D8.1 Report on the link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data in Pharmacotherapeutic Audit (M22)

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 D8.1: Report on the link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data in Pharmacotherapeutic Audit (M22) Version: 1.0 Status: Final Dissemination Level1: PU Due date of deliverable: 30.09.2021 Actual submission date: 30.09.2021 Work Package: WP8: IDMP for Clinical Care, Patients, Pharmacies, Research and Pharmacovigilance Lead partner for this deliverable: I-HD Partner(s) contributing: DWIZ, FOUND, SNOMED, CEB/CBG, HAR Deliverable type2: R Main author(s): Vander Stichele Robert I-HD Other author(s): Dipak Kalra Lucia Comnes I-HD Datawizard Anja Van Haren Yuri Quintana CEB/CBG Harvard Maurizio Taglialatela Found Joseph Roumier I-HD Jane Millar Snomed International Dirk Van Nimwegen I-HD 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 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 2 of 61 Revision history Version Date Changes made Author(s) 0.1 05.07.2021 Table of content Robert Vander Stichele I-HD 0.2 01/O9/21 Revised table of content Robert Vander Stichele I-HD 0.3 O5/09/21 First 5 chapters Robert Vander Stichele I-HD 0.4 09/09/21 First inner circle draft Robert Vander Stichele, I-HD 0.5 14/09/2021 Draft for internal review Robert Vander Stichele,Dipak Kalra, Josepth Roumier, Dirk Vannimwegen; I-HD 0.6 16/09/21 Internally reviewed version Luc Nicolas Final 29/09/2021 Final version Robert Vander Stichele Statement of originality This deliverable contains original unpublished work except where clearly indicated otherwise.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. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 3 of 61 Deliverable abstract In this deliverable, crucial steps are reported in linking precise global identification of individual medicinal products from different countries (using the IDMP standard) to international drug classifications, such as the World Health Organisation Anatomical Chemical Therapeutic Classification (ATC), SNOMED-CT, RxNorm, and other simplified educational classifications. The clinical value of such linking requires to be demonstrated. Ir will be illustrated by showing the usefulness of a link between the global identification number (the Pharmaceutical Product Identifier or PHPID) and the ATC classification. This link will then facilitate the application of internationally validated rules for decision support in pharmacotherapeutic audit. The report provides a detailed analysis of the coding systems for substances, dose form (EDQM and RxNorm), and the possible business rules for representing strength according to the granularity of substances, and the patterns of dose form. This work was set up in preparation of establishing a procedure for the production of the Pharmaceutical Product Identifier (PHPID), with codificati on rules for substance, dose form and strength, fed into a Hash function, to produce a global unique identification number for a group of identical medicinal products from different countries, sharing the same 3 items. A public repository of such PhPIDs would be instrumental for applications in clinical care, research, patient information, pharmacovigilance, and precision medicine. Keywords: Substance, Dose Form, Strength, Pharmaceutical Product Identifier, International Drug Classifications, Pharmacotherapeutic Audit, (In)Appropriate Prescribing 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. The participants of the UNICOM project. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 4 of 61 TABLE OF CONTENTS Deliverable abstract ................................................................................................................................. 3 List of abbreviations ................................................................................................................................. 7 List of figures ........................................................................................................................................... 9 List of tables .......................................................................................................................................... 10 List of annexes ...................................................................................................................................... 11 1 Executive summary ........................................................................................................................ 12 2 Rationale for the use of the ISO/CEN IDMP Standards in clinical care ......................................... 13 2.1 Introduction to the use of IDMP in clinical care ..................................................................... 13 2.2 The need to precisely identify medicinal products within one country .................................. 14 2.3 The need to precisely identify similar medicinal products from different jurisdictions ........... 14 2.3.1 Cross-border prescribing and dispensing .......................................................................... 14 2.3.2 Evidence-based Medicine crossing the borders ................................................................ 14 3 Building a solid foundation for international grouping: The Pharmaceutical Product Identifier (PHPID) ................................................................................................................................................. 16 3.1 Rationale ................................................................................................................................ 16 3.2 The 3 crucial elements of the Pharmaceutical Product Identifier .......................................... 16 3.3 Two major applications relevant for clinical care ................................................................... 16 3.4 Focus on PHPID in Work Package 8 of the UNICOM project ............................................... 17 3.5 Responsibility for PhPID production ...................................................................................... 17 4 Moving to pharmacotherapeutic groups, by gradually building groups of higher level of abstraction. 18 4.1 Starting with the level of PhPID ............................................................................................. 18 4.1.1 Identifying the triplets for amlodipine as an example ........................................................ 18 4.1.2 Looking up the available coding systems .......................................................................... 20 4.1.3 Looking up available codes for modified substances ........................................................ 20 4.1.4 Looking up avaible codes for dose forms .......................................................................... 20 4.1.5 Looking up available codes for strength Units ................................................................... 21 4.1.6 Completion of collection of codes in preparation of PHPID production............................. 21 4.2 Building up to the level of Virtual Medicinal Product Group (VMPGroup) ............................. 21 4.3 Building up to the level of Virtual Therapeutic Moiety Level (VTM) ....................................... 23 4.4 Dealing with combinations ..................................................................................................... 24 4.5 Bringing it all together ............................................................................................................ 26 4.6 Moving to pharmacotherapeutic groups ................................................................................ 28 5 Starting to get real. The Unicom Pilot Product List ........................................................................ 29 5.1 Rationale ................................................................................................................................ 29 5.2 Upstart of the PPL project ..................................................................................................... 29 5.3 Selection criteria for the PPL list............................................................................................ 29 5.4 Result of the selection: 35 substances .................................................................................. 30 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 5 of 61 5.5 Next step: Cleansing of the list .............................................................................................. 30 5.6 Gathering medicinal products from 4 countries (NL, BE, UK, GE) ........................................ 31 5.7 Preparing further work with the UNICOM Pilot Product List .................................................. 31 6 Analysis of the 3 basic elements of IDMP (substance, dose form, strength) ................................. 32 6.1 The hierarchy of substances ................................................................................................. 32 6.1.1 Physical reality, abstraction, and levels of hierarchy in the representation of substance . 32 6.1.2 Levels of hierarchy of substances ..................................................................................... 35 6.1.3 The case of moieties without a modifier ............................................................................ 36 6.2 Dose forms and their characteristics ..................................................................................... 37 6.2.1 Existing coding systems for dose forms ............................................................................ 37 6.2.2 The value of EDQM ........................................................................................................... 37 6.2.3 Possible improvements for EDQM in the transition to a global terminology ..................... 38 6.2.4 Possibility of creating a dose form ontology. ..................................................................... 39 6.2.5 Comparison to RxNorm and Snomed-CT dose forms....................................................... 39 6.2.6 Application to the UNICOM Pilot Product List ................................................................... 39 6.3 The complexity of determining strength of medicinal products ............................................. 40 6.3.1 Strength determination and substance .............................................................................. 40 6.3.2 Strength and dose form ..................................................................................................... 42 6.4 Application to the UNICOM Pilot Product List ....................................................................... 43 7 The procedure for production of PhPIDs ........................................................................................ 46 7.1 The need for global governance ............................................................................................ 46 7.2 What has already been done ................................................................................................. 46 7.3 Pathway to a consensus on the procedure for PhPID production ......................................... 47 8 Building a UNICOM Repository of validated Pharmaceutical Product Identifiers - PhPIDs .......... 48 8.1 Rationale ................................................................................................................................ 48 8.2 ICT-Operationalisation ........................................................................................................... 48 8.2.1 Conceptual model .............................................................................................................. 48 8.2.2 Logical model ..................................................................................................................... 49 8.2.3 Technical model................................................................................................................. 50 8.3 Fostering the impact of UNICOM on clinical care and research ........................................... 50 8.4 Prospects on sustainability of this UNICOM repository, beyond completion ........................ 51 9 Linking drug classifications to IDMP .............................................................................................. 52 9.1 A short and incomplete inventory of international drug classifications .................................. 52 9.2 The need for a simple classification of indications and products for Medical Education and Patient Information ............................................................................................................................ 53 10 Linking decision rules in pharmacotherapy to drug classifications and IDMP .......................... 54 10.1 Explicit rules for (in)appropriate prescribing from internationally validated PIM-lists. ........... 54 10.2 Sources of clinical data and medications to apply the criteria ............................................... 55 10.3 Coding the rules to disease and medication identifiers ......................................................... 55 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 6 of 61 10.4 Piloting IDMP decision support in UNICOM .......................................................................... 55 10.5 Using explicit criteria to test the quality of clinical data in the International Patient Summary (iPS) 56 11 Conclusions ............................................................................................................................... 57 12 References ................................................................................................................................ 59 13 List of Annexes .......................................................................................................................... 61 13.1 D8.1_Annex 1. List of Medicinal products from BE for the 35 substances from the UNICOM Pilot Product List ................................................................................................................................ 61 13.2 D8.1_Annex 2. Result of data cleansing of substances on the Unicom Pilot Product List ... 61 13.3 D8.1_Annex 3. Report of the WHO-UMC/FDA Pilot for PhPID production ........................... 61 13.4 D8.1_Annex 4. Analysis of EDQM terminology ..................................................................... 