercial use of any information contained in this document may require a license from the owner of that information. This document reflects only the views of the authors, and the European Commission is not liable for any use that may be made of its contents. The information in this document is provided “as is”, without warranty of any kind, and accept no liability for loss or damage suffered by any person using this information. © 2019-2023. UNICOM consortium members. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 5 of 62 TABLE OF CONTENTS Revision history ....................................................................................................................................... 3 Deliverable abstract ................................................................................................................................. 4 Deliverable review ................................................................................................................................... 8 List of abbreviations ................................................................................................................................. 9 1 Executive summary ........................................................................................................................ 11 2 Content of the deliverable .............................................................................................................. 12 2.1 Contents of the deliverable .................................................................................................. 12 2.2 Authorship and responsibilities ............................................................................................ 13 3 Methodology to describe Pharmacotherapeutic Drug Classes ...................................................... 14 3.1 Overview of drug classifications ........................................................................................... 14 3.1.1 Examples of high level international classifications ......................................................... 14 3.1.2 Multilingual examples ...................................................................................................... 15 3.1.3 A classification for medical education and patient product information ........................... 15 3.1.4 Other examples................................................................................................................ 16 3.1.5 Standardized drug groups (SDGs) .................................................................................. 16 3.2 Methodology for linking International Drug Classifications to National Medicinal Products ..... 17 3.2.1 From national medicinal product pack to global pharmaceutical product ....................... 17 3.2.2 From global pharmaceutical product to the ATC level V class ........................................ 27 3.2.3 From ATC Level V to Clinical Classifications .................................................................. 36 3.2.4 Closing the loop from medicinal product pack to pharmacotherapeutic drug classes .... 36 4 Attempts to produce complete samples of medicinal products, standardised to IDMP, for a limited number of substances. .......................................................................................................................... 39 4.1 The UNICOM Pilot Product List ........................................................................................... 39 4.1.1 Rationale .......................................................................................................................... 39 4.1.2 Methodology .................................................................................................................... 39 4.1.3 Results ............................................................................................................................. 40 4.2 UFIS and the UNICOM T6.1 FHIR SERVER ....................................................................... 40 4.3 The minimal data set on 4 substances, based on “data as is“, created in Work Package 8 41 4.3.1 Motivation for the initial decision to create a minimal data set. ....................................... 41 4.3.2 Constraints of this work ................................................................................................... 41 4.3.3 Countries involved ........................................................................................................... 42 4.3.4 Sources per country ......................................................................................................... 42 4.3.5 Methodology of the data collection process and central standardisation ........................ 42 4.3.6 Further development of the data collection process ........................................................ 43 4.3.7 Overview of the Minimal data sets per substance and per country ................................. 43 5 Demonstrate the use of this repository in the application of decision support systems across Europe, and in the tools for patients to navigate drug class names in product labelling .................................... 44 5.1 IDMP and Drug information.................................................................................................. 46 D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 6 of 62 5.1.1 Electronic Product information (ePI) ................................................................................ 46 5.2 Drug Information Centres ..................................................................................................... 47 5.3 IDMP and Guidelines ........................................................................................................... 48 5.4 IDMP and Computerized Decision Support (CDSS) ............................................................ 49 5.5 IDMP and substitution rules and INN prescribing ................................................................ 52 5.6 IDMP and patient Facing Apps ............................................................................................ 56 5.7 IDMP and the quality of clinical data .................................................................................... 57 5.7.1 Using the medication list to improve the quality of medical documentation .................... 57 5.7.2 IDMP and the Patient Summary ...................................................................................... 58 6 Conclusions for IDMP and Clinical Care ........................................................................................ 60 7 Annexes .......................................................................................................................................... 61 7.1 Annex 1. Drug Classification for Medical Education ........................................................... 61 7.2 Annex 2. Repository of Potentially Inappropriate Medication (PIM) ................................... 61 7.3 Annex 3. Administrable dose forms in EDQM..................................................................... 61 7.4 Annex 4. List of Substance Codes for the UNICOM Pilot Product List ............................... 61 7.5 Annex 5. ATC codes pertaining to the UNICOM Pilot Product List .................................... 61 7.6 Annex 6. Presentation to the Pistoia Alliance (August 2023) ............................................. 61 7.7 Annex 7. Presentation to the WHO Collaborating Centre .................................................. 61 7.8 Annex 8. The Duran List of Anticholinergics ....................................................................... 61 7.9 Annex 9. Triggers for completing clinical data based on PIM ............................................. 61 8 Publications from this work ............................................................................................................. 62 LIST OF FIGURES Figure 1. WHO Standardised Drug Groups (SDGs) .......................................................................................... 17 Figure 2. Actual and Virtual concepts ............................................................................................................. 20 Figure 3. Example of a Medicinal Product ...................................................................................................... 21 Figure 4. Expression of strength and patterns of dose form ........................................................................... 22 Figure 5. Methods for PhPID Definition .......................................................................................................... 22 Figure 6. Proposed WHO/UMC/PhPID Repository .......................................................................................... 23 Figure 7. Legacy conversion versus prospective implementation .................................................................... 25 Figure 8. RxNorm Virtual drug concepts ......................................................................................................... 26 Figure 9. SNOMED-CT virtual concepts ........................................................................................................... 27 Figure 10. Linking ATC to PhPID ..................................................................................................................... 29 Figure 11. Moiety and moiety+modifier codes ............................................................................................... 30 Figure 12. Value Set of EDQM ISI characteristic .............................................................................................. 33 Figure 13. Dose form aggregation .................................................................................................................. 34 Figure 14. Link from PhPID to ATC .................................................................................................................. 36 D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 7 of 62 Figure 15. The 5 connected levels of the ATC Classification ............................................................................ 36 Figure 16. Construction of concepts ................................................................................................................ 37 Figure 17. Double loop towards the ATC ........................................................................................................ 38 Figure 18. UNICOM Pilot Product List ............................................................................................................. 40 Figure 19. A chain of concepts ........................................................................................................................ 44 Figure 20. Levels of specificity in drug information ......................................................................................... 49 Figure 21. An example of pharmacodynamic nephrotoxic interaction ........................................................... 51 Figure 22. Diversity in rules for substition and INN Prescribing in UNICOM EU member States ...................... 53 Figure 23. Repairing the quality of medical record through the medication list ............................................. 57 Figure 24. Quality control of semantic operability of electronic health records. ............................................ 59 D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 8 of 62 Deliverable review Internal reviewer: Luc Nicolas (ETHEL) External reviewer: ......................................... Answe r Comments Type * Answe r Comments Type * Is the deliverable in accordance with the Description of Action? ☒ Yes ☐ No The document refers clearly to the tasks and sub-tasks of the DoA ☐ M ☐ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a the international State of the Art? ☒ Yes ☐ No Make sure to provide external links when referring to external sources. ☐ M ☒ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a Is the quality of the deliverable in a status that allows it to be sent to European Commission? ☐ Yes ☒ No A number of clarifications need to be made; please also check carefully all editing proposed. Some ☐ M ☒ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a that needs improvement of the writing by the originator of the deliverable? ☒ Yes ☐ No Aside from comments and editing, please also make sure that all abbreviations are listed and ae exported from the ☐ M ☒ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a that needs further work by the Partners responsible for the deliverable? ☒ Yes ☐ No The annexes cannot be provided “as is” and need to be reformatted so that included in the document (with a possible open link to the full resource) ☒ M ☐ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a Is the structure and contents of the deliverable structured, logical and easy to understand? ☒ Yes ☐ No The document is well structured and to the point but wording is sometimes a bit fuzzy. ☐ M ☒ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a suitable to meet its intended scope? ☒ Yes ☐ No ☐ M ☐ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a Is in** conformance with UNICOM deliverable template? ☒ Yes ☐ No ☐ M ☐ m ☐ a ☐ Yes ☐ No ☐ M ☐ m ☐ a * Type of comments: M = Major comment; m = minor comment; a = advice D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 9 of 62 List of abbreviations Abbreviation Complete form AMP Actual Medicinal Product AMPP Actual Medicinal Product Pack API Application Programming Interface ATC Anatomical therapeutic Chemical Classification ATM Actual Therapeutic Moiety BCFI Belgian Centre for Pharmaco-therapeutic Information CAP Centralised Authorisation Procedure CAS Chemical Abstract Service CD Clinical Drug CDF Complex Dose Form CEF Connecting Europe Facility CEN European Standards Commission / Comité Européen de Normalisation CMT Combined Term DCP Decentralised Procedure DDD Defined Daily Dose DPP Defined Daily Dose per Package EDQM European Directorate for the Quality of Medicines eHDSI eHealth Digital Services Infrastructure EMA European Medicines Agency EPC Established Pharmaceutical Classes ePI Electronic Product Information EU European Union FMD Falsified Medicines Directive FDA Food and Drug Administration FHIR Fast Healthcare Interoperability Resources GTIN Global Trade Item Number EHDS European Health Data Space EHR Electronic Health Record HL7 Health Level 7 IDMP Identification of Medicinal Products D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 16 of 62 The Drug Class entries in this classification have been translated in English, Greek and Italian for the purpose of the work on Patient Facing Apps in Task T8.3. These entries are transferrable to a Terminological Markup Framework Multilingual Technology (ISO/CEN Terminological Markup Format, ISO 16642).12 3.1.4 Other examples Finally, examples exist where complicated drug classes have been formally expressed as listings of corresponding ATC codes at the fifth level of the ATC (the level of the substance). One example is the class of anticholinergic drugs, with examples in many therapeutic indications, either deliberately used for their anticholinergic effect, or related to anticholinergic side-effects. One such example is the international Duran List, labelling a thorough selection of 100 anticholinergic substances (50 with strong and 50 with weak effect)13 Another example is the operationalization of validated international lists of explicit criteria for (in)appropriate prescribing based on Beers List, STOPP/START and EU-7 PIM list. The substances and drug classes mentioned in these lists were all defined in terms of ATC V codes or combination of codes. (See Annex 2). Recently, a systematic analysis of drug brand names in FAERS was conducted, with linking to the active ingredient and to the ATC classification.14 3.1.5 Standardized drug groups (SDGs) Finally, the WHO collaborating Centre for Pharmacovigilance, the Uppsala Monitoring Centre (UMC) has produced the Standardized Drug Groups (SDGs),15 defined as any grouping of medicines having one or several properties in common. These Groups are defined in standardized WHO Drug terminology for substances (linked to the variants), and in terms of ATC level V codes. 12 Federica Vezzani, Giorgio Maria Di Nunzio. Multilingual digital terminology: Introduction to the special issue. Digital Scholarship in the Humanities, Volume 38, Issue Supplement_1, June 2023, Pages i1–i5, https://doi.org/10.1093/llc/fqad028 13 Durán CE, Azermai M, Vander Stichele RH. Systematic review of anticholinergic risk scales in older adults. Eur J Clin Pharmacol. 2013 Jul;69(7):1485-96. 14 Fusaroli M, Giunchi V, Battini V, Puligheddu S, Khouri C, Carnovale C, Raschi E, Poluzzi E. Enhancing Transparency in Defining Studied Drugs: The Open-Source Living DiAna Dictionary for Standardizing Drug Names in the FAERS. Drug Saf. 2024 Mar;47(3):271-284. 15 Lagerlund O, Strese S, Fladvad M, Lindquist M. WHODrug: A Global, Validated and Updated Dictionary for Medicinal Information. Ther Innov Regul Sci. 2020 Sep;54(5):1116-1122.. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 17 of 62 Figure 1. WHO Standardised Drug Groups (SDGs) All these different approaches to drug grouping can be connected to the ATC classification, and from there to National Drug Dictionaries, who very often code all their medicinal products using this classification. This link to the medicinal products is a crude link, based on substance, and does not take into account granular dose form and strength, although for a number of substances, the ATC Classification assigns more than one Route of Administration, with sometimes different ATC codes. IDMP is expected to be instrumental to add more precision and robustness to this linking. Linked Activity 1: Drug Classifications 3.2 Methodology for linking International Drug Classifications to National Medicinal Products 3.2.1 From national medicinal product pack to global pharmaceutical product Discussion on the concept of Medicinal Product Pack In most developed countries, medicinal products are distributed in discrete packs (outer package), containing an inner package, enclosing the manufactured item, and the written drug information as a folded leaflet. This is in contrast with bulk dispensing where medicinal products are present in the pharmacies in large containers and dispensed to the patient in a white paper bag or orange plastic bottle, with a written posology instruction (e.g.to be taken twice a day). In 2002, the European Union, including the United Kingdom, moved from bulk dispensing (pills in large containers shipped to the pharmacy and then dispensed in white paper bags) fto unit-of-use distribution (with an outer package, containing an inner package, the manufactured items, and a User Leaflet in the Standardised Drug Groups in WHODrug Terminology 1. Analgesia producing opoids 24. Drugs interacting with CYP2C19 2. Antiangiogenic drugs 25. Drugs interacting with CYP2C8 3. Antiarrhythmics 26. Drugs interacting with CYP2C9 4. Antidepressants 27. Drugs interacting with CYP2D6 5. Antiemetics and antinauseants 28. Drugs interacting with CYP2E1 6. Antihaemorrhagic drugs 29. Drugs interacting with CYP3A 7. Antihistamines 30. Drugs interacting with OATP 8. Antihypertensives 31. Drugs interacting with P-glycoprotein (P-gp) 9. Antithrombotic drugs 32. Drugs interacting with UGT 10. Benzodiazepines 33. Drugs used in diabetes 11. Blood and related drugs 34. Drugs used in pain therapies 12. Corticosteroids 35. Essential fatty acids 13. Disease-modifying antirheumatic drugs (DMARDs) 36. Hormone replacement therapy 14. Diuretics 37. Immunomodulators 15. Drugs acting on gonadotropin-releasing hormone (GnRH) receptors 38. Monoclonal antibodies 16. Drugs acting on NMDA receptors 39. Nonsteroidal anti-inflammatory drugs (NSAIDs) 17. Drugs for gastric acid related disorders 40. Phosphodiesterase (PDE) inhibitors 18. Drugs for obstructive airway diseases 41. Psychoanaleptics 19. Drugs for ulcerative colitis 42. Radiopharmaceuticals 20. Drugs interacting with BCRP 43. Statins 21. Drugs interacting with CYP1A2 44. Systemic anti-infectives 22. Drugs interacting with CYP2A6 45. Vaccines 23. Drugs interacting with CYP2B6 D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 18 of 62 package. The requirement of a user leaflet in understandable language for the public is an example of a legal obligation for an industry to communicate with the customer in an understandable way. In most countries with unit-of-use distribution of medicinal products, each package has its own number (as such and/or represented by a barcode). This number can originate from the distribution sector, from the marketing research sector, or from the regulatory authorities. It is to be distinguished from the GTIN number, which is a number given by the industry, and related to the specific combination of inner package and manufactured Item. There is also the batch number, from the industry, allowing back tracing to the production process. In addition, there can also be serialisation numbers (unique number per pack as 2D barcodes, to check against Falsified Medicinal Products. Most Medicinal Product Dictionaries, whether maintained by the regulatory authorities, the pharmacists unions, the software vendors, independent drug information centres, the industry, or private publishers will list the National Drug Code, identifying each individual pack. The number is unique and identifies the company marketing the product, the substances (combinations of active and