61 13.5 D8.1_Annex 5. Comparison EDQM / RxNorm ..................................................................... 61 13.6 D8.1_Annex 6. Specifications for the Draft Procedure for PhPID production ....................... 61 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 7 of 61 List of abbreviations Abbreviation Complete form ATC Anatomical Therapeutic Chemical Classification CAS Chemical Abstract Services CD Precisely Clinical Drug Precisely CSD Clinical Semantic Drug DDD Defined Daily Dose DID Defined Daily Doses per 1000 inhabitants per day Dm+d Dictionary of Medicines and Devices (UK) EDQM European Directorate for the Quality of Medicines EMA European Medicines Agency EU European Union EU-SRS European Substance Registration System (EMA) EUTCT European Union Telematics Controlled Terms FDA Food and Drug Administration GSRS Global Substance Registration System (FDA) ICD International Classification of Diseases ICPC International Classification of Primary Care IDMP Identification of Medicinal Products suite of ISO Standards INN International Non-Proprietary Name INNM International Non-proprietary Name Modified LOD Linked Open Data MeSH Medical Subject Headings MOH Ministry of Health MP Medicinal Product MP only Medicinal Product Only (SNOMED-CT) NDA New Drug Application NLM National Library of Medicine OHDSI “Observational Health Data Sciences and Informatics” consortium D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 8 of 61 OMOP-CDM Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) PHPID Pharmaceutical Product Identifier PIM-list Potentially Inappropriate Medication-list PREMs Patient Reported Experience Measures PROMs Patent Reported Outcome Measures SAM Source Authentique de Médicaments (BE) SDO Standard Developing Organization SNOMED-CT Systematized Nomenclature of Medicine -- Clinical Terms SPOR Substance/Product/Organisation/References services UMLS Unified Medical Language System UNII Unique Ingredient Identifier VMP Virtual Medicinal Product VMPGroup Virtual Medicinal Product Group VTM Virtual Therapeutic Moiety WHO_UMC World Health Organisation Uppsala Monitoring Centre D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 9 of 61 List of figures Figure 1. Three ways of dealing with aggregation of combination medicinal products. ........................ 25 Figure 2. Collation of concepts in representation of medicinal products in different systems .............. 27 Figure 3. The UNICOM Pilot Product List: 35 selected substances (incl. 4 combinations) ................. 30 Figure 4. Substance as a molecule and substance as a "scattered aggregate" ................................... 32 Figure 5. Molecular structures of amlodipine moiety and modifiers ...................................................... 34 Figure 6. Ambiguity in the terms "substance" and "modified substance" .............................................. 35 Figure 7. Carbamazepine, a molecule without a modifier ..................................................................... 36 Figure 8. Moiety and Precise Active Ingredient: ambiguity of concepts or of words ? .......................... 36 Figure 9. Determining strength based on business rules (pattern) dependent on type of dose form and nature of substance. ......................................................................................................... 42 Figure 10. Haiku on what binds and seperates almost similar things ................................................... 46 Figure 11. Conceptual model of a repository of PhPIDs ....................................................................... 48 Figure 12. Conceptual model of a repository of PhPIDs ....................................................................... 50 Figure 13. Connecting national drug dictionaries with international classifications .............................. 51 Figure 14. Selection process of criteria in a repository from 3 validated lists of explicit criteria of potentially inappropriate prescribing ................................................................................. 54 Figure 15. Audit of quality of clinical documentation with clinical rules for appropriate prescribing ...... 56 Figure 16. Pharmaceutical Product Identifier (PHPID-Level IV) as the connecting element between domains and use cases (courtesy of Christian Hay, UNICOM, WP1). ............................. 58 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 16 of 61 3 Building a solid foundation for international grouping: The Pharmaceutical Product Identifier (PHPID) 3.1 Rationale IDMP provides the means to accurately identify and describe each individual medicinal product in a given jurisdiction (e.g. a member state of the European Union, the USA, Japan). It is preferable that the elements to identify and describe medicinal products are standardized, which means that the list of elements used is the same in all jurisdictions, and that these elements (and their value sets) are expressed by terms from controlled multilingual vocabularies and reliable coding systems. The more accurate the description of individual medicinal products is made, the more reliable the international grouping of (almost (identical) medicinal products on the basis of one or more characteristics can be performed. The possibilities of groupings are endless, but some are more relevant than others. 3.2 The 3 crucial elements of the Pharmaceutical Product Identifier Within IDMP, a grouping on the basis of 3 crucial variables is established: 1. The Substance (the modified substance in case of chemicals) 2. The Dose Form (more specifically the administrable dose form) 3. The Strength (of the administrable dose form) With global, universally accepted codes for each of these 3 elements, a unique identifier can be created which combines these 3 elements, using a HASH function. In IDMP this unique identifier of a group of medicinal products, containing the same 3 elements is called the PhPID (Pharmaceutical Product identifier). There are 4 levels in this identifier: 1. PhPID_L1: Only substance 2. PhPID_L2: Substance + Strength (little practical value) 3. PhPID-L3: Substance + Dose Form 4. PhPID-L4: Substance + Dose Form + Strength Concepts similar to PHPID-L4 already exist in RxNorm (Semantic Clinical Drug), in Snomed-CT (Clinical Drug – CD Precisely), in the UK Dictionary of Medicines and Devices -Dm+d - (Virtual Medicinal Product -VMP), albeit the granularity of the value set of substance and of dose form, and the way of expressing strength might slightly differ between these information systems. 3.3 Two major applications relevant for clinical care With these 3 basic elements (substance, granular dose form, and referenced strength, two major applications can be realised: 1. a solid operationalisation of INN prescribing (prescribing by International Non-Proprietary Name) can be facilitated, 2. links can be established with • Classes for drug utilisation research in the taxonomy of the WHO Anatomical Chemical Therapeutic Classification, D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 17 of 61 • SNOMED-CT drug Classes, • WHODRUG Standardized Drug Classes, • RxNorm (and the common data model OMOP from OHDSI), • the table of content of pharmacological handbooks and pharmacopoeias • the drug classes defined in decision rules of decision support systems. 3.4 Focus on PHPID in Work Package 8 of the UNICOM project In this deliverable, the focus is on the grouper PhPID (the Pharmaceutical Product Identifier). This will provide the solid foundation for sensible, robust, and precise grouping of medicinal products across jurisdictions. It is based on uniform description of substance, dose form and strength in an abstraction of (almost) identical medicinal products from different jurisdictions, independent of the country and the marketing authorisation holder. This basic grouping can be the starting point for further clinical grouping of these atomic PhPID groups, guaranteeing correct information of the 3 basic constituents (substance, dose form, strength), needed for a correct connection to higher pharmaco-therapeutic classes. Because precision is needed for this basic grouping, the three constituent elements must be well defined: • the substance needs to be described at its most relevant granularity level (meaning the modified active substance for ionized chemicals) • the dose form should be at the most relevant granular level for administrable dose forms, and supported by solid characteristics and definitions (as in EDQM and in SNOMED-CT). • The strength needs to be a standardized strength, with a Basis of Substance Strength (BoSS) that is the same for all members of the group. This precision may sometimes be too far reached for clinical purposes, but is needed for regulatory purposes and pharmacovigilance use cases. It provides the basis for solid processes of aggregation to higher levels of abstraction, which might be more relevant for clinical care use cases. • Note : For more specific groupings, based on more intricate details of the medicinal products, open access to the standardized drug databases of the jurisdiction will be needed, to provide details, such as presence of inactive ingredients of clinical interest, other inactive ingredients, shape, colour, taste, range of permitted routes of administration, pack size, as well as regulatory information such as belonging to controlled substances, Over the Counter (OTC) or on prescription (Rx), subject to reimbursement rules, prices, etc. For a description of what is needed beyond this basic identification to run adequate decision support systems, we recommend a publication of the pharmacological department of Heidelberg University, Germany. (Senger et al., 2011) 3.5 Responsibility for PhPID production The exact and standardized determination of the 3 main elements in the identification of medicinal products is of course the responsibility of each local jurisdiction. However, super-national governance and validation process of these crucial attributions will be needed, if interoperability for this basic and pivotal concept is to be achieved. Increased cooperation between industry, agencies, and institutions such as the WHO Uppsala Monitoring Centre for Pharmacovigilance can ensure a smooth validation of this basic identification of any (new and old) medicinal product (see further). Once a limited number of countries have identified and validated the PhPID of all their medicinal products, and provided that this information becomes publicly available, it will be much easier for other agencies to follow the same pathway. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 18 of 61 4 Moving to pharmacotherapeutic groups, by gradually building groups of higher level of abstraction. 4.1 Starting with the level of PhPID The Pharmaceutical Product Identifier (PhPID) groups medicinal products from different countries and different companies that share the same substance(s), dose form and strength. These 3 elements are expressed in standardized terms from controlled terminologies, at a high level of granularity (modified substance, when available; granular dose form (EDQM or SNOMED-CT); and strength. This triplet of elements pulls together the (almost) identical medicinal products of different countries and companies, and provides a solid foundation for further aggregation. In this deliverable, we will use the example of the medicine amlodipine (for the treatment of hypertension and angina pectoris), and for illustrations with authorized medicinal products, we will use products from the Belgian Market, as these data are readily and publicly available in the SAM (Source Authentique de Médicaments) database of the eHEALTH system. 4.1.1 Identifying the triplets for amlodipine as an example As an example, Table 1 lists the available combinations of substance / dose form/ strength for medicinal products containing only the calcium antagonist amlodipine (so not combination products). This medication is marketed in 3 modified substances (amlodipine besylate, amlodipine mesylate, amlodipine maleate). Hence all these products contain the single moiety “amlodipine”, but with different modifiers. There are 2 strengths (5 mg and 10 mg) and three different dose forms available (capsule,hard; tablet; coated tablet). That leads to 9 different triplets. Looking at the data for Belgium, we can see that there is amlodipine besylate / capsule, hard / 5 mg and amlodipine besylate / capsule hard /10 mg (both represented with 1 brand, marketed by the originator). No other company markets amlodipine in this dose form. There are two other triplets with amlodipine besylate / tablet, one with 5 mg (marketed by 6 generic companies) and one with 10 mg (marketed by 7 generic companies). Triplets with amlodipine mesylate are not (anymore) available in Belgium. Triplets with amlodipine maleate come in tablets for 5 mg, and in tablets and coated tablets for 10 mg;one company has a coated table that can be split (a divule) of 10 mg, which can easily be broken in two halves. Most companies bring their medicinal products in a small (28 to 30 pack units) and a large pack (98 to 100 pack units). (See Table 1 for an overview). In Annex 1, the complete list of medicinal product packs in Belgium is given for the 35 substances of the UNICOM Pilot Product List). D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 19 of 61 Table 1. Overview of Pharmaceutical Product Groups in the Belgian national market), containing the single moiety amlodipine. In the UNICOM Pilot Product List small team, a comparison was undertaken with the UK, the Netherlands, and Germany, with respect for the non-public character of the drug data in the Netherlands and Germany. In the UK there are about 40 medicinal product packs (all oral forms and 5 or 10 mg) with 1 originator and 19 generic companies (data with courtesy from Julie James). In the Netherlands, there are 30 medicinal products (number of packs not clear) from 1 originator and 9 generic companies, including unlicensed paediatric doses (1 mg) and liquid dose forms (data with courtesy from Leonora Grandia). In Germany, there are 55 medicinal product packs, from 1 originator and 29 generic companies; one company licenses a 7.5 mg strength (data with courtesy from Ursula Tschorn). Extending this analysis to these and other countries will lead to a further (but moderate) increase of number of triplets needed to represent amlodipine. Pharmaceutical Product Group (virtual medicinal product) Data from Belgium Modified substance Granular dose form Strength Orginator Company Generic Companies amlodipine besilate capsule, hard 5 mg 10 amlodipine besilate tablet 5 mg 0 6 amlodipine mesilate tablet 5 mg 0 0 amlodpine maleaat tablet 5 mg 01 amlodipine besilate capsule, hard 10 mg 1 0 amlodipine besilate tablet 10 mg 05 amlodipine mesilate tablet 10 mg 00 amlodipine maleate tablet 10 mg 0 1 amlodipine maleate coated tablet 10 mg 01 Note: There is 8 active companies in Belgium (1 orginator, 7 generic companies) There are 16 medicinal Products and 28 Medicinal Product Packages available for amlopdine These belong to 7 different Pharmaceutical Product Groups, each to be defined by a PhPID For the prescriber there are 2 basic options: amlopdipine oral 5mg and amlopine oral 10 mg D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 20 of 61 4.1.2 Looking up the available coding systems Once the available triplets for a given substance are determined, it is possible to look for the exact codes, that represent each of the three basic elements. Coding systems are needed for the modified substances or inert moieties, for the dose forms, and for the units of measurement of strength. For “modified substance” at least 5 coding systems are available WHODRUG, EUTCT, UNII, CAS, and SNOMED-CT. For dose form, 3 coding systems are available:EDQM, SNOMED-CT, and RxNorm For strength only one, widely recognised coding system is available, namely UCUM (Unified Code for Units of Measure, from the Regenstrief Institute. 