inactive ingredients), the dose form, the type of pack, the pack size. There are national rules which need to be applied when a National Drug Code needs to change when variations of any of these elements occur. In some less developed countries, it could occur that the national drug code from the distribution sector is not be known by the regulator or by the Medicinal Product Dictionary publisher, because of a lack of communication between systems. In the drug models, explained in WP9 (D9.1), this lowest concept of identification is named the Actual Medicinal Product Package (AMPP). There may be subtle differences in the delineation of this concept in each country depending on whether differences in pack type or manufactured items are taken into consideration. It is at this level that the IDMP ontology provides the Pack identifier (PCID), to be constructed and maintained by rules that are common to all countries. The lifecycle and rules that govern the updating of these concepts may show subtle differences. In some countries, this concept is defined by the mathematical pack size (number of tablets in a pack in case of presentation strength or total volume of a syrup in case of concentration strength). In other countries, the type of package will also be taken into account and then the National Drug Code (NDC) will differ for a pack with 5 blisters of 6 tablets, and a pack of 6 blisters of 5 tablets, both with a pack size of 30), In the USA, the NDC will even differ when the same medicinal product is allowed to be marketed with two different manufactured items (e.g. grey oval tablets or white round tablets, coming from two different generic bulk manufactures), a situation that will not or seldom occur in Europe. For the description of the pack, which is the level of the record in a database, it is important to have a unique ID (a primary key). However, the nature of this primary key can differ country by county (taking into account or not taking into account small differences in pack type or manufactured item). For comparison between countries, it is therefore important to precisely identify a comparable concept for pack (AMPP) that is based solely on the mathematical pack size, ignoring differences in pack type. The issue of pack type is not an issue of great clinical relevance and does not or very seldomly play a role in the drug choice process of the physician. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 19 of 62 Discussion on the concept of Medicinal Product A Medicinal Product is a concept which aims at the aggregation of medicinal product packs, with the same marketing company, the same substance(s), the same dose form, and the same strength, within the jurisdiction of a national marketing authorisation authority. In IDMP the substance must be specified if a modifier is present. The dose form must be termed from the value set of the granular EDQM dose form terminology. The strength must be normalised according to agreed business rules (e.g. a choice to be made between a gel of 3%;100 ml gel with 3 grams substance; or a gel of 30 mg/ml). These business rules for normalisation of the expression of strength are complicated, depending on the pattern of dose form, and still debated with the partners of the GIDWG (Global IDMP Working Group), under the leadership of the WHO. In the Drug Models this concept would be called the Actual Medicinal Product concept (AMP). In IDMP ontology the concept is the Medicinal Product, with again universal rules to build and maintain the Medicinal Product ID (MPID). The labelling of the different packages aggregated under this concept may or may not be the same for all packages. Sometimes small packages with a low strength may have different indications and different labelling. Determining the characteristics of a company is not always easy, as there are many differences: multinational company, regional subsidiary, national company, manufacturer, distributer, re-packager, etc. One can focus on the national marketing authorization holder. These names are not always standardised. For example in the US database RxNorm, where the information is entered by the company at the moment of an application for a new National Drug Code, the company name may be misspelled and this will be propagated throughout the concept tree, and may hamper the correct creation of the concept of medicinal product. In Europe, with the provision of centralised SPOR Terminology Services for organisations (the Organisation Management System OMS) this problem can be minimised. Discussion on the concept of Actual Therapeutic Moiety An actual therapeutic moiety is an aggregation of medicinal products, marketed by the same company and with the same substance(s). The label in the country is usually a phantasy brand name without the company name (e.g. Tenormin®) or with the company name (e.g. Norvasc® Pfizer®). In case of generics, the label consists of the generic substance name + the company name (e.g. simvastatin TEVA® or amlodipine Sandoz®). Linguistic similarity between these brand name labels in different countries do not necessarily mean that the products in these countries are identical. The marketing authorisation and the link to the official label can be on this level, but also be more specific on lower levels, depending on the medicinal product. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 20 of 62 Discussion on the concept of Pharmaceutical Product Description of the concept In the IDMP architecture the Pharmaceutical Product is a concept subordinated to Medicinal Product. It is the global representation of the elements in a Medicinal Product that are not country-specific or company-specific. The Pharmaceutical Product Set is the combination of 3 elements. • Substance (with modifier if any) • Administrable dose form (granular value set of EDQM) • Strength (normalised by business rules) As such it is a concept that groups all medicinal products from different countries that carry these 3 elements, and will be identified by a global Identifier: the PhPID. Four levels are distinguished for different combinations between these 3 elements: • PhPID Level IV : Substance + dose form + strength • PhPID Level III : Substance + dose form • PhPID Level II : Substance + strength • PhPID Level I : Substance PhPID Level I is a concept that binds all actual therapeutic moieties with the same substance (+ modifier if any) from any country. PhPID Level IV is the full global representation of the essence of any national medicinal product from any country. Figure 2 provides a summary of the concepts and their relationships. Figure 2. Actual and Virtual concepts In figure 3, this is illustrated by a concrete medicinal product: Actual Therapeu�c Moiety Actual Medicinal Product Actual Medicinal Product Pack Virtual Medicinal Product Pharmaceu�cal Product ID PhPID Na�onal Medicinal Product Pack Code Na�onal Medicinal Product Code Na�onal Brand code MPID PCID PS. The marke�ng authorisa�on (number) can span all of these na�onal concepts PhPID D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 21 of 62 Figure 3. Example of a Medicinal Product Creation of the global identifier The debate is still raging on the level of granularity that is needed for the creation of the PhPID. The most strict and granular definition is the one mentioned above: substance (with modifier if any); granular EDQM dose form; strength (normalised according to business rules). But several alternatives have been proposed, based on relaxing the obligation to explicitly name the modifier for the substance, if any, and considering less granular value sets for the dose forms. The argument is that for regions outside Europe the attribution of the granular administrable EDQM dose formis too difficult, too cumbersome, and too error prone. An aggregation is proposed based on the unique combinations of 4 characteristics of the administrable dose form (basic dose form, method of administration, release-characteristics, intended site). This proposal has been explored in depth in an analysis of the EDQM terminology and the results of this analysis have been published16. In the conclusions of this research, it was clearly stated that the mechanical grouping based on unique combinations of these characteristics was not always creating distinct, clinically relevant groups. Many groups needed to be concatenated or split. The level of granularity proves to become similar to the granularity of RxNorm, but insufficient for clear distinction. Within the context of the work in the GIDWG working group, however, this proposal is still sufficient, under impulse of the FDA and the non-European Regions of the WHO. For the normalisation of strength, a global solution seems however to be in the making, based on business rules for strength expression according to patterns of dose form. These business rules provide guidance on which dose forms must be expressed in presentation strength and which dose forms in concentration strength. In addition, the rules indicate the right choice for various ways to express concentration in percentage, as weight over volume (125 mg/5ml), or as a true concentration (25mg/ml). In figure 4, it is indicated how the expression 50 units/ml is transformed in codes to be fed into the algorithm, creating the PhPID. 16 Vander Stichele, R.H.; Roumier, J.; van Nimwegen, D. How Granular Can a Dose Form Be Described? Considering EDQM Standard Terms for a Global Terminology. Appl. Sci. 2022, 12, 4337. https://doi.org/10.3390/ app12094337 Norvasc Pfizer Norvasc 10 mg CPS Norvasc 10 mg 30 cps amlodipine besilate capsule, hard 10 mg Na�onal Drug Pack Code Na�onal Product Code Na�onal Brand code ITA amlodipine besilate PFIZER ITALIA Srl capsule, hard 30 x 10mg/ ITA amlodipine besilate PFIZER ITALIA Srl capsule, hard 10mg/ D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 22 of 62 Figure 4. Expression of strength and patterns of dose form The choices made for the granularity in the representation of those 3 elements (substance, dose form, strength) has of course a direct impact on the resulting global identifier. Figure 5. Methods for PhPID Definition The decision to opt for one of these methods has a direct impact on the workload related to the legacy conversion of older medicinal products to IDMP and on the maintenance of the identification process. But it has also important consequences in terms of granularity and precision. It can be argued that if IDMP wants to make a difference as a system for medicinal Product Identification, the emphasis should be on precision, and hence the first method should be preferred. This method of producing PhPID with a high level of granularity and precision creates the rock-solid foundation for precision in medicinal product identification and for the solid building of higher levels of aggregation that might be more clinically relevant. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 23 of 62 Coding issues Substances are currently coded in Europe through the EU-SRS, with output to the NCAs through the SMS SPOR services. The cleansing of this database has been a tremendous work, conducted in collaboration with the US Food and Drug Administration, and with the vision to promote a Global Substance Identifier (GSID), both for the moiety and for the moiety+modifier. For the dose form, based on EDQM terminology, there are codes proper to EDQM, but the SPOR Services create a new coding system for each value in the EDQM value sets, with an automatic mapping between both. This SPOR code is available to NCAs through the EMA portal but may be more difficult to access by other stakeholders. Mapping lists between the term labels and the two coding systems can be easily provided. However, a decision is to be made on which of the two codes will be fed into the PhPID creating algorithm. On a global scale, in the GIDWG, the tendency is to work with the original EDQM Code. EMA is more inclined to work with the equivalent SPOR code. Again, this choice of code systems to be fed in the algorithm for creation of PhPIDs can lead to different numbers, depending on the choice made. Governance It is obvious that the creation and maintenance of a coding system such as PhPID will require a lot of coordination. The WHO CC for Pharmaco-vigilance Uppsala Monitoring Centre (UMC) has volunteered to take up the role of keeping a repository of PhPIDs and coordinate the consistency and maintenance rules for this endeavour. Figure 6. Proposed WHO/UMC/PhPID Repository However, some National Competent Authorities in Europe claim that the rules for the production of PhPID can be formulated in a simple way, and that the responsibility for producing PhPIDs can be kept at the national level, with consistent results, and possibly with some international curating method. Scaling up the PhPID production -organizing data streams of structured drug information from volunteer countries UMC-PHPID internal service New PHPID Valida�on Update PhPID Interna�onal WG Publish PHPID Assignment solved Map PhPID Country data Country data Country data Country data Data extract D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 24 of 62 Discussion on the advantages of granularity in IDMP in defining the Pharmaceutical Product For the creation of substitution rules in cross-border exchange of ePrescriptions and the medication lists in the International Patient Summary, the granular description of the administrable dose form and the naming of the modifier for the active ingredient of a medicinal product provides a more solid basis to decide by which national medicinal product a medicinal product from another EU country can be substituted. The same is true for conducting cross-national comparisons of the extent and variability of national medicinal product dictionaries, and for conducting detailed price comparisons and comparisons of the reimbursement rules, which can vary by substance modifier, dose form, and pack size. For pharmaco-epidemiology research, it is important that the information on drug exposure can be interpreted at a granular level. Safety profiles of products can be influenced by subtle differences in dose form or modifiers of active ingredients. In D8.4 of UNICOM, a protocol is described for a study comparing the cardiovascular side effects of diclofenac sodium with diclofenac potassium. Finally, for clinical reasons, the details on the granularity of dose forms or the modifier of active substance may not always be of relevance. However, to build in a robust way clinical meaningful aggregation concepts for prescribing by international non-proprietary name (INN prescribing) it is necessary to start from a precise and global description of the national medicinal products, made possible by the implementation of IDMP standards. IDMP implementation for new and older medicinal products For new medicinal products evaluated by EMA in the Centralized Authorisation Procedure (CAP), IDMP compliance will be guaranteed thanks to the introduction of the Product Management Service (PMS) and the creation of the electronic application form (eAF). Currently this is still work in progress, as only variations are considered by now. By contrast, for the older medicinal products, and for Decentralized Procedures (DCP) the national NCAs face the formidable task to convert the description of these products to the new IDMP standards. National medicinal Dictionaries contain easily between 8.000 to 15.000 Medicinal product packs per country. And there are 27 Member States. Hence, this legacy conversion is a formidable task. For very old medicinal products, the existing labelling may lack some of the essential information such as the nature of a salt or ester (modifier) of the active ingredient (moiety). Digging up such information has been labelled with the term „pharmacoarchaeology“. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 25 of 62 Figure 7. Legacy conversion versus prospective implementation Discussion on the relationship with other virtual drug schemes In this project, we have explored the relationship between IDMP Pharmaceutical Product concept and the virtual drug models of RxNorm and SNOMED CT.17 In the US, RxNorm, produced by the National Library of Medicine, the concept that is most closely related to the PhPID is the concept “Clinical Drug“. However, this concept is built upon the active moiety of the molecule with therapeutic effect, without the modifier (salt of ester). Although this information is present in the RxNorm system, it is not used for the construction of the concept of “Clinical Drug“. In addition, for dose form, another building block of the concept, the terminology of dose form from RxNorm is used, which is much less granular then the EDQM dose forms, and without the formalized definitions, descriptors, and characteristics. The expression of strength might be decided by companies during the application for a National Drug Code, and might not always be normalised between companies. 17 Vander Stichele R, Kalra D. Aggregations of Substance in Virtual Drug Models Based on ISO/CEN Standards for Identification of Medicinal Products (IDMP). Stud Health Technol Inform. 2022 May 25;294:377-381. doi: 10.3233/SHTI220478. Index Date Prospec�ve DADI-Project (industry => Agency) IDMP-Compliant Registra�on Retrospec�ve Substance cleansing EDQM standardiza�on Strength Normalisa�on Pharmaco-archeology Perspec�ve on future and history of IDMP implementa�on D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 32 of 62 73 values (65 pharmaceutical dose forms (60 administrable) and 8 combined pharmaceutical dose forms, each in combination with a container or administration device) SNOMED-CT works with pharmaceutical dose forms and the value set is slightly less than 400. The same apparatus of descriptors and characteristics is present in the system, albeit with slightly different value sets. This should come as no surprise because EDQM dose terms and SNOMED CT dose terms both follow the logic of the ISO/CEN standard EN ISO 11239 on Pharmaceutical Dose Forms. RxNorm has a much lower level of granularity with a value set of dose forms not exceeding 200; no distinction between manufactured dose form and administrable dose form is made, and there is a limited aggregation in dose form groups. There is a formal mapping from SNOMED-CT dose forms to EDQM dose forms, initiated in the UNICOM project28. The mapping of EDQM Dose Forms to SNOMED CT is based on March 2023 EDQM version and January 2023 International Edition of SNOMED CT. Individual mapping from EDQM to SNOMED terms were classified as "Equivalent", "Narrower than", or "Broader than“. The successful mapping rate was estimated to be 60%. There was no mapping attempt in the opposite direction, from SNOMED-CT to EDQM dose form terms. RxNorm has been connected to the SNOMED virtual Drug model, including for dose forms. It has however been acknowledged that more work is needed to align the dose forms of RxNorm and SNOMED-CT.29 An attempt was made during the UNICOM project to align the EDQM dose forms with RxNorm dose forms, illustrating the less precise value set of RxNorm.30 The preceding development of the ontology of EDQM dose forms (which is detailed in the next chapter) was crucial for this alignment process. A similar attempt could be made to align also with the dose forms of SNOMED-CT. Four limitations of the EDQM terminology were observed during this work. • EDQM (and IDMP) does not work with the property of “splitability” for several solid oral dose forms, such as tablets or capsules that can be split in two (divules). Hence, standardisation leads here to some loss of information, illustrating the continued need for national information, added by the National Agency or the National Medicinal Product Dictionary. • The notion of systemic effect of the dose form is not acknowledged. While this may be an attribute to be given at the level of the medicinal product (e.g. nasal sprays with local effect, and nasal sprays with systemic effects), it is probably possible to assign to most of the dose forms the notion of “systemic effect“, “local effect“, leaving a few dose forms without the label ”unclear effect“. • The issue of splitable, crushable, and mixable dose forms is not considered by EDQM, as this may be properties that vary at the product level. However, these issues can be important for drug safety31. Maintenance of this information is currently at the national level, either in medicinal product dictionaries or in hospital pharmacy systems. • In the EDQM system, for each Pharmaceutical Dose form that can (or must) be transformed, the nature of the transformation process is described in a characteristic (transformation). The result of that process, the administrable dose form, is not formally described. Although this is often obvious it is not always the case. The EDQM organisation reacted positively to this remark and prepared a formal file where for all Pharmaceutical Dose form the administrable dose form, resulting from a transformation 28https://confluence.ihtsdotools.org/display/USRG/Mapping+Guidance+for+EDQM+to+SNOMED+CT+Pharmaceutical+Dose+Fo rm+Mapping 29 Nikiema J, Bodenreider O. Comparing the representation of medicinal products in RxNorm and SNOMED CT - Consequences on interoperability. CEUR Workshop Proc. 2019 Aug;2931:F1-F6. 30 Karapetian N, Vander Stichele R, Quintana Y. Alignment of two standard terminologies for dosage form: RxNorm from the National Library of Medicine for the United States and EDQM from the European Directorate for the Quality in Medicines and Healthcare for Europe. Int J Med Inform. 2022 Sep;165:104826. doi: 10.1016/j.ijmedinf.2022.104826. 31Senger C, Seidling HM, Quinzler R, Leser U, Haefeli WE. Design and evaluation of an ontology-based drug application database. Methods Inf Med. 2011;50(3):273-84. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 33 of 62 process is given. A first draft was recently updated, with the intention to continue the updating process (see Annex 3). The value set of the Intended Site (characteristic) is rather small (N=27, with some combinations possible), and may be insufficient for the construction of an INN prescribing instruction with substance, a dose form, and strength section. The precision of the EDQM dose forms (N=443) may be, on the one hand, very instrumental for a precise global identification, but, on the other hand, too elaborated for clinical purposes, such as INN prescribing. Figure 12. Value Set of EDQM ISI characteristic Note: the values "cutaneous/transdermal” and “intravesical/urethral” were split to the separate values - resp. cutaneous and transdermal, and intravesical and urethral - following discussions initiated in UNICOM. Also note the differentiation between oral, buccal, oromucosal, and sublingual. Finally, we identified a need to create an intermediate level of aggregation for the EDQM dose forms, for the use cases of INN prescribing and the alignment with RxNorm Dose forms. Value Set of EDQM Intended Site Characteristic Auricular Nasal Buccal Ocular Cutaneous Oculonasal Dental Oral Endocervical Oromucosal Environmental Parenteral Extracorporeal Pulmonary Gastric Rectal Gastroenteral Sublingual Intestinal Transdermal Intramammary Unknown/Miscellaneous Intraperitoneal Urethral Intrauterine Vaginal Intravesical D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 34 of 62 Figure 13. Dose form aggregation Hence, a draft proposal for such a dose form ontology was elaborated within WP8, based on a thorough analysis of the unique combinations of the characteristics of the dose forms, with their formal definitions (more details in the next chapter). The result of this work was published as a simple ontology of dose forms in WebProtégé.32 This ontology was used for an alignment of EDQM dose forms with RxNorm.33 A final validation round with experts within UNICOM is currently under way. Normalisation of strength Although the expression of strength may seem to be the least problematic one for constructing a global PhPID, there are several tricky issues which need to be dealt with in IDMP. National medicinal Products have a strength printed on the outer package. That is the strength expression as authorised by the National Regulatory Authority. For most medicinal products this reference strength is the strength of the moiety (not considering the molecular weight of the salt/ester (modifier)). However, this is not always the case within one specific country and this may differ also between countries. Medical departments of pharmaceutical companies may also have different strategies in this regard, now or in the past. Hence, IDMP requires to determine the strength of the moiety and the strength of the moiety+modifier. To know the second strength when one strength is known, requires the knowledge of the molecular weight of both instances. One also needs to know to which of the two instances of the molecule the authorized strength refers. And that is often also part of the “pharmaco-archaeology“ process. 32 Vander Stichele, R.H.; Roumier, J.; van Nimwegen, D. How Granular Can a Dose Form Be Described? Considering EDQM Standard Terms for a Global Terminology. Appl. Sci. 2022, 12, 4337. https://doi.org/10.3390/ app12094337 33 Karapetian N, Vander Stichele R, Quintana Y. Alignment of two standard terminologies for dosage form: RxNorm from the National Library of Medicine for the United States and EDQM from the European Directorate for the Quality in Medicines and Healthcare for Europe. Int J Med Inform. 2022 Sep;165:104826. doi: 10.1016/j.ijmedinf.2022.104826. Value Set Intented Site Characteris�c Of EDQM Dose Forms (N=28) Value Set EDQM Pharmaceu�cal Dose Forms PDF/CDF/CMT (N=443) Administrable dose forms (N=305) Proposed ontology for Intermediate Level of Aggrega�on for EDQM Dose Forms (N=65) D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 35 of 62 Moreover, for each product a choice must be made between a presentation strength and a concentration strength. For products with presentation strength, only the value and the unit of the numerator need to be known for a specific unit of presentation (that needs to be determined in EDQM terminology). The pack size will in that case refer to the number of units of presentation in the medicinal product pack. For products with a concentration strength, a numerator value and unit need to be determined and a denominator value and unit The denominator value will be default 1 and not mentioned (e.g. 5mg/ml). Here the numerical pack size will be the total volume or weight of the product e.g. 80ml syrup or 30g dermal gel). A 3% dermal gel can be expressed as 3 gr per 100ml or 30mg/ml. It is not acceptable to represent these different strengths in different PhPIDs as they are essentially the same. There is thus a strict need to normalize different expressions of the same strength in a global way. This is not an easy task and has been the focus of long and intense technical debate on how to express the strength in function of different patterns of dose forms. This debate has been held in the GIDWG (Global IDMP Working Group), with WHO UMC, EU, FDA, with an intense contribution of the Norwegian Agency NOMA, and other stakeholders. Several versions of what is called „business rules for strength expression“ have been edited, and circulated internally, but no official draft has yet been published. Issues with multi-dose and uni-dose vials, issues with inhalers, dermal products remain to be solved. It is possible but not guaranteed that a first draft will be made public by the end of the UNICOM project in May 2024. Controlling the expression of strength will be one important aspect of the governance role of WHO UMC when establishing the Repository of PhPIDs. Linking the PhPID to ATC Level V through the Virtual Medicinal Product Group With these components (specified substance, granular administrable dose form and 2 ontologies) a clinical meaningful stepping stone can be established to link the global representation of national medicinal products (PhPID) to the ATC Classification Level V. Two ontologies can govern a higher level of aggregation: an ontology of substance can aggregate the substance with the role of PAI to the grouper of substances with the same moiety (to be labelled with the INN terminology). The EDQM administrable dose form can be aggregated to the value set of the Intended Site Characteristic. The strength expression can remain unchanged in this effort. These 3 aggregated elements will together establish the concept of the Virtual Medicinal Product Group (VMPGroup). And this concept can be linked to the ATC classification Level V, almost always on a one to one basis. We suggest this to be a international academic collaborative effort, in close cooperation with the WHO CC for Drug Statistics in Oslo (global governor of the ATC Classification), and with the WHO CC for Pharmacovigilance (Uppsala Monitoring Centre UMC) (global governor of the PhPID Repository). D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 36 of 62 Figure 14. Link from PhPID to ATC 3.2.3 From ATC Level V to Clinical Classifications When Medicinal Products are linked to the ATC level V they are automatically linked to 4 higher hierarchical classes of this Classification, widely used for Drug Utilisation research. Figure 15. The 5 connected levels of the ATC Classification In the previous chapter we discussed the link between several international pharmacotherapeutic classifications and the ATC Level V code. 3.2.4 Closing the loop from medicinal product pack to pharmacotherapeutic drug classes Medicinal product packs in almost all countries bear bear a distinctive national identification code known as the National Drug Code (NDC). This code is unique to each package available in the market and serves to identify authorized packages and distribution units within the supply chain. Functioning as a primary key in the medicinal product dictionary, this code is linked to the medicinal product but may vary depending on the pack size, such as small, larger, or the largest package available. In the United States, the NDC code may vary for each medicinal product, particularly if there are multiple manufacturing options available, for instance, white round pills from an Indian generic provider or oval grey pills from a Chinese provider, both containing the same substance and strength. The regulations governing the life cycle of such a code, such as updating when there is a change of ingredients, may differ in each country. Link Table from National Identifiers of Medicinal products Link Table from ATC and International Classifications Multilingual Management of labels Conceptual Model of the Link between aGlobal Repository of PHPIDs and ATC VIrtual Medicinal Product Group VIrtual Therapeutic Moeity Group v ISI Dose Form INN substance Granular substance v Granular Administrable dose from Strength Virtual medicinal Product, identified by the PhPID Strength Substance Ontology Dose Form Ontology D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 37 of 62 For clinicians, distinguishing between a pack of 5 blisters with 6 tablets or a package of 6 blisters with 5 tablets may not be relevant, nor the material of the blister, whether it be aluminum or plastic. What matters is the numerical pack size, indicating the quantity of tablets in a package. Moreover, information on pack type is not consistently available from all countries and may only be presented as a text string. IDMP has a whole set of rules for creating and maintaining PCIDs (Pack control identifiers), which could potentially replace or complement National Drug Codes. However, only a few countries have made significant progress in implementing IDMP to generate PCIDs for all their medicinal product packs. In this project, we introduced the concept of “medicinal product pack“, defined by the country ISO-3 country code, the marketing authorisation holder, the specified substance, the granular EDQM administrable dose form, the strength (presentation of concentration strength), and, finally, the numerical pack size. This concept facilitates the comparison of the occurrences of this concept across different countries. Furthermore, we devised the concept of "Medicinal Product," which encompasses the aforementioned elements excluding the numerical packsize. This enables a comparison of the variety of different medicinal products across various countries. It was imperative to rectify company names in the US due to subtle spelling variations that were initially entered during the NDC application process and subsequently propagated in the RxNorm system. Subsequently, we formulated the concept of "Pharmaceutical Product," defined by the specified substance, the granular EDQM code, and the strength, omitting the company and country, as it represents a global concept. Figure 16. Construction of concepts We finally introduced the concept of „Virtual Medicinal Product Group (VMPGroup), delineated by the moiety, the value of the Intended Site characteristic of the dose form, and the strength. For instance, the label for the example depicted in the figure is „amlodipine oral 10 mg,“ with the ATC-Code Level V identified as C08CA01. No numerical identifiers were utilized, as the string of values within the concepts uniquely defines the concept. In the subsequent sections, we will elucidate how this methodology was applied to analyze a subset of the therapeutic arsenal comprising medicinal products across various countries. The method of linking a medicinal product pack to the ATC level V code and subsequently to any pharmacotherapeutic class in multiple international drug classifications is best exemplified in the forthcoming figure. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 38 of 62 Figure 17. Double loop towards the ATC ISO/CEN IDMP standards govern the national Marketing authorisation identification from pack to product at the national level. In addition, it provides a global identifier for the medicinal products which remains neutral regarding their therapeutic categorization. The intermediate layer of the Virtual Medicinal Product Group provides further aggregation of substance and dose form based on substance and dose form but remains detached from therapeutic classification. The connection to the Anatomical Therapeutic Chemical (ATC) classification is crucial for situating products within a clinical context. This typically involves a one-to-one relationship, although there are exceptions, such as aspirin tablets 100mg, which can be used as both a preventive cardiovascular drug and a pediatric pain reliever, resulting in two distinct ATC codes.. In many countries, there exists a registry of medicinal product packs, which connects the pack identifier to the ATC level V code, its ATC route of administration, the Defined Daily Dose (DDD), and calculates the number of DDDs per package (DPP). This registry serves as a critical component within administrative and scientific databases' ICT systems for conducting drug utilization research. Upon the completion of IDMP implementation, this will establish a dual linkage between medicinal product packs and ATC level V, presenting numerous opportunities for enhancing the quality assurance of the registry. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 39 of 62 4 Attempts to produce complete samples of medicinal products, standardised to IDMP, for a limited number of substances. Several tasks outlined in Work Packages 5 to 8 were designed to operate with complete data sets on IDMP standardised sets of medicinal products It quickly became evident that this would pose a significant challenge for many National Competent Authorities (NCAs) involved in the project, particularly given the emergence of the COVID-19 pandemic, which diverted the agencies' attention elsewhere. Hence, the leadership of UNICOM, and the work packages 1,6,7,8 and 9 have proposed to focus on a pilot product list of a limited number of substances. 4.1 The UNICOM Pilot Product List 4.1.1 Rationale  The objective of this endavour was to choose a set of substances to focus on for generating IDMP-compliant descriptions of medicinal products across various countries.  Ideally, this selection would encompass a range of substances that present challenges in medicinal product identification information and cross-border prescribing. 4.1.2 Methodology The following criteria were selected: • Frequently used substances • Substances that exemplify challenges in product identification. • Substances from the Connecting Europe Facility (CEF) eHealth Digital Service Infrastructure (eHDSI) Critical Test Data (the list of medicinal products for the cross border pilots) • Focus on chemical substances, and a limited number of combinations (amoxiclav and an anti-conceptive combination (drospirone+ethinylestradiol)) For all selected substances, a working group of experts from WP9, WP8 and WP2 determined all available modifiers for each substance-if anyand collected the codes for active moiety and for moiety+modifier-if anyin the following coding systems: • EU-SRS • WHODrug • UNII • CAS • SNOMED-CT This list was also shared with the WHO CC Centre of OSLO, with the request to add available ACTcodes (sometimes more than one per substance). Based on this list of ATC codes, several countries were asked to provide a list of available medicinal products in their country. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 40 of 62 4.1.3 Results A list of 35 substances was established (37 when combinations are split) Figure 18. UNICOM Pilot Product List The coding of these substances and their modifiers -if anywas completed by a group of UNICOM experts,34 resulting in a list of these substances and their codes in EU-SRS, WHODrug, UNII, CAS, and SNOMED-CT (see Annex 4) The list of relevant ATC-codes for the UNICOM Pilot Product List was provided by the WHO CC for Drugs Statistics Methodology (DMS) (see Annex 5).35 The list of medicinal products corresponding to these ATC codes was compiled for Belgium, Greece, Italy, Norway, and Finland. For the pilot with Patient Facing App (T8.3), full collections of medicinal products for a least some substances in Greece, Italy and the United States of America were needed. Four substances were selected from the minimal dataset of 35 substances, namely amlodipine, carbamazepine, ibuprofen, simvastatin. This selection was based on frequency, and suitability for demonstrating decision support applications. 4.2 UFIS and the UNICOM T6.1 FHIR SERVER From that point the work around this Pilot Product List split up. WP9 focussed on describing for each substance one or a few medicinal product completely in IDMP, and to store that information in the UFIS database (a FHIR server, following the guidance of the Biomedical Research and Regulation Working Group36). 34 In collaboration with Ursula Tschorn, Julie James, Annet Rozema, Leonora Grandia, Robert Vander Stichele 35 Courtesy of Mohammad Nouri Sharikabad, director of WHO CC DSM in Oslo, Norway 36 https://www.hl7.org/Special/committees/rcrim/index.cfm The Unicom Pilot Product List List of substances • metformin • amlodipine • perindopril • tramadol • ciclosporine • itraconazole • goserelin • clotrimazole • varenicline • ibuprofen • tafluprost • calcium carbonate • ergocalciferol • paracetamol • diazepam • morphine • enoxaparin • hydrocortisone • lidocaine • trastuzumab • chloroquine • clomipramine • carbamazepine • simvastatin • enalapril • omeprazole • diclofenac • cefuroxime • salbutamol • amoxicillin • clavulanate • insulin glargine • teriparatide • drospirenone • ethinylestradiol • glyceryl trinitrate D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 41 of 62 Another group established the UNICOM FHIR Server D6.1, following the guidance of the HL7 FHIR Pharmacy Working Group).37 Ultimately, some 300 Medicinal Products, mainly from Sweden and Estonia, were transferred to UFIS. When this database was terminated by the NCAs, the data were transferred to the UNICOM T6.1 FHIR Server. 4.3 The minimal data set on 4 substances, based on “data as is“, created in Work Package 8 After two years within UNICOM, it became evident that the WHO Collaborating Centre for Pharmacovigilance, Upsala Monitoring Centre (UMC), would undertake the responsibility of maintaining a repository of PhPIDs once consensus was reached regarding the principles of applying the ISO/CEN IDMP standards for this purpose. Only a few National Competent Authorities, namely those of Sweden, Estonia, and Portugal, had generated IDMP-compliant data for approximately 300 medicinal products, selected to exemplify substances listed in the UNICOM Pilot Project List. Examples included combination products such as amoxiclav, and also complex products with more than one manufactured item (e.g. the three-phasic contraceptive oestrogen/progestogen pill), products for Helicobacter treatment with several manufactured items, parenterals with medical devices. A FHIR server was built to harbour those IDMP-compliant data, and they were later transferred to the UNICOM FHIR SERVER (database T6.1). For this approach, the NCA participating in the provision of data needed to comply to IDMP-internally, and to provide the data in FHIR format, according to the EU IDMP Implementation Guide. While this approach was very useful for the participating NCAs to test and demonstrate their internal IDMP-compliant systems, the collection of medicinal products did not serve the needs of the pilots and tasks of Work Package 8, nor the needs for testing the substitution component, as envisioned in T6.2. For that, it was imperative to have complete samples of all available medicinal product packs for each active substance. Moreover, complete samples of medicinal product packs for one substance, needed to be also available in at least one, and in preferably more countries. This was necessary to facilitate testing of a substitution module in cross-border services and to conduct experiments involving patients traveling to foreign countries with a patient-facing app containing their personal list of national medications. 4.3.1 Motivation for the initial decision to create a minimal data set. By the third year of UNICOM, it became clear that such samples would not be provided by the participating National Competent Authorities (NCAs); it was then decided to start a parallel data collection of “data AS IS”, meaning that a request would be made to a national data provider to provide complete samples of medicinal product packs for a limited, predefined list of substances. It was not necessary for the data to have undergone prior attempts at standardization to IDMP. The local descriptions of the products would suffice. 