4.1.3 Looking up available codes for modified substances In Table 2, we provide an overview of the coding numbers for the 3 modified substances of amlodipine. It remains to be seen if one of these codes can be the best candidate for a global solution or whether a new global code needs to be defined. It is not clear yet, whether through the SPOR system a European code will be imposed in Europe (either the EUTCT or a new SPOR-code for substance). In this stage of the development of IDMP implementation this might be problematic. Table 2. Available codes for the 3 modified substances of amlodipine. Note: • In EUTCT the label is amlodipine besilate and amlodipine mesilate (with besylate and mesylate as synonyms). • For Amlodipine mesylate, very few pharmacovigilance reports, indicating very limited commercial activity. • Theoretically, there is also amlodipine benzoate, but not marketed in the EU (and hence, no pharmacovigilance reports) • It must be noted that the codes in the WHODRUG Dictionary and the terms in the INN vocabulary are currently not connected. 4.1.4 Looking up avaible codes for dose forms It is clear that the IDMP standard promotes the use of EDQM terminology for dose forms. Table 3 lists the available codes for 3 dose forms of amlodipine. Other coding systems, such as SNOMED-CT and RxNorm also provide coding and terminology, but both systems would have only one entry to describe the 3 EDQM dose forms listed here, namely: Oral Tablet. There has been a proposal from the FDA to use the characteristics of the EDQM terminology as the basis for a new, global coding system. Coding possibilities for amlodipine Modified substances WHODRUG INNM EUTCT UNII CAS SNOMED-CT amlodipine besylate 00972401001 100000090079 864V2Q084H 111470-99-6 84976003 amlodipine mesylate 00972404001 100000089571 291Y33EZHA 246852-12-0 not present amlodipine maleate 00972403001 100000089370 CQ27G2BZJM 88150-47-4 421048000 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 21 of 61 Table 3. Available codes for 3 dose forms of amlodipine. 4.1.5 Looking up available codes for strength Units In the UCUM system, the code for mg is “408”. (table 4) Table 4. Ucum code for milligram 4.1.6 Completion of collection of codes in preparation of PHPID production With this approach all elements for starting a HASH function that will result in an PhPID identifier are gathered. This work will have to be completed by looking at the medicinal products that contain amlodipine in combination with other substances (combination preparations). Similar work must be performed for each of the 35 selected substances of the UNICOM Pilot Product List, and ultimately for all medicinal products from the national drug dictionary, country by country. It may come as a surprise that at this stage of the development of the implementation of IDMP on a global scale, crucial issues of uncertainty about the procedure to follow in the production of the Pharmaceutical Product Identifiers are still not settled, neither in the EMA IDMP Implementation Guide, neither within UNICOM, neither at the global level. Intense concertation on the global level, but also experience in pragmatic pilots, facilitated by UNICOM, and maybe coordinated by WHO-UMC, may be useful to achieve consensus on this matter. 4.2 Building up to the level of Virtual Medicinal Product Group (VMPGroup) The group of medicinal products with the same PhPID is defined by the modified substance (INNModified) for most of the chemical substances, the most important type of substances in pharmacology.3 For clinical use cases, this may not always be relevant. Some may argue that there are clinical differences between the different salts of amlodipine, while others may consider their therapeutic activity equivalent, making it irrelevant to know which of the salts is used. The prescriber may accurately express his/her therapeutic intent by specifying that the pharmacist must dispense • a medicinal product containing amlodipine (without specifying the salt), • an oral dose form • choosing the 5 mg strength, rather than the 10 mg 3 For a full discussion of the characterisation of different types of substances, see the documents prepared for WG6 of ISO, by Jean-Jacques Gonzages and Herman Diederik. Granular Dose form EDQM capsule, hard 10210000 tablet 10219000 coated tablet 10220000 Unit of strength UCUM mg 408 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 22 of 61 In Table 5, the possible expressions for these choices are given. Table 5 . Triplets of higher aggreation for amlodipine Further details (e.g. tablet that can be split or not, coated tablet or not) can be specified or left to the discretion of the pharmacist, in dialogue with the patients. This grouping might also provide the basis for transparent and fair business rules for substitution regulation. In Belgium the principles for establishing these VMPGroups for INN Prescribing have been operationalised in regulation from the agency. (Van Bever et al., 2014) Again, it is quite possible to gather the codes from official terminologies (in casu WHO for the label of the INN, EDQM for Intended Site, and UCUM for strength, in preparation for a hash function to produce the identifiers for these concepts. For dose form, it is proposed to use the high-level categorisation of the intended site characteristic of EDQM, rather than the very granular EDQMP dose form. It is possible to create intermediate classes (see further) to assist the prescriber in adequate choices with regard to dose form. Snomed-CT has a similar concept as attribute to the dose form. Virtual Medicinal Product Group INNsubstance Intended Site Strength amlopidine oral 5 mg amlodipine oral 10 mg D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 23 of 61 4.3 Building up to the level of Virtual Therapeutic Moiety Level (VTM) Finally, it is possible to make abstraction of the dose form and strength, and only look at the INNsubstance as a criterion to group substances and medicinal products. This grouper concept collects all the available modified substances. Table 6 provides the possible coding for this concept, in the WHODRUG-INN coding of the INN nomenclature. Table 6. Potential code for the grouper of modified substances of amlodipine. Using this concept, we can now construct a collection of medicinal products that all contain this INNsubstance (hence any form of modified amlodipine) in medicinal products with amlodipine as a single substance or in combination products, where amlodipine is one of the constituents. In Snomed-CT, the most related concept would be “Medicinal Product ONLY or MP ONLY”. According to the Snomed-CT definition: “MP ONLY is: an abstract representation of a medicinal product based on description of only and exclusively the active ingredient substance(s) that it contains but regardless of any modification of those active ingredient substance(s). The definition indicates that this concept and other related concepts are meant to make an abstract description of individual medicinal products and is part of the international release (not the national extension). (SNOMED DRUG MODEL description). Note: For the purpose of naming groups of medicinal products by their substances, SNOMEDCT also uses a general concept, called “structural grouper”, defined as: A concept grouping together medicinal products based on the chemical structure of their active ingredient nnsubstance(s). This level is very close to the fifth level of the ATC classification, which is one way to standardize the concept, serving as an entry to the 5-level taxonomy for Drug Utilisation Research of the World Health Organisation (see further). Situated at the same level is also the European Union Reference Dates (EURD) list, that governs the coordination of Periodic Safety Update Reports (PSURs), by assigning to each active substance or combination of active substances a responsible agency and a responsible company, at the European level, with the task to periodically review on fixed dates the safety reporting of the therapeutic arsenal. Note: o PhPID_L1 is also an aggregation of medicinal products but characterised by the modified substance (if pertinent). Its equivalent in Snomed-CT is Medicinal Product Precisely, defined as: an abstract representation of a medicinal product based on description of only and exclusively the precise active ingredients it contains o Snomed-CT also has the concept Medicinal Product Form (as ONLY and as CONTAINING), which is intended as a combination of substance and form. o Another concept is the “therapeutic role grouper”, defined as: a concept grouping together medicinal products based on a broad description of their use in treatment of disease. That is a very general concept, but laying the bridge to pharmacotherapeutic classification. In addition, the rigorous ontological structure of Snomed-CT allows complex multi-axial classification. For a full discussion of the grouper concept . o The IDMP standard proposes the concept “cluster”, a vague and general concept, that is not further elaborated. INNSubstance WHODRUG-INN amlodipine 00972401001 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 24 of 61 4.4 Dealing with combinations At this level it becomes important to be able to handle single active substances and combination of active substances, and more importantly the relation between single substances and combinations. In the examples described up to now, only medicinal products were used, containing a single substance. Yet, in pharmacology, a number of combination products exist. While clinical pharmacologists in general are not in favour of combining more active substances in one medicinal product, the number of combinations is high in ill-regulated markets, with many questionable combinations.(Wirtz et al., 2013) In the past decade, mainly in the field of cardio-vascular primary and secondary prevention, the number of combinations of cardiac medicines has substantially increased, mainly because the scientific associations of cardiologists advocated them. The case of amlodipine is a good example, as it is present in many combination products (see Table7). Table 7. Amlodipine single and combinations Medicinal products with only amlodipine Amlodipine (single) amlodipine oral 5 mg amlodipine oral 10 mg Medicinal products with amlodipine in combination ACE inhibitor + Calciumantagonists Perindopril + amlodipine perindopril + amlodipine oral eq. 4 mg + 5 mg perindopril + amlodipine oral eq. 4 mg + 10 mg. perindopril + amlodipine oral eq. 8 mg + 5 mg perindopril + amlodipine oral eq. 8 mg + 10 mg Ramipril + amlodipine ramipril + amlodipine oral 5 mg + 5 mg ramipril + amlodipine oral 5 mg + 10 mg ramipril + amlodipine oral 10 mg + 5 mg ramipril + amlodipine oral 10 mg + 10 mg Sartans + Calciumantagonists Olmesartan + amlodipine olmesartan + amlodipine oral 20 mg + 5 mg olmesartan + amlodipine oral 40 mg + 5 mg olmesartan + amlodipine oral 40 mg + 10 mg Telmisartan + amlodipine telmisartan + amlodipine oral 80 mg + 5 mg telmisartan + amlodipine oral 80 mg + 10 mg Valsartan + amlodipine valsartan + amlodipine oral 80 mg + 5 mg valsartan + amlodipine oral 160 mg + 5 mg valsartan + amlodipine oral 160 m + 10 mg Ace-inhibitors + Calciumantagonists + Diuretics Perindopril + amlodipine + indapamine perindopril + amlodipine + indapamide oral eq.4 mg+5 mg+1,25mg perindopril + amlodipine + indapamide oral eq. 4 mg + 10 mg+1,25 mg perindopril + amlodipine + indapamide oral eq. 8 mg + 5 mg + 2,5 mg perindopril + amlodipine + indapamide oral eq. 8 mg + 10 mg + 2,5 mg Sartans + Calciumantagonists + Diuretics Olmesartan + amlodipine +HCT‡ olmesartan + amlodipine + HCT oral 20 mg + 5 mg + 12,5 mg olmesartan + amlodipine + HCT oral 40 mg + 5 mg + 12,5 mg olmesartan + amlodipine + HCT oral 40 mg + 5 mg + 25 mg olmesartan + amlodipine + HCT oral 40 mg + 10 mg + 12,5 mg olmesartan + amlodipine + HCT oral 40 mg + 10 mg + 25 mg. Valsartan + amlodipine +HCT‡ valsartan + amlodipine + HCT oral 160 mg + 5 mg + 12,5 mg valsartan + amlodipine + HCT oral 160 mg + 5 mg + 25 mg valsartan + amlodipine + HCT oral 160 mg + 10 mg + 12,5 mg valsartan + amlodipine + HCT oral 160 mg + 10 mg + 25 mg valsartan + amlodipine + HCT oral 320 mg + 10 mg + 25 mg Statins + ACE-Inhibitors + Amlodipine Atorvastatatine + Perindopril + Amlodipine atorvastatine + perindopril + amlodipine oral 0 mg + eq. 4 mg + 5 mg atorvastatine + perindopril + amlodipine oral 20 mg + eq. 4 mg + 5 mg atorvastatine + perindopril + amlodipine oral 20 mg + eq. 8 mg + 5 mg atorvastatine + perindopril + amlodipine oral 20 mg + eq. 8 mg +10mg atorvastatine + perindopril + amlodipine oral 40 mg + eq. 8 mg+10mg D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 25 of 61 There are several ways to deal with the complexity of combinations. (see figure 1). One can split a group of medicinal products only containing amlodipine, and another group with combinations all containing amlodipine but specifying the other constituent substances. One can make a collection of all products containing amlodipine (without specifying