4.3.2 Constraints of this work Several constraints were agreed to limit the amount of work to a feasible extent: • First, medicinal products for only 4 substances (amlodipine, carabamazepine, ibuprofen, simvastatin were requested. • The request was solely for the provision of national identifiers and descriptors, which were required to encompass information on active substances (with the role of Precise Active Ingredient), dose form and strength (as stated on the national package), the marketing authorization holder, pack size, national drug code for the pack, and, optionally, pack description 37 https://www.hl7.org/Special/committees/medication/index.cfm D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 48 of 62 5.3 IDMP and Guidelines Guidelines play an increasingly pivotal role in shaping the daily practices of healthcare providers. Typically, these guidelines are tailored to address specific diseases for particular healthcare professions. However, they can also be designed to address multimorbidity with a multidisciplinary approach. Guidelines can be translated in specific practice recommendations, gathered in evidence-based pointof-care information summaries40.41 These guidelines contain a lot of references to pharmacotherapy. Examples of such point of care summaries with international penetration is EBM-Guidelines®42, from Finland, distributed in several European Countries in several languages.43 Linking such resources to the national Medicinal Product Dictionaries will be greatly facilitated by the implementation of IMDP in the 27 Member States of the European Union. Also, the internationalisation of MPDs such as BNF in the UK, ZN-Index in the Netherlands and Vidal in France may be fostered by IDMP. Similar argumentation can be developed for producers of care pathways. Tools to evaluate the quality of these point-of-care resources have been developed.44 Lately, FHIR resources have been created to support Evidence-based Medicine, with the aim to represent evidence from clinical trials (the cornerstone for the assessment of efficacy of medicines, and for real world data, instrumental for pharmacovigilance and pharmacoepidemiology).45 An important consideration here is the specificity, which can vary across different levels: the level of individual medicinal products, the level of pharmaceutical products (comprising all medicinal products with the same substance, dose form, and strength), and the level of pharmacotherapeutic class. For 41 Banzi R, Liberati A, Moschetti I, Tagliabue L, Moja L. A review of online evidence-based practice point-of-care information summary providers. J Med Internet Res. 2010 Jul 7;12(3):e26. 42 https://www.ebm-guidelines.com/apps/dtk/ebmg 43 https://www.duodecim.fi/english 44 Lenaerts G, Bekkering GE, Goossens M, De Coninck L, Delvaux N, Cordyn S, Adriaenssens J, Aertgeerts B, Vankrunkelsven P. A Tool to Assess the Trustworthiness of Evidence-Based Point-of-Care Information for Health Care Professionals (CAPOCI): Design and Validation Study. J Med Internet Res. 2021 Oct 5;23(10):e27174 45 Vorisek CN, Lehne M, Klopfenstein SAI, Mayer PJ, Bartschke A, Haese T, Thun S. Fast Healthcare Interoperability Resources (FHIR) for Interoperability in Health Research: Systematic Review JMIR Med Inform 2022;10(7):e35724 D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 49 of 62 clarity and conciseness, it's crucial to manage and differentiate these distinct levels effectively. Figure 20. Levels of specificity in drug information 5.4 IDMP and Computerized Decision Support (CDSS) Computerized Decision Support has been introduced decades ago, from the beginning of the introduction of Electronic Health Records.46 It was often hailed as a motive for using computers in medical documentation of clinical practice. However, user satisfactions and acceptance could not be taken for granted, as alert fatigue was a phenomenon that quickly kicked in.47,48 All over the world many CDSS have been developed for numerous subjects: • Pharmaco-kinetic interactions • Pharmaco-dynamic interactions • Indications • Contra-indications • Lactation and pregnancy • Ability to drive and work machines, • Allergies • Posology support • Detection of cascade therapy (prescriptions for alleviating side-effects of medicinal products) For pharmaco-kinetic interactions, systems abound, also directed at patients, and accessible through the, internet. The consistency of these systems is however limited, and there is little consensus on what are clinically relevant alerts. 46 Beeler PE, Bates DW, Hug BL. Clinical In decision support systems. Swiss Med Wkly. 2014 Dec 23;144:w14073. 47 McCoy AB, Thomas EJ, Krousel-Wood M, Sittig DF. Clinical decision support alert appropriateness: a review and proposal for improvement. Ochsner J. 2014 Summer;14(2):195-202. 48 Carli D, Fahrni G, Bonnabry P, Lovis C. Quality of Decision Support in Computerized Provider Order Entry: Systematic Literature Review. JMIR Med Inform. 2018 Jan 24;6(1):e3. Specificity of drug information Information specific for a particular Medicinal Product Information specific for a group of medicinal products with the same substance(s) Information specific for the relevant pharmacotherapeutic class D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 50 of 62 Pharmaco-dynamic interactions are complex to assess, especially in older adults and residents in nursing homes with long lists of medication. Their frequency and impact on quality of life could be more important. For example, the determination of anti-cholinergic load to which a poly-medicated patient is exposed is a difficult and time-consuming work. There are many anticholinergic scales, some of them integrating this information in a review of scales49, and some transformed in an automated assessment tool.50,51 The Duran List of anticholinergics is a global list of 100 weak and strong anticholinergics, provided with the corresponding ATC codes (see annex 8). Full implementation of IDMP would make the complex connection to all the relevant medicinal products easier, for one country, and certainly also for multiple countries. It would also allow to consider the dose form and strength more easily. Decision support has also been developed to combine computerized assessment of appropriateness of prescribing with interprofessional evaluation in the context of Long-Term Care Facilities (Nursing homes). Electronic platforms allow nurses, pharmacists, and physicians to cooperate and evaluate automatically generated suggestions for correcting inappropriate prescribing and for deprescribing.52 Within UNICOM, in Task T8.5, for the first time the impact of pharmaco-genetics on CDSS has been explored (see D8.5 and D8.10). One example of complex pharmacodynamic interaction was elaborated in the pilot of the Patient Facing Apps around ibuprofen, to illustrate automatically generated alerts for nephrotoxicity, induced by the concomitant use of Non-Steroidal Anti-inflammatory Agents, ACE-inhibitors and diuretics. 49 Durán CE, Azermai M, Vander Stichele RH. Systematic review of anticholinergic risk scales in older adults. Eur J Clin Pharmacol. 2013 Jul;69(7):1485-96. 50 Wauters M, Klamer T, Elseviers M, Vaes B, Dalleur O, Degryse J, Durán C, Christiaens T, Azermai M, Vander Stichele R. Anticholinergic Exposure in a Cohort of Adults Aged 80 years and Over: Associations of the MARANTE Scale with Mortality and Hospitalization. Basic Clin Pharmacol Toxicol. 2017 Jun;120(6):591-600. 51 Wehran T, Eidam A, Czock D, Kopitz J, Plaschke K, Mattern M, Haefeli WE, Bauer JM, Seidling HM. Development and Pilot Testing of an Algorithm-Based Approach to Anticholinergic Deprescribing in Older Patients. Drugs Aging. 2024 Feb;41(2):153-164. 52 Wauters M, Elseviers M, Vander Stichele R, Dilles T, Thienpont G, Christiaens T. Efficacy, feasibility and acceptability of the OptiMEDs tool for multidisciplinary medication review in nursing homes. Arch Gerontol Geriatr. 2021 Jul-Aug;95:104391. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 51 of 62 Figure 21. An example of pharmacodynamic nephrotoxic interaction This illustrates the complexity to express decision rules in broad pharmacotherapeutic classes, to precise ATC code collections, and from there to PhPIDs and national Medicinal Products. In Europe the penetration and commercial success of such tools is hampered by the barriers of language (24 official languages), differences in national regulations (27 Member States), and by the almost impossible effort to connect to the sheer number of ever evolving Medicinal Product Dictionaries (on average 5 per country). Hence full implementation of IDMP in Europe will contribute to cross the border also for evidence-based drug information and decision support, and not only for ePrescriptions. Example of a rule in a decision support system for medication management Decision Rule in Human Language IF the patient is on an agent acting on the angiotensine system treatment AND on a diuretics, THEN He/She should not make use of NON-steroidal antiiflamatory drugs (especially when the kidney function is already compromised BECAUSE There is a serious risk of renal failure, leading to heart failure, hospitalisation, and possibly death Decision Rule operational in code IF the medication list has C09 AND CO3; C01AA ;C02L ;C07B ;C07D; C08G AND new prescription is requested : M01A; M01B; N02AJ02 (dihydrocodeine and acetylsalicylic acid); N02AJ07 (codeine and acetylsalicylic acid); N02AJ08 (codeine and ibuprofen); N02AJ09 (codeine and other non-opioid analgesics) ; N02AJ16 (tramadol and celecoxib); N02AJ18 (oxycodone and acetylsalicylic acid); N02AJ19 (oxycodone and ibuprofen) THEN LAUNCH ALERT 567 Do not add an NSAID to the medication list of this patient. He/She is in serious danger of renal failure, which might lead to hospitalization and possibly death. Consider using an non-inflammatory pain killer D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 52 of 62 5.5 IDMP and substitution rules and INN prescribing During the UNICOM project, information was collected on the national regulations for substitution and for prescribing by International Non-Proprietary Name. The initial plan, as outlined in Linked Activity (iii), was to evaluate the impact of these rules on the initial experiments for Cross-Border ePrescription and eDispensation. However, neither the eHSDI test environment nor the implementation of IDMP had progressed sufficiently to allow for the planning and execution of such tests. Nevertheless, it was possible to use the UNICOM Patient Facing App test environment (D8.5) to simulate testing of semantic interoperability of cross border transfer of medication lists. The 3 patient facing apps involved were: • HealthPass from Gnomon in Greece: • PharmaWizard from DataWizard in Italy • InfoSage from Beth Israel Deaconess Medical Center in the USA. The medication lists used for the test would have to be limited to the 4 substances, for which all Medicinal Products Packs were standardised to IDMP in the Minimal Data (instrumental to tasks T8.1, T8.2, and T8.3) For that purpose, 236, 220, and 5167 medicinal product packs from Greece, Italy, and the USA were centrally standardised to IDMP by a small team (a clinical pharmacologist, a terminological expert, and a drug database expert). The standardisation of the medicinal products was implemented for the variables of the Minimum Attribute List, developed in UNICOM D5.7, listing the variables that are needed for cross-border ePrescriptions, in eHDSI (except for the variables describing Pack Type). This standardization of IDMP variables was supplemented with the construction of the following concatenated labels: • pharmaceuticalProductLabel: Specified Substance + granular EDQM dose form + normalised strength • virtualMedicinalProductGroupLabel: Substance + EDQM characteristic Intended Side + normalised strength. To provide a global label to each national medicinal product the following labels were constructed: • medicinalProductLabel: Country + Specified Substance + Company + granular EDQM dose form + normalised strength • medicinalProductPackLabel: Country + Specified Substance + Company + granular EDQM dose form + normalised strength As the EDQM Code for granular Administrable dose form was present in the Minimal Data Set, all other characteristics of the dose form (including the Intended Site characteristic) could be added automatically. This not only allowed to test different algorithms to create the PhPID (see figure 5), but also made the testing of cross border substitution possible in the patient facing apps, a private environment, outside the realm of the eHDSI infrastructure. In addition, a substitution module was developed in T6.2 to be demonstrated in Patient Facing Apps in 3 countries (see UNICOM D8.5). Based on the results of the survey of Substitution rules and INN Prescribing, the following 3 levels of substitution were identified: • Brand substitution: the medicinal product of the ePrescription in sending Country A can only be delivered as an eDispensation by the pharmacist (without confirmation by a physician), in case an identical medicinal product from the same company is marketed in country B D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 53 of 62 • Pharmaceutical Product substitution: the medicinal product of the ePrescription in sending Country A can be delivered as an eDispensation by the pharmacist (without confirmation by a physician), by picking from a list of medicinal products with the same Pharmaceutical Product (specified substance, granular EDQM administrable dose, normalised Strength) • INN prescribing: the medicinal product of the ePrescription in sending Country A can be delivered as an eDispensation by the pharmacist (without confirmation by a physician), by picking a suitable candidate from a list with medicinal products (marketed in the country) within the same Virtual Medicinal Product Group (same substance, same EDQM Intended dose form, same normalised strength) in country B. Based on the results of a survey among UNICOM NCAs on Substitution rules and INN prescribing it beame possible to determine what the possibilities were in the different countries for choosing a replacement for an ePrescription of country A for an eDispensation in country B. Figure 22. Diversity in rules for substition and INN Prescribing in UNICOM EU member States To illustrate this approach, we can provide an example. Let’s imagine an Italian Patient with a prescription of the Italian Medical Product NORVASC*14 cpr 10 mg (with the national Drug Code 27428022) traveling to Greece, and requesting an eDispensation in Greece. The global label of this medicinal product is: ITA amlodipine besylate PFIZER ITALIA Srl tablet 14 x 10mg/ This is an IDMP standardized label, informing the Greek healthcare provider that this is an Italian medicinal product, containing amlodipine besylate as active substance, licenced by Pfizer, with “tablet” as dose form, in a strength of 10 mg and containing 14 tablets. Each of these elements is coded, and stored separately so that the selection processes can be applied. First option In this example the first option for brand substitution is not available as Pfizer no longer markets amlodipine in Greece. Brand Equivalent same company, same specifiedsusbstance, same granular administrabledose form, same strength Pharmaceu�calProduct Equivalents same specifiedsubstance, same granular administrabledose form, same strength Interna�onalNonProprietaryPrescribing Equivalents same ac�ve moiety, same high leveldose form group (e.g. oral) same strength Note :Posology instructions can induce more choice in strength , by splitting (if dose form permits)or by double dosing Austria Kroa�a Sweden Ireland GE,BE, FI,NL, NO,EE, GR,ITA, ES,PT Fig 1. Differencesin limita�ons for finding equivalentsfor crossborderePrescrip�onsin EU countries DE,BE, FI,NL, NO,EE, GR,IT, ES,PT D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 54 of 62 Second option The second option is to go for exact substitution based on the Pharmaceutical Product: amlodipine besylate tablet 10mg/ The substitution module will then generate a list of medicinal product packs: Third option: INN substitution The third option is based on the elements constituting the Virtual Medicinal Product group : This leads in this case to a much more extended list and more choice for the eDispensation. amlodipine oral 10mg/ This approach was implemented for each of the 4 substances and for the 3 countries implicated in Task T8.3, and demonstrated at several events, reported in D8.5. with a pilot in 25 traveling residents, and an evaluation of the concordance of the (e)Prescription and the (e)Dispensation, on private smartphones (and not in the cross-border eHSDI infrastructure) D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 55 of 62 Hence, we were able to perform the tasks formulated in Linked activity 3: Testing semantic interoperability of cross border transfer of( medication lists D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 56 of 62 5.6 IDMP and patient Facing Apps As described above, the activities around the drug class grouping and decision support were applied in the development of the three Patient Facing Applications in UNICOM • HealthPass (Greece) • PharmaWizard (Italy) • InfoSage (USA) During the demonstrations, cross linking of drug groups to decision support were tested in the interactions between Italian – Greek residents, Italian - American residents, and Greek – American residents. For the decision support, the knowledge database of RxNorm was mainly used. To identify the Italian Products, IDMP was used, and to identify the American products outside the Minimal Data Set, the ATC Code was used. The following examples were used: • An American resident traveling to Italy dispensed an Italian brand of ibuprofen as a pain killer for an ankle strain. On the medication list of this patient an American diuretic is present, and also an ACE-inhibitor. When integrating the prescription of the Italian product in the medication list, the app generates a safety request to RxNORM, and gets in return an alert for possible renal failure form by concomitant use of 3 nephrotoxic substances. (pharmacodynamic interaction) • An American resident traveling to Greece, with simvastatin (a statin) on the medication list, gets a prescription in Greece for amlodipine (a calcium anta)agonist for hypertension). Safety request to RxNorm issues a warning of rhabdomyolysis (degradation of muscle) due to a pharmacokinetic interaction, leading to steep increase of the concentration of simvastatin in the blood, and to toxicity. (pharmacokinetic interaction) • An American resident traveling to Italy, a known poor metabolizer for CYP2C9, receives an OTC prescription for a painful tooth. He receives a warning from RxNorm that it would be better to refrain from taking ibuprofen, or at least half the dose, or take an alternative paracetamol. The risk of toxicity of ibuprofen it too high for this person. (risk of toxicity because of pharmacogenetic variant) Hence, we were able to perform tests using the resources developed for linked activity 4: cross linking drug groups and for linked activity 3: testing semantic interoperability of cross border transfer of medication lists. D8.4: Application of IDMP in Drug Labelling and Drug Information, in Clinical Decision Support and in QA of Clinic Data Page 57 of 62 5.7 IDMP and the quality of clinical data 5.7.1 Using the medication list to improve the quality of medical documentation Drawing from experience in assessing polypharmacy within nursing homes, it became evident that highly reliable data on medication usage can be obtained from the eMAR systems (electronic Medication Management) utilized by nurses to prepare medication administration rounds. However, in many countries, the quality of medical records in this sector may present challenges. For example, there may be issues such as the absence of a current and comprehensive list of medical problems or indications. This is particularly a problem for the automatic application of the more complex explicit criteria for (in)appropriate prescribing, as in the international STOPP/START list and the Beers‘ List. These complex criteria require the presence of interoperable essential clinical data. A proposal was made to use these criteria, designed for assessment of the quality of prescribing, for the assessment of the completeness of the medical record. We therefore studied the criteria involving medication (expressed in ATC) and clinical diagnoses. A list was devised of medications that necessitated clinical data criteria for evaluating their appropriateness. For each medication on the list, the specific nature of the required clinical data was documented. A program was developed to analyze individual medication lists for the inclusion of such medications. Following an assessment of the presence or absence of specific medical data, questions could be prompted regarding the presence or absence of missing information on particular clinical conditions. Physicians could respond with a simple "yes" or "no" to these questions (anticipated to be limited in number per patient), after which the problem list would be automatically updated in a structured and coded format. Figure 23. Repairing the quality of medical record through the medication list This work was developed during the UNICOM project, as a master thesis at Ghent University,53 resulting in a list of triggering medications with the corresponding inquiry for clinical data (see Annex 9). • 535353 Bos M. Using Explicit Criteria for potentlially (in)appropriate prescribing to improve medical documentation. [Master Thesis] Ghent University 2023 Medication Chart of an individual patient Clinical data from an individual patient Clinical Rule data base List of clinical elements needed to run the relevant clinical rules List of clinical elements missing to run the relevant clinical rules Question list to patient and/or treating physician EHR