the other constituent substances) and with an included subset of medicinal products that only contain amlodipine. Finally, it is possible to create a database with medicinal products, either single or combinations, but all containing amlodipine, with another database where all single components are listed, with pointers from each component to the relevant medicinal products in the main database). Figure 1. Three ways of dealing with aggregation of combination medicinal products. Threeways of dealing with aggregation of medicinal product combinations Medicinal Products only containing amlodipine amlodipine amlodipine + a amlodipine + b amlodipine + c amlodipine + d amlodipine + e amlodipine + a + b amlodipine + a + f amlodipine + a + g + i Amlodipine besylate Amlodipine maleate a b c d e f g i VTMGroup Precise active ingredients All Medicinal Products containing amlodipine (either as single or in combination, without specifying the other components) Medicinal Products only containing amlodipine All Combination Medicinal Products containing amlodipine D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 32 of 61 6 Analysis of the 3 basic elements of IDMP (substance, dose form, strength) In this chapter, we describe analysis and research performed to clarify remaining uncertainties, surrounding the 3 basic concepts of the IDMP standard 6.1 The hierarchy of substances 6.1.1 Physical reality, abstraction, and levels of hierarchy in the representation of substance During the development of the Drug Ontology in the OBO FOUNDRY, it was recognised that single molecules, even if they pertain to strong medicines, do not heal. The single molecule may have functional properties (a disposition) for instance to bind to a specific receptor. For therapeutic effect to occur (at least as believed in allopathic medicine), a huge number of these molecules in a single pill is needed to expect some therapeutic effect. (Hanna et al., 2016). While each individual molecule has a molecular mass, all the molecules together in a pill (named here somewhat peculiar “a scattered aggregate”) carry a weight, which is a key element of the expression for the strength of the pill. As an example, the active substance of amlodipine is a white crystalline powder, which can be processed into a pill. The weight of the active substance might only be a fraction of the total weight of a pill (a few grams). (See figure 4) Figure 4. Substance as a molecule and substance as a "scattered aggregate" In the previous chapters we have learned that amlodipine has 3 different modifiers, amlodipine besylate, amlodipine mesylate, and amlodipine maleate. These substances exist under the form of a powder, which is constituted of two parts: first the moiety (the active part of the molecule) and second the salt. Together they form a physical reality as a white crystalline powder, and can be compounded in solid tablets, and will dissolve in enteric fluids, when ingested. After ingestion and dissolving, the molecule will split in the active moiety (the base) and the salt anion. The active part will then be absorbed, flow into the blood stream, and reach the intended sites of the body. The distinction between active ingredient as a molecule or as an scattered aggregate (Drug Ontology, OBO Foundry) Amlodipine besylate as a molecule Has a molecular mass Has a mechanism of action (disposition) (calcium antagonism) Amlodipine besylate as a “scattered aggregate” of molecules in a tablet Has a weight (as part of the tablet weight) Has a therapeutic role (lowers hypertension and relieves angina pectoris) Hanna J, Bian J, Hogan WR. An accurate and precise representation of drug ingredients. J Biomed Semantics. 2016 Apr 19;7:7. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 33 of 61 In Table 8, we listed the structural formulas and the molecular mass of the moiety and of the 3 modified substances, to illustrate the differences, which justifies that in the Chemical Abstract System, each of these physical entities receives a different code. Table 8. Description of moiety and modified substances of amlodipine As can be seen from the figures the molecular mass of the moiety is smaller than the modified substances, and each of the modified substances has a different mass. Description of modified susbstances and moiety of amlodipine Formula Molecular mass CAS-number amolodipine C 20 H 25 Cl 1 N 2 O 5 408.9 g/mol 88150-42-9 amlodipine besylate C 20 H 25 ClN 2 O 5 · C 6 H 5 SO 3 H567.0 g/mol 111470-99-6 amlodipine mesylate C 20 H 25 Cl 1 N 2 O 5 . C 1 H 4 SO 3 H505.0 g/mol 246852-12-0 amlodipine maleate C 20 H 25 ClN 2 O 5 . C 4 H 4 O 4 524.9 g/mol 88150-47-4 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 34 of 61 When looking in fig. 5 at the chemical structure of three modified substances, we can see that there is a common element in the structure, the biggest part is the same in the three modified substances. It is the moiety. Each of the modified substances has a second part, called the modifier. The modifiers are different, and because of that the molecular mass of the modified substance will be bigger than the moiety and different for all three modified substances. Note: To determine the strength of a medicinal product, we will need to know what will be the basis of that calculation. What will be the Basis of Strength Substance (BoSS)? Will it be the moiety or will it by the modified substance. If it is the moiety, products with different modifiers can still be equalised in strength (amlodipine 5 mg per tablet). If the modified substance is the Basis of Strength Substance then there can be subtle differences in the expression of strength of different modified substances for the same moiety. An example is the substance perindopril, marketed by different companies in tablets of 4 or 5 mg, while there is no real difference in the mass of the active moiety, without the salt. This distinction is often misunderstood and can be and was the reason for a lot of confusion in marketing authorisation. Figure 5. Molecular structures of amlodipine moiety and modifiers Amlodipine Besilate Amlodipine Mesilate Amlodipine Maleate D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 35 of 61 6.1.2 Levels of hierarchy of substances Having made these fundamental observations, we can now turn to the meaning of the terms that are used to describe different concepts related to substance. Again, we will look at the example of the term “amlodipine”, providing our definition for the different meanings of the same term. Figure 6. Ambiguity in the terms "substance" and "modified substance" Three meanings of a substance term Amlodipine (1) Term for the physical reality of chemical molecule, which consitutes the active part of an ingredient with therapeutic role. This molecule has a chemical structure, molecular mass, a code in the CAS-system, and a mechanism of action. Amlodipine (2) Term for the collection of modified substances (amlodipine besilate, mesilate and maleate), which all contain amlodipine (1) Amlodipine (3) Term for the collection of medicinal products that contain any one of the 3 modified substances (named with amlodipine (2)), and no other ingredients with an active role. A medicinal product can be entered in the collection even is the modifier is unknown. Two meanings of a modified substance term Amlodipine besylate (1) Term for the physical reality of a chemical molecule, consisting of the active part and the salt. This molecule has a chemical structure, molecular mass, a code in the CAS-system, and a mechanism of action Amlodipine besylate (2) Term for the collection of medicinal products containing this specific modified substance D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 36 of 61 6.1.3 The case of moieties without a modifier Most of the active substances are chemicals, and most chemicals consist of a moiety and a salt or an ester. Some chemicals do not have a modifier. An example is carbamazepine. Figure 7. Carbamazepine, a molecule without a modifier This molecule will exist on its own, will not split in two parts when dissolved, and has only one CAS number (no CAS-number for modified substances). It is a moiety, but not in the sense of the moiety of amlodipine. It is this concept that is needed to accurately describe the substance. If it needs to be presented at the level of a grouper concept, it will be by a collection with only one item. This situation results in a problem, an unresolved ambiguity in the representation of substances, that is recognised within EMA, FDA, WHO_UMC, and Standard Developing Organisations (SDOs) and hopefully adequately dealt with. The problem can be best illustrated with an image (see Fig. 10). In case of modified substance it must be determined what the Basis of Strength Substance (BoSS) is: the moiety or the modified substance. In case of a moiety without modifier there can be no confusion. Figure 8. Moiety and Precise Active Ingredient: ambiguity of concepts or of words? Two kinds of precise active ingredients modified substance amlodpine besilate moeity without modifier carbamazepine Two kinds of moeities (+) the ionized moeity amlodpine moeity without modifier carbamazepine D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 37 of 61 The intense cooperation between experts in WP2, WP8, WP9 of the UNICOM Pilot Product List Project and the international cooperation with WHO-UMC and FDA contributed to raise awareness on these issues. There is a willingness to come to a consensus, and to provide the substance registers of the USA and EU with a solid framework to address the problem of the hierarchy of substances. 6.2 Dose forms and their characteristics 6.2.1 Existing coding systems for dose forms There are 3 coding systems for standardizing dose forms of medicinal products: Snomed-CT, RxNorm, and EDQM (advocated by the IDMP standards). The dose form model of Snomed-CT is similar to EDQM, with terms, definitions, characteristics, and an ontological structure. The granularity is slightly less (which may be beneficial), the value sets for the characteristics are slightly different (which is remediable), and the definitions are more consistently formalized. In RxNorm, the dose form is actually a value set to describe characteristics of products. It is much less granular (179 dose forms in RxNORM, vs 428 in EDQM). Definitions are not very precise. There is a rudimentary collection of dose form types, which tend to overlap. Note: WHODrug uses New Form Code of the European Pharmaceutical Marketing Research Association, another standardised dose form coding system. Also CDISK has controlled terminology for dose forms. 6.2.2 The value of EDQM EDQM Standard terms for Pharmacology originates from the European Pharmacopoeia, with a long tradition of excellence in maintaining the controlled vocabularies. It is a terminology for Europe, but it has been recognised internationally by the International Committee for Harmonisation (a platform for the EU, US, and Japan) for global standardization in pharmaceutical issues of regulation and research. The ISO standard regarding dose forms is currently under revision in ISO WG6 of TC125, coinciding with the analyses performed in UNICOM. For the global identification of medical products, ISO has opted for a granular description of the dose form, and the granularity of EDQM (428 human terms) is much deeper than the terms list of RxNorm (179 terms). The value of the standardized use of a core set of characteristics for the dose form (needing transformation, release characteristics, intended site, and administration method) is widely recognised also by the FDA, and ICH. A pilot project has been conducted by the FDA to use the characteristics of the EDQM Dose Forms, as the basis for PhPID production, and this has been explored in collaboration with WHO-UMC (See Annex 3) Hence the time is right for a serious reflexion on how this excellent European resource can grow into a global terminology. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 38 of 61 6.2.3 Possible improvements for EDQM in the transition to a global terminology Within WP8 in UNICOM (IDMP and Clinical Care), an analysis was performed of EDQM, in preparation of pilots to use IDMP in decision support systems. This analysis was initiated by Robert Vander Stichele, I-HD, lead of WP8, together with a semantic expert (Joseph Roumier) and a drug database expert (Dirk Vannimwegen). It is an integral part of this deliverable, and can be consulted in Annex 4. Here we will limit ourselves to a brief summary. First, a selection was made of the terms for human use, leading to an analysis of 428 current dose forms. The following minimal interventions were explored: • All pharmaceutical dose forms that need transformations were explicitly aligned with their resulting administrable dose form (and basic dose form and state of matter). This work was particularly important as the IDMP standard stipulates explicitly that for the calculation of the PHPID-L4, the code of the administrable dose form must be used, and that the strength must be expressed as the strength of the administrable dose form. • All definitions were analysed and compared with a set of the characteristics for definitional completeness and consistency • Value sets of the characteristics were analysed for multiplicity (more than one value for a characteristic of a dose form). A new value was created combining the values of the constituent values (instead of storing the possible values in separate columns). • The dose forms with the value “cutaneous/transdermal for intended site were critically analysed, and the replacement with a single value (either cutaneous or transdermal), was explored which was feasible in almost all the cases. • The intended site of sublingual dose forms was oro-mucosal, while there is a clear intention to reach the systemic circulation with these dose forms (sometimes intentionally sublingual to bypass the first-liver pass). The value “sublingual” was suggested. • An exploration was made concerning the possibility to also characterise the dose forms as having a systemic or local effect. A few dose forms were identified where that was problematic, in the cutaneous/transdermal, the nasal, and the rectal dose forms, but for the vast majority of dose forms this was perfectly possible. We came to the conclusion that a thorough analysis was needed of the definitional value of the characteristics and its use in the PhPID production process (as proposed by the FDA). To be able to perform this analysis, an experimental new version of the EDQM database of terms was created, implementing the changes described above. This database and the implicit structure were transferred to WebProtegé, an online ontology manager, also used during the transition development from ICD10 to ICD11.5 This allowed ontological work by collaborating experts, using the groupware facilities of WebProtégé. In addition, the data were also transferred in an SQL database, for further analysis. The 428 lines of the EDQM Human Dose Form Database were then ordered on the four characteristics (transformation, Release Characteristics, Intended Site, Administration Method). Collections of dose forms with the same four values for their characteristics were constructed. These collections were then analysed separately for all collections with the same intended site, and by needing transformation or not. The result was manually analysed and meaningful groups were formed or split. Basic criteria for group formation were the following ones: similarity of characteristics, clinical relevance, and similarity of the way strength would be expressed for such a group. This also resulted in further suggestions for improvement of the basic data, but also in an initiation of the steps to develop an ontology of dose forms. 5 https://protegewiki.stanford.edu/wiki/WebProtege D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 39 of 61 6.2.4 Possibility of creating a dose form ontology. The 428 dose forms of EDQM can already be divided into groups sorting them on the 25 values of the intended site (oral, nasal, rectal, etc). We proposed and developed one intermediate level of granularity, depending on the similarity of the characteristics as described above. The result is described in the Annex 4, and will be discussed further with experts of EDQM and formally presented to WG6 of ISO/CEN. 6.2.5 Comparison to RxNorm and Snomed-CT dose forms For the comparison with RxNorm we refer to the study report in Annex 5. It is obvious that the dose terms of RxNorm are used as descriptive variables, and do not constitute an ontology on their own. The level of granularity is much less (only 144 dose forms), and the definitions are broad. There is no formal system of characteristics, but the FDA is contemplating to use the EDQM characteristics (see WHO_UMC/FDA Pilot Report in Annex 3). There is a rudimentary but overlapping list of drug form groups. Preliminary discussions were initiated with the Dose Form experts of SNOMED-CT, mainly to explain the approach, and exchange ideas on mapping or alignment between the 2 systems. 6.2.6 Application to the UNICOM Pilot Product List As is opted for the granular EDQM dose form codes, any dose form encountered during analysis of national medical products can now be standardized to an EDQM dose form term, with the corresponding codes and characteristics. The question remains whether in the SPOR data register the code for dose forms will be the existing EDQM code, or a new proprietary code from EMA. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 40 of 61 6.3 The complexity of determining strength of medicinal products Historically, in the regulatory process there are many examples of inconsistencies and discrepancies in the determination of the strength of similar medicinal products, and when considering the emergency authorisation of the first Covid-19 vaccines, this proved to be agin the case.. In the report of the WHO_UMC/FDA report (see Annex 3) also inspired by the EU IDMP Implementation guide6, many challenges have been listed with remedial suggestions. The following tables and figures of this paragraph were extracted from the WHO_UMC/FDA Pilot report to demonstrate the approach to find a practical way to implement the complex framework developed in IDMP, to be able to represent it in a standardized way all the possible variations of strength expression. 6.3.1 Strength determination and substance First, 3 different kinds of strength are defined: presentation strength, concentration strength, and reference strength (see Table 9). Table 9. Strength definitions Strength definitions Strength (Presentation) When the strength of a substance is described as a qualitative term describing the discrete unit in which a Pharmaceutical Product is presented Strength (Concentration) When the strength of a substance is expressed as the amount of substance per unit of measurement, such as millilitre or gram Reference Strength The strength for the active moiety. If there is no reference substance, the active substance and its strength must not be repeated in the data object reference strength It is clear that correct and standardized determination of strength depends predominantly on a thorough insight in the molecular mass of moieties and modified substances, and for that concrete guidance is given in the report with examples (See table 10). 6 https://www.ema.europa.eu/en/documents/regulatory-procedural-guideline/products-management-services-implementationinternational-organization-standardization-iso-standards_en.pdf D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 41 of 61 Table 10. Calculation of the reference strength If in the SmPC the active substances are given as salts/esters or pro-drugs and the strength corresponds to the salts/esters or pro-drugs, the reference substance strength is being calculated as follows: o Substance (mg)/(substance molecular weight(mg/mol)* reference substance molecular weight(mg/mol)) If in the SmPC the active substances are given as salts/esters or pro-drugs and the strength corresponds to the active moiety, the salts/esters or pro-drugs substance strength is being calculated as follows: o Reference substance (mg)/(reference substance molecular weight(mg/mol)* substance molecular weight(mg/mol)) If in the SmPC the active substance is the active moiety. In this case the active substance is identical to reference substance and no calculation is performed. This can again be illustrated with the example of amlodipine. (See table 11) Each of the 3 modified substances (salts) of amlodipine have a different molecular mass. But in the end the authorized strength that will appear on the package is 5mg. It is the reference strength, in this case weight of the moiety of amlodipine in one tablet. In the production process of pills, each company using a particular modified substance will have to calculate (using the formula’s described above) how much of the modified substance needs to be sprinkled into a tablet, to reach the exact reference strength of 5 mg. In table below, the result of this calculation is given up to 4 decimals of the weight (as this has to be an exact calculation for a precise production process). Table 11. Calculation of the weight of different modified substances for a given reference strength for the moiety amlodipine Amlodipine besylate 5mg 409 g/mol 567 g/mol 6,9315 mg Amlodipine mesylate 5mg 409 g/mol 505 g/mol 6,1736 mg Amlodopine maleate 5mg 409 g/mol 530 g/mol 6,4792 mg Molecular mass of the moiety Molecular Mass of the Modified Substance Reference strength Weight of the scattered aggregate of the modified substance in the tablet D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 48 of 61 8 Building a UNICOM Repository of validated Pharmaceutical Product Identifiers - PhPIDs 8.1 Rationale The result of the UNICOM Pilot Project in terms of draft identifiers (with preliminary validation) will need to be collected. WP9 has proposed a FHIR resource for assisting agencies in experiments to implement full implementation of IDMP in their databases, by means of test FHIR server, with procedures based on compliance with the EU Implementation Guide, now in V2, and soon to be updated. However, it is worthwhile to organize a limited data collection of the basic elements which are needed for PhPID production. In the Pilot Product project, triplets of basic data have already been collected for medicinal products of 3 substances on the list, to be submitted to a draft production of PhPID by WHOUMC, during the UNICOM project. The resulting PhPIDs can be stored in a public repository (preferably in LOD (Linked Open Data), containing: • The draft PhPID identifiers • The codes and terms of their 3 basic constituents • The substance hierarchy • The dose form ontology • The links to pharmacotherapeutic classifications 8.2 ICT-Operationalisation A small team in WP8 (Task T8.1) worked out the ICTmodel for this repository, with a conceptual model (see figure11), a logical model (see figure 12), and a technical model 8.2.1 Conceptual model Figure 11. Conceptual model of a repository of PhPIDs D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 49 of 61 In this conceptual model, the database of draft PhPIDs is the kernel of this system. It contains the PhPIDs and their constituencies (modified substances, dose forms, strength). National drug dictionaries can provide a link to this database for those products where there is a PhPID available, and provided actual medicinal products for this PhPID are on the national market. There is a possibility to group similar PhPIDs to a higher level of abstraction, called Virtual Medicinal Product Group (e.g. amlodipine oral 5 mg). This abstraction is based on very simple ontologies. One aggregation is proposed for substance to make the aggregation of modified substances into a class of substances, containing the same moiety. Another simple ontology is for dose form, based on the characteristic for Intented Site, but with an intermediate level grouping to the granular EDQM dose form (see Annex 4). The link with international classifications will be established through this Virtual Medicinal Product Group, which will greatly reduce the workload of constructing the chain from high level international classifications to actual medicinal products in a country. The model also foresees a Virtual Therapeutic Moiety concept, which is an abstraction of the PhPID_level 1 (only substance). In these models no abstractions are foreseen for PhPID-Level 2 (substance + strength, considered as not really relevant), and for PhPID_Level 3 (Substance + dose form). However, concepts similar to these dose forms in Snomed-CT and RxNorm are easily programmable with this approach. The model foresees multilingual management of the value sets of the crucial variables, such as substance, dose form, units of measurement, of course, using as much as possible the EUTCT, EDQM, and SPOR facilities. 8.2.2 Logical model The logical model deals with: • The handling of combination products (in the PhPID concepts, Virtual medicinal Product Group Concept, and the Virtual Therapeutic Moiety Concept) • The precise implementation of the substance and dose form ontology. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 50 of 61 Figure 12. Conceptual model of a repository of PhPIDs 8.2.3 Technical model The full technical model of this database has been elaborated, in the aftermath of this deliverable, and will be populated with the results of the first experiments of the PhPID production. The repository can be populated with data from medicinal products, pertaining to the substances on the PPList, originating from a limited number of countries, to start with. 8.3 Fostering the impact of UNICOM on clinical care and research The ultimate goal of this approach is the very objective of this deliverable: finding a robust interoperable way of connecting precise medicinal product identification in the different constituencies with the international classifications. Each country in Europe and also elsewhere in the world has in most cases several databases with their medicinal products. The Marketing Authorisation Agency usually has a database collecting data on the process of accepting New Drug Applications and changes to the labelling (variations). Countries may have an official Drug Database, governing the processes of eHealth. In some countries, there is an independent drug information Centre, which produces web sites and printed material for health professionals and patients. Pharmacist associations, vendors of Electronic Health Records may all have their own drug database. International and national publishers of scientific information may also provide drug dictionary services. Some international publishers coordinate the maintenance of the connections to a number of national drug databases in an internal proprietary system, which is a costly endeavour. Developers of Decision Support Systems rely on international classifications to build their rules for decision support. However, connecting these rules to different national drug dictionaries is not an easy task, which hampers the diffusion of these valuable resources through the European market. If UNICOM can make a contribution here, the efforts will have been worth it. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 51 of 61 Figure 13. Connecting national drug dictionaries with international classifications 8.4 Prospects on sustainability of this UNICOM repository, beyond completion Such an accessible, open repository can provide valuable examples of linking of national medicinal products to existing PhPIDs and to international classifications. As the content grows, it can become more and more interesting for agencies and publishers to consult. It would also be instrumental to the pilot project of WHO_UMC and FDA, to publish their results. In a spirit of transferability, this repository can be planned to be handed over to supranational organisations, by the end of UNICOM. N C A M P D e H E A L T H E H R Draft PhPID Database VMP Group ATC+ROA SNOMED-CT RX-NORM Drug Ontology N C A M P D e H E A L T H E H R Country A Country B CHARACTERISTiCS Substance INN Dose form Aggregated Strength CHARACTERISTiCS Substance Granular Dose form EDQM Strength D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 52 of 61 9 Linking drug classifications to IDMP 9.1 A short and incomplete inventory of international drug classifications Pharmacotherapeutic classifications play a very important role in the organisation of drug information, labelling, pharmacovigilance, clinical care and research. Pharmacopoeia and handbooks of pharmacology may have complicated tables of content, while the Anatomical Chemical Therapeutic (ATC) Classification of the World Health Organisation (WHO) is a straightforward taxonomy with only 5 levels7. While clinical classification reflects the medical culture and therapeutic thinking, the ATC provides a tool for correct recording of the consumption of medicinal products (grouped in broad classes) during specific time periods in specific regions. At the world level, for pharmacovigilance applications, the Uppsala Monitoring Centre of the World Health Organisation maintains the WHODRUG dictionary, including Standardised Drug Groups (SDG)8 : Drug Classifications linked to national medicinal product dictionaries exist in many countries. In the USA, RxNorm is maintained by the National Library of Medicine, and integrated with the Medical Subject Headings (and its pharmaceutical classes) and UMLS (Unified Medical Language System)9. First Databank (FDB Med knowledge is a proprietary medicinal product dictionary, integrated with decision support systems for the USA, with international ambitions10 . Other example are the British National Formulary11, Vidal in France (partner in UNICOM)12, Rote Liste in Germany13 , the Belgian Centre for Pharmacotherapeutic Information14, and in the Netherlands Z-index15 and Farmacotherapeutic Kompas16, a compilation of labelling information at the level of the INN substance. Snomed-CT has an important drug class component and its own model to represent medicinal products with the possibility to link to national medicinal product dictionaries through national extensions. It provides the key to integrating drug information into medical documentation17. The Drug Ontology DrOn, related to the OBO Foundry of biomedical ontologies, is an embryonic attempt to integrate semantic web ontologies with description of medicinal products and drug classification18. In the Linked Open Data Network, the information from many pharmaceutical and chemical databases are intricately linked19. Creating reliable links between international drug classifications and national medicinal products Several international attempts exist to link the information resources mentioned above with a broad range of national medicinal product dictionaries. On the semantic web, there is a Linked Open Data Network LOD , linking together all semantic web databases, including life sciences20. 7https://www.whocc.no/ 8 https://www.who-umc.org/whodrug/whodrug-portfolio/whodrug-global/ 9 https://www.nlm.nih.gov/research/umls/rxnorm/overview.htm 10 https://www.fdbhealth.com/solutions/medknowledge-drug-database 11 https://www.bnf.org/ 12 https://www.vidal.fr/ 13 https://www.rote-liste.de/ 14 https://www.cbip.be 1515 https://www.z-index.nl/g-standaard 16 https://www.farmacotherapeutischkompas.nl/ 17 https://confluence.ihtsdotools.org/display/DOCMPM), 18 http://www.obofoundry.org/ontology/dron.htm 19 https://lod-cloud.net/ 20 https://lod-cloud.net/#life_sciences D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 53 of 61 Numerous national drug databases for regulation, science and clinical practice have links to the ATC classification, as it is the standard for producing drug statistics for international comparison. This (almost) ubiquitous link will probably be very important to build the bridge between national medicinal products and the proposed repository of PhPIDs. In addition, accurate description of pack size will help national actors to calculate more precisely the number of Defined Daily dose per package. This is a crucial operation to ensure standardised international comparison of drug consumption21 . The Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) is based on a conformational process of national drug data to the common standard of RxNorm. It is widely used in international pharmaco-epidemiological research22. The promise of IDMP implementation is that it will bring ease and precision to the process of linking national medicinal products and product packages to all the International classifications. 9.2 The need for a simple classification of indications and products for Medical Education and Patient Information Medical students, pharmacy students and nursing students must acquire a minimal working knowledge of the therapeutic arsenal, during their graduate and post-graduate training. Universities have selected a basic set of ailments, with which their students must become familiar, and also selected the pharmacotherapeutic classes, suitable for more intense teaching and for testing the achievement of educational objectives. The Netherlands and Belgium share an electronic platform for case-based, problem-oriented teaching of pharmacotherapeutics. At the kernel of this education instrument is a compact database of indications (the 300 most important in primary care, organised in 2 levels) and pharmaceutical classes (350 in also organised in two levels), linked to each other, so that choice of a particular indication will evoke a list of potentially pertinent pharmaceutical classes and substances. Currently this information resource is linked to the 5th Level of the ATC classification (the level of substance). In the second half of the UNICOM project, it is planned to link also to PhPIDs, to facilitate the international implementation of this electronic educational platform, in cooperation with the European ENLIGHT program23. Finally, for patients, names of pharmacotherapeutic classes may be an enigma. And yet, drug labelling contains many cross-references to pharmacotherapeutic classes. For instance, the labelling of amlodipine mentions that patients who are allergic to dihydropyridines should not take amlodipine. But how can the patient actually know whether one of the medications on his/her allergy list is a dihydropyridine? Building a controlled vocabulary of drug group names, with professional terms and laymen equivalents (either laymen terms or descriptions), has been realised in an old European Project, called Multilingual Medical Glossary, still available on the Web24.) Combined with the indication/pharmaceutical class resource described above, and linked to PhPIDs, this could provide the basis for international apps that allow patients to navigate drug labelling, and to interprete their medication list. These resources (currently in development) will be made available through the software factory, organised by WP6 in Unicom. 21 https://www.whocc.no/use_of_atc_ddd/ 22 https://www.ohdsi.org/data-standardization/the-common-data-model/ 23 https://enlight-eu.org/ 24 https://users.ugent.be/~rvdstich/eugloss/welcome.html D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 54 of 61 10 Linking decision rules in pharmacotherapy to drug classifications and IDMP 10.1 Explicit rules for (in)appropriate prescribing from internationally validated PIM-lists. In the past decades tools have been developed to assist researchers and health care professionals in analysing the quality of their prescribing. With the greying of society and the extension of the preventive and therapeutic arsenal, polypharmacy has become more and more prevalent in the population in general, among the older adults, and especially among the oldest old. More people reach advanced age, are confronted with multi-morbidity, and also with polypharmacy. For prescribers, finding the balance between appropriate treatment of multiple ailments and avoidance of the downside of polypharmacy is a delicate exercise.(Hoffmann et al., 2020) Several international validated tools for assessing the quality of prescribing have been developed, either as survey tool for field research, either as ICT-applications for clinical care. The latter are software tools that will perform a pre-analysis of medication lists of patients, resulting in a number of comments to the prescribing physician who will then decide to follow the advice or not. Explicit criteria from 3 internationally validated lists of explicit criteria for potentially (in)appropriate medication Lists were collected and brought together in a web-based repository. (Ivanova et al., 2018) • the EU-7 PIM list (geared to medications on the European markets; • the Beers’ list (developed in the US and used widely in the world), • the STOPP/START list (developed by Irish geriatricians, also looking at underuse). Figure 14. Selection process of criteria in a repository from 3 validated lists of explicit criteria of potentially inappropriate prescribing D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 55 of 61 10.2 Sources of clinical data and medications to apply the criteria Looking at the literature of studies, having used these tools, the main terminology systems were International Classification of Diseases and the ATC for medications, with the algorithms working on the medication list and on the problem list of the Electronic Health Record. As most criteria are rather crude, the information captured in an International Patient Summary could be sufficient for operating these criteria in a valid way. 10.3 Coding the rules to disease and medication identifiers The criteria are in the following format: IF the patient has disease X THEN Drug Y should not be taken BECAUSE Drug Y is contra-indicated for Disease X. Each mention of medication in the criteria in the repository has been recoded to the pertinent list of ATC codes. Hence, these links could be extended to IDMP; using the repository of PhPIDs, as proposed above. For the disease element in the rules, several classification systems (ICPC, ICD, SNOMED-CT) can and should be used, given the variety of implementation in clinical systems. Criteria with medications pertaining to the UNICOM Pilot Product List will be expressed in collections of PhPIDs and this information will also be made available in the software factory of WP6 by the second half of the UNICOM project. Knowledge drug databases will have to consider IDMP as a pivot between decision rules and medicinal product dictionaries (Hoffmann et al., 2020) (Hoffmann et al., 2020)(Eiermann et al., 2010) . 10.4 Piloting IDMP decision support in UNICOM It should be possible in the second half of the project to select a number of explicit PIM criteria, pertaining to medication on the UNICOM Pilot Product List, for experiments with International Patient Summaries, to be demonstrated in HL7 Connectathons. D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 56 of 61 10.5 Using explicit criteria to test the quality of clinical data in the International Patient Summary (iPS) In many health institutions (such as nursing homes), the medication list is amongst the most reliable data on the patients (often the only reliable data, available in code form). The quality of the clinical date (in the list of problems, diagnoses, indications) is often more problematic. The element of medication in the explicit criteria could be used to scan the medication list for relevant medicines. This can trigger requests for relevant clinical information (e.g. what is the indication for this medication? Is a well-known contra-indication for this medication present in the problem list?) These requests can then help the health care provider to complete the information in International Patient Summary, to make sure that it becomes a document that better reflects the current condition of the patient, and is a better basis for fair medical audit of the prescribing quality. Figure 15. Audit of quality of clinical documentation with clinical rules for appropriate prescribing 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 D8.1: Link between IDMP and Pharmacotherapeutic Groups and the Need for Medical Data Page 57 of 61 11 Conclusions The objective of UNICOM is the implementation of the ISO/CEN standards across the continuum of the Research and Development departments of the industry, the marketing authorisation eHealth drug Databases, the Medicinal Product Dictionaries, and the databases of vendors of Electronic Health Records and International Patient Summaries, and finally the data warehouses of the health care sector. The precise and standardised description of many essential details of each individual medicinal product in a given constituency is a crucial prerequisite. Three elements are of fundamental importance: substance, dose form, and strength. UNICOM will be instrumental in the pilots where coding systems will be chosen, decisions on granularity of information will be made, and procedures developed to feed HASH-functions to produce global identifiers, called Pharmaceutical Product Identifiers PhPID level 1 to 4. Full implementation of IDMP for the detailed description of all aspects of a medicinal product requires a full toolbox to service the requirements of specific use cases. A PhPID_Level IV (combining substance, dose form and strength) will groups medicinal products from different constituencies. The PhPID aspect of UNICOM can be the Swiss knife approach within IDMP, to allow application of the standards across different domains and use cases, while assuring interoperability between domains and systems. In this deliverable and in the Gap Analysis of WP1, a number of unresolved issues have been identified that stand in the way of operation routing production of global identifiers, such as the hierarchy of substance, the granularity of dose forms, the representation of strength according to substance definition and dose form pattern, and the handling of combination products. Methodological approaches to solve the remaining issues have been suggested, while stressing the need for international cooperation, involving also the national marketing authorisation agencies. Next steps are to insure demonstration of draft procedures for the production of the PhPID, based on examples from the UNICOM Pilot Product List. Building and maintaining a trustworthy repository of Pharmaceutical Product Identifiers will be crucial to initiate the applications that will demonstrate the value of IDMP. 3 Introduction The ISO standards for Identification of Medicinal Products (IDMP) describe the use of standardised definitions for the identification and description of medicinal products, such as, identification of Substances, Medicinal Product (MPID), Medicinal Product Package (PCID) and Pharmaceutical Product (PhPID) aiming to facilitate the exchange of medicinal product information in a robust and consistent manner. Implementation of the standards on a global level would improve interoperability across global regulatory and healthcare communities, ensuring unambiguous communication across jurisdictions. To have consistent identification according to IDMP, the substance ID and PhPID should be maintained at a global level. The Uppsala Monitoring Center (UMC) has, after discussions with WHO and regulators, agreed to generate and maintain global PhPIDs for pharmacovigilance processes as well as other applicable use cases. The PhPID is generated based on information about substance, administrable dose form and strength, all defined in the ISO IDMP standards. The pilot project will focus on an evaluation of predefined global dose form characteristic codes as input into the generation of PhPID. The original ISO 11239:2012 standard, Regulated Information on Pharmaceutical Dose Forms, Units of Presentation, Routes of Administration and Packaging, was based on the use of a single controlled vocabulary (i.e., EDQM) that regions could use to communicate with each other. Further, the technical specification, TS 20440:2016 guide for ISO 11239:2012, assumes the use of a single controlled vocabulary or a mapping to it. However, there is currently no agreement on such a central vocabulary, in particular, for pharmaceutical dose forms. Various regions are using their own set of terminologies for dose form, which are not harmonised and show different levels of granularity between regions, making a one-to-one mapping between a regional terminology and a centrally controlled vocabulary of low quality. Investigations have shown around 20-45 % one-to-one matches between the ISO 11239 compliant EDQM terminology and terminologies used by U.S. FDA (FDA) and Health Canada. Similar results were also shown for mappings to SNOMED, CDISC and EPMRA dose form terminologies. To solve the issues with mapping between different dose form terminologies, a proposal was made at International Standards organisation TC 215 WG6 October 2020 meeting to use a centrally maintained set of dose form characteristics to describe a dose form term and code, for use in global IDMP, and in generation of PhPID. To evaluate this new concept for description of dose form, U.S. FDA and UMC agreed to perform a pilot study. The pilot would: 1. Assign chosen EDQM dose form characteristics for US marketed medicinal products corresponding to their selected substances, a. Substances identified in the UNICOM Pilot product list1 2. Evaluate the performance of the dose form characteristics for generation of corresponding PhPID a. Generation will use numerical representations of dosage form together with substance and strength. 1 The UNICOM pilot product list has been created within the UNICOM project, a European Commission supported Innovation Action focused on the implementation of the ISO IDMP standards. The list contains medicinal substances that represent the range of challenges that exist for medicinal product identification information and the objectives for the list is to provide actual exemplar content for cross border perscription pilot as well as other areas of work within UNICOM. 4 Objective The objective of the pilot is to demonstrate that a selected set of dose form characteristics from EDQM and other potential characteristics can describe dose forms and be utilized as input in the generation of PhPID. Project Scope According to the ISO standard for Pharmaceutical product, ISO 11616, PhPID shall be presented for both active substance and specified substance, each containing four PhPID identification levels. The PhPID shall be generated using the corresponding ISO standards and technical specifications: Substance ISO 11238 and ISO/TS 20440 Administrable dose form ISO 11239 Units of measure ISO 11240 Table 1. Four levels of PhPID This project focuses mainly on exploring PhPID for active substance on the fourth level, calculating PhPID_SUB_L4 according to table 1. The data on medicinal products included in the pilot has been provided by U.S.FDA based on the substance data from UNICOM Pilot product list. 5 The dose forms used for the PHPID calculation are expressed according to four of the centralized core EDQM dose form characteristics (and their codes); Release characteristics, Intended Site, and Administration Method and Basic administrable dose form. The process for generation of PHPID includes the following steps: 1. Assignment by the FDA of the centralized core dose form characteristics for US marketed medicinal products corresponding to the substances in the UNICOM Pilot product list. 2. The UMC validate the FDA data and assign the relevant strength to generate corresponding PhPID, using numerical representations of dose form characteristic together with substance and strength; 3. FDA and UMC will perform a data equivalency assessment on the use of centralized core dose form characteristics and other potential characteristics for generation of PhPID. Data Processing Procedures The purpose of this section is to describe the process applied to collect and validate the data to be used for PhPIDs construction. The data, provided by U.S. FDA, included 1167 US marketed medicinal products with assigned centralized core dose form characteristics (Release Characteristics, Intended Site, Administration Method and Basic Administrable Dose Form) and information on active ingredient, trade name, Marketing Authorization Holder, route, state of matter, transformation information, UNII code, approved strength. In order to generate necessary input data for PhPIDs construction, information on substance/reference substance, dose form, unit of presentation, unit of concentration, and strength has been collected. In order to validate the data, the following process has been applied: 1. Substance information has been verified in SmPC, selected based on Trade Name, Marketing Authorization Holder, and Dose Form information, and relevant salts/esters of the substance have been defined appropriately; 2. Dose form information has been evaluated further in order to select a suitable pattern, using Pattern Framework. 3. The strength information has been evaluated depending on pattern selected in the step above - in some cases the strength (presentation) or strength (concentration) or both have been entered; 4. Strength information has been further evaluated in the SmPC, to determine if the strength corresponds to the active substance or to the reference substance (active moiety); If the strength corresponds to the active substance, the strength value has been filled in the substance field and reference substance strength (active moiety strength) value has been calculated in the field for reference strength and vice versa. Figure 1. Medicinal product verification process using SmPC. 6 Combination products (combination of substances) In case of combination products and how their respective substances and strength will be presented, each combination product has been assigned a unique ID number to identify the input data for all substances and their respective strength in the combination. Substance This section provides and overview on the procedures applied to analyze information on active substance, assumptions made, and challenges faced during the validation process. The first task in the validation process was to identify the active substance with therapeutic intent at the most appropriate level of granularity. For chemicals this means that the active substance must be defined, for example, as an active moiety or a salt/ester. If the active substance is a salt/ester or similar, a reference substance needs to be identified, i.e. active moiety. For biological substances and vaccines, a simplification was made where the substance was identified as the active moiety and no reference substance was considered. For identification of the substance to be submitted to the HASH function in this pilot, the WHODrug ID was used. For future use of a global substance ID, the level of granularity needed for unique identification of a substance will be based on the current investigation of minimal fields by ISO WG6. There were several challenges regarding substances (Table 2) that have been faced and corresponding assumptions taken for this pilot to mitigate the risks of PhPID generation inconsistency. 7 Table 2. Challenges and their description vs mitigation related to substance evaluation process Challenge Description Challenge Mitigation In case substances have been inadequately described in the label If a substance is inadequately described in the label, the active moiety is chosen as an active substance. Examples: 1. Ibuprofen and ibuprofen potassium, Trade name=Ibuprofen from SOFGEN PHARMACEUTICALS LLC and HUMANWELL PURACAP PHARMACEUTICAL WUHAN CO LTD. The substance was presented as a mixture between the free acid and salt, but no ratio was given. Hydrates When there is any inconsistency in the substance description, if it is a hydrate or not, a pharmacopeial monograph or a structure connected to an official name such as INN or USAN can be used as a reference. If there is no clear guidance, the non-hydrate form is used. The active moiety is used as the reference substance in either case. 1. Morphine sulfate – according to both USAN and Ph Eur the name Morphine sulfate is the pentahydrate, hence the pentahydrate is used in this pilot. 2. Amoxicillin – according to USAN the name Amoxicillin is the trihydrate and the Ph Eur only have a monograph for the trihydrate, hence the trihydrate is used in this pilot. 3. Amoxicillin (unspecified), the non-hydrate was chosen 4. Omeprazole Magnesium –according to USAN and Ph Eur there is only the non-hydrate variant and is hence used instead of the trihydrate mentioned in the label text. 5. Lidocaine HCl – the non-hydrate was used unless a clear description in the label (if investigated) stated otherwise see example Zingo https://www.accessdata.fda.gov/drugsatfda_docs/label/2007/022114lbl.pdf According to USAN the name Lidocaine HCl is the monohydrate and the Ph Eur has Lidocaine HCl Monohydrate, but since there were labels as in the example above the name sent by FDA was used. 6. Estradiol, Trade name= ANGELIQ. The substance was described as the nonhydrate in the text, as well as the name and molecular weight. But the structure was represented as the hemihydrate. INN named salts If a salt is assigned an INN name, like Enoxaparin sodium then the salt is considered the active moiety and no reference substance was used for the PHPID calculation. 8 Covid vaccines approach For the COVID-19 vaccines included in the dataset, the substance was identified as the active moiety and no reference substance was used. Similar biological substances approach It has been decided to consider these substances as having the same substance identifier. Further specifications are possible according to more detailed substance information, but have not been taken into account for the PHPID production within pilot´s scope (i.e. substances defined as identical, for all trastuzumab containing products). Product name vs active ingredients data source FDA Only substances included in the active ingredient list are considered as a basis for the PHPID. Example: The product name included both Lidocaine and Dextrose (i.e. 5% Lidocaine Hydrochloride and 7.5% Dextrose Injection), but only lidocaine was included as an active ingredient in the label. Table 3: An example on differences between pharmacopeias and naming bodies regarding the naming of hydrates. Substance Pharmacopeia/naming body Name Structure and additional information Lidocaine hydrochloride, CAS 73-78-9 Ph Eur NA USAN Lidocaine hydrochloride, anhydrous As part of the (USAN name) CAS 6108-05-0 record JAN Lidocaine hydrochloride KP NA Lidocaine hydrochloride monohydrate, CAS 610805-0 Ph Eur Lidocaine hydrochloride monohydrate USAN Lidocaine hydrochloride JAN NA KP Lidocaine hydrochloride hydrate 9 There are several similar examples for other hydrates where the pharmacopeias and naming bodies don’t align. There are also inconsistencies regarding naming within the different pharmacopeias. Dose Form Core Characteristics This section provides an overview on how centralized core EDQM dose form characteristics (Release Characteristics, Intended Site, Administration Method and Basic Administrable Dose Form and their codes) and other potential characteristics can be used to describe dose forms for use in the generation of PhPID, assumption made, and challenges faced. There are several challenges regarding dose form (please see table 4) that have been faced and corresponding assumptions taken for this pilot in order to mitigate the risks of PhPID generation inconsistency. Table 4. Challenges and their description vs mitigation related to dose form evaluation process Challenge Description Challenge Mitigation Dose Form expression variations When dose forms are expressed differently within different jurisdictions as seen for some of the Covid-19 vaccines, a clear base for decision of which dose form should be used to create a global PHPID is necessary. Please find example of dose forms assigned for the Covid-19 vaccine Comirnaty within different authorities: Authority of approval Administrable dose form EMA dispersion for injection FDA suspension for injection UK solution for injection Label information variations When medicinal product dose form description is twofold in SmPC, in case the product consists of a capsule that should be opened and swallowed and is described as granules - the dose form is treated as a capsule similar to the concept for a solid dose form in a container (i.e. Alkindi sprinkle) Label information variations When medicinal product dose form description is twofold in SmPC, in case the product is described as a system and as a patch – the patch has been selected as a dose form (i.e. Ztlido) Unit of Presentation For intradermal injection system select “system” as unit of presentation according to EDQM definitions (i.e. Zingo) 10 Pattern Selection Pattern A has been selected for nasal sprays that deliver its entire contents upon activation (i.e. Valtoco) Pattern selection Pattern A has been selected for enema products, where applicator should be considered as UOP (i.e. Cortenema) To summarize, a framework for handling both dose form variations between different regions as well as label variations needs to be developed for consistent PHPID generation. Strength This section provides guidance on how to record and/or express information on strength, and reference strength of active ingredients present in medicinal products, assumptions made, and challenges faced. As outlined in ISO 11616, in order to unambiguously link the strength to the product - both strengths for the substance and reference substance (when the active ingredient is a salt/ester/pro-drug) are deemed. The identification of substance or reference substance and their corresponding strength has been verified in the SmPC. The reference strength is derived from active moieties of an active substances(s). •If in the SmPC the active substances are given as salts/esters or pro-drugs and the strength corresponds to the salts/esters or pro-drugs, the reference substance strength is being calculated as follows: oSubstance (mg)/(substance molecular weight(mg/mol)* reference substance molecular weight(mg/mol)) •If in the SmPC the active substances are given as salts/esters or pro-drugs and the strength corresponds to the active moiety, the salts/esters or pro-drugs substance strength is being calculated as follows: oReference substance (mg)/(reference substance molecular weight(mg/mol)* substance molecular weight(mg/mol)) •If in the SmPC the active substance is the active moiety. In this case the active substance is identical to reference substance and no calculation is performed. Strengths are defined according to ISO 11240 describing the use of the UCUM standard, where possible. When the strength of a medicinal product needs to undergo a transformation (for example reconstitution) for administration, the strength resulting from the transformation accordance with the regulated product information (i.e. in the SmPC) is stated. There are several challenges regarding substances (please see table 4) that have been faced and corresponding assumptions taken for this pilot in order to mitigate the risks of PhPID generation inconsistency. Table 5. Challenges and their description vs mitigation related to strength evaluation process Challenge Description Challenge Mitigation Strength expression variations In order to create consistent PHPIDs, the strength expression must be harmonized for similar products. When there is any inconsistency in the strength expression for similar products the UCUM standards and/or SI units where selected when possible: Strength given in % (i.e. Diclofenac epolamine) is expressed in mg 11 Strength given in IU (i.e. Ergocalciferol) is expressed in mg when possible Strength given as mg/g or mg/mL (i.e. locoid solution) is expressed in mg/g Unit variations for biosimilars (i.e. Hercipin Hylecta) is expressed in IU HydratesLabel information variations affecting strength expression The value of the reference strength may vary depending on if the substance is a hydrate or non-hydrate and this information is often not clear as shown in the examples in table 2. A harmonized expression of hydrates needs to be developed to ensure consistent PhPID assignments Strength for reference substances The strength for reference substances has been not expressed when the active substance is an active moiety. Patches or similar without a delivery rate Patches and systems that do not include a delivery rate are placed under the pattern A and the strength is expressed as presentation. Other patches/film, extended release that have delivery rate are placed under pattern C. Please find below some examples on strength expression for patches/systems, sourced from FDA data file: Active Ingredient Trade Name Dosage Form Strength Pattern Strength Expression Substance Strength Ref Subs Strength Diclofenac epolamine Flector Patch 1,3% A Strength by Presentation 180mg 130mg Lidocaine Lipoderm Patch 5% A Strength by Presentation 700mg N/A Lidocaine Ztlido Patch 1,8% A Strength by Presentation 36mg N/A Lidocaine hydrochloride monohydrate Zingo System (intradermal) 0,5 mg A Strength by Presentation 0,5mg 0,4mg glyceryl trinitrate Nitro-Dur Film, extended release 0.4mg/hr C Strength by Concentration 400ug/h N/A Strength expression variations When strength is expressed differently within different jurisdictions as seen for some of the Covid-19 vaccines, a clear base for decision of which strength expression should be used to create a global PHPID is necessary. 12 Please find example of expression of strength for Covid-19 vaccine from AstraZeneca within different authorities: Authority of approval Strength per dose (0.5 ml) EMA 2.5 × 108 infectious units UK 5 × 1010 viral particles Australia 5 × 1010 viral particles Concentrate´s strength In cases medicinal product is formulated as a concentrate and shall be diluted with an unknown amount of liquid (i.e. water, juice) for oral administration, it is challenging to obtain an accurate concentration after the dilution. Therefore, concentrate’s strength of the concentrate before dilution has been used for PhPID generation (i.e. Diazepam Intensol™ Oral Solution (Concentrate)). Strength interval The use cases for expression of strength interval needs to be clarified. Would it for example be applicable to products that are administered according to parameters such as age and weight or products dissolved differently depending on administration (in intramuscul ar or intravenous)? Pattern Framework According to ISO 11616, the strength used for the PhPID is primarily the presentation strength. For liquid preparations, the standard suggests that both the presentation strength (expressed as total volume of the container) and the concentration strength (per unit volume, giving a value of 1 in the denominator) should be taken into account. A separate PhPID called the product code concept is suggested to represent the strength concentration and should be mapped to the strength presentation at all applicable PhPID levels. This pilot has evaluated the above concepts for strength and their extent of integration in the PhPID. Table. 6 Strength definitions Strength definitions Strength (Presentation) When the strength of a substance is described as a qualitative term describing the discrete unit in which a Pharmaceutical Product is presented 19 LOVENOX – Pattern B and/or D Solution for Injection (prefilled syringes) Strength EDQM dose form Administrati on method Pattern B PHPID by concentration only Pattern D PHPID by presentation + concentration Strength Amount of Diluent to be Added (mL) Volume to be Withdraw n Approximate Concentratio n (mg/mL) Presentati on strength Concentrati on strength 30mg/0,3mL N/A Total 100 30mg/prefill ed syringe 100 mg/1mL Solution for injection injection XXYY AAXXYY 40mg/0,4mL N/A Total 100 40mg/prefill ed syringe 100 mg/1mL Solution for injection injection XXYY BBXXYY 60mg/0,6mL N/A Total 100 60mg/prefill ed syringe 100mg/1mL Solution for injection injection XXYY CCXXYY 120 mg/0.8 mL N/A Total 150 120 mg/syringe 150mg/1mL Solution for injection Injection WWYY DDWWYY 100mg/1ml N/A Total 100 100 mg/ampoul e 100mg/1ml Solution for injection Injection XXYY EEXXYY Lovenox is supplied as prefilled syringes and ampoules, in two different concentrations. Applying pattern B will result in aggregation into two PHPID, while pattern D result in unique PHPID for all above compositions. SOLU-CORTEFPattern B range expressed Powder for Solution (vials) Strength EDQM dose form Administrati on method Pattern B PHPID by concentration only Pattern D PHPID by presentation + concentration Strength Amount of Diluent Added (mL) Volume to be Withdraw n Approximate Concentratio n (mg/mL) Presentati on strength Concentrati on strength range 20 100mg/2mL 2mL +100 to 1000 mL Total 0,1 100mg/vial 0,1mg/mL0,99mg/mL Solution for infusion Infusion YYZZ AAYYZZ 250mg/2mL 2mL+25 0 to 1000 mL Total 0,25 250mg/vial 0,1mg/mL0,99mg/mL Solution for infusion infusion YYZZ BBYYZZ 500mg/4mL 4mL+50 0 to 1000mL Total 0,5 500mg/vial 0,1mg/mL0,99mg/mL Solution for infusion infusion YYZZ CCYYZZ 1000mg/8mL 8mL +1000m L Total 0,99 1000mg/via l 0,1mg/mL0,99mg/mL Solution for infusion infusion YYZZ DDYYZZ Solu-cortef is supplied as single-dose vials in four different concentrations. Applying pattern B will result in aggregation into one PHPID, while pattern D result in unique PHPID for all above compositions. CALDOLOR – Pattern D Solution for Injection (vials) Strength EDQM dose form Administrati on method Pattern B PHPID by concentration only Pattern D PHPID by presentation + concentration Strength Amount of Diluent to be Added (mL) Volume to be Withdraw n Approximate Concentratio n (mg/mL7) Presentati on strength Concentrati on strength 100mg/mL TBC TBC 4mg/mL 800mg/vial 4mg/mL Solution for Injection Injection YYZZ AAYYZZ 4mg/mL N/A TBC 4mg/mL 800mg/bag 4mg/mL Solution for Injection Injection YYZZ BBYYZZ The product Caldolor comes in two strength where the 100 mg/ml vial is diluted before use to the same concentration as the 4 mg/ml vial. Applying pattern D to these products would result in two unique PHPIDs while pattern B generate identical PHPID. Generally, it could be discussed how to take dilution into account for PHPID and which strength expressions should be used. Some products are diluted sequentially and/or according to weight, resulting in many different final concentrations. The effort to find information about the dilutions in the SmPCs are substantial and a lot of manual calculations needed to express strength for the different dilutions. Including dilution in the strength expression has an impact on the PHPIDs and needs to be expressed consistently for harmonized PHPID generation. 21 Advantages with pattern B is aggregation of data to get an enlarged dataset for evaluation while pattern D results in unique PHPIDs which could ensure easier identification of for example medication errors. The procedure for presentation to the chosen MD5 HASH function Order of input data The order for input into the MD5 hash function was: substance/ strength/ dose form for this pilot to keep the substance information with its relating strength (Note: strength values have been rounded to two decimal places). Example of numerical representations Table 7. Numerical representation of 3 Covid -19 vaccine using the UMC simplified substance ID and EDQM characteristics codes. The reference substance and reference strength is identical to substance and strength in this case and have been excluded from the table. Product Substance ID Strength by presentation Strength by concentration Basic Dose Form code Administration Method Form Code Intended Site Code Release Characteristics Code Covid-19 vaccine Pfizer (Comirnaty) 36667 NA 100 µg/ml 0085 0011 0022 0047 Covid-19 vaccine Moderna 35854 NA 200 µg/ml 0085 0011 0022 0047 Covid-19 vaccine AstraZeneca 35853 NA 5 × 108 infectious units 0085 0011 0022 0047 Outcome analysis The purpose of this section is to analyze the PhPIDs generated and understand if a selected set of dose form characteristics from EDQM and other potential characteristics can describe dose forms and be utilized to solve the issues with mapping between different dose form terminologies. In order to perform a thorough outcome analysis, the following methods were applied: Dose Form Mapping exercise and Findings In order to investigate how using Release, Intended Site, Administration Method and Basic Administrable Dose Form would affect harmonization of Dose Forms - EDQM Administrable Dose Form were mapped to Dose Form core characteristics. 22 Please find below some examples and related findings on cases where PhPIDs will be the same using core characteristics even if there are differences in EDQM Adminstrable Dose Forms. TABLETS PhPID EDQM Administrable Dose Form Basic Dose Form Administration Method Intended Site Release Characteristics Same PhPID Sublingual tablet Tablet Administration Oromucosal Conventional Buccal tablet Tablet Administration Oromucosal Conventional Unique PhPID Muco-adhesive buccal tablet Tablet Application Oromucosal Prolonged Unique PhPID Chewable tablet Tablet Chewing Oral Conventional Unique PhPID Chewable/dispersible tablet Tablet Chewing/ Swallowing Oral Conventional Unique PhPID Implantation tablet Tablet Implantation Parenteral Prolonged Unique PhPID Inhalation powder, tablet Tablet Inhalation Pulmonary Conventional Same PhPID Vaginal tablet Tablet Insertion Vaginal Conventional Effervescent vaginal tablet Tablet Insertion Vaginal Conventional Unique PhPID Orodispersible tablet Tablet Orodispersion Oral Conventional Same PhPID Tablet Tablet Swallowing Oral Conventional Coated tablet Tablet Swallowing Oral Conventional Film-coated tablet Tablet Swallowing Oral Conventional Tablet with sensor Tablet Swallowing Oral Conventional Unique PhPID Gastro-resistant tablet Tablet Swallowing Oral Delayed Unique PhPID Prolonged-release tablet Tablet Swallowing Oral Prolonged Unique PhPID Modified-release tablet Tablet Swallowing Oral Modified Finding: SmPC does not always provide the details on administrable dose form. EDQM Administrable Dose Form such as tablet, coated tablet, film-coated tablet, and tablet with sensor will have same PhPID cased on chosen characteristics. EDQM Administrable Dose Form mapping for sublingual tablet and buccal tablet will also result in same PhPID. 23 From pharmacovigilance perspective generating same PhPIDs for the above EDQM terms would be beneficial. From prescription perspective this approach would represent a potential risk for the patient receiving inappropriate pharmaceutical form. CAPSULES PhPID EDQM Administrable Dose Form Basic Dose Form Administration Method Intended Site Release Characteristics Unique PhPID Chewable capsule, soft Capsule Chewing Oral Conventional Unique PhPID Oromucosal capsule Capsule Chewing Oromucosal Conventional Unique PhPID Inhalation vapour, capsule Capsule Inhalation Pulmonary Conventional Same PhPID Vaginal capsule, hard Capsule Insertion Vaginal Conventional Vaginal capsule, soft Capsule Insertion Vaginal Conventional Unique PhPID Rectal capsule Capsule Insertion Rectal Conventional Same PhPID Capsule, hard Capsule Swallowing Oral Conventional Capsule, soft Capsule Swallowing Oral Conventional Same PhPID Gastro-resistant capsule, hard Capsule Swallowing Oral Delayed Gastro-resistant capsule, soft Capsule Swallowing Oral Delayed Same PhPID Prolonged-release capsule, hard Capsule Swallowing Oral Prolonged Prolonged-release capsule, soft Capsule Swallowing Oral Prolonged Same PhPID Modified-release capsule, hard Capsule Swallowing Oral Modified Modified-release capsule, soft Capsule Swallowing Oral Modified Findings: SmPC does not always provide the details on if capsule is hard or soft. From pharmacovigilance perspective generating same PhPIDs for the above EDQM Administrable Dose Form would be beneficial grouping EDQM Administrable Dose Form should group “hard” and “soft” capsules together. From prescription perspective this approach would represent a potential risk for the patient receiving inappropriate pharmaceutical form. 24 SOLUTION – INFUSION/INJECTION PhPID Administrable Dose Form Basic Dose Form Administration Method Intended Site Release Characteristics Unique PhPID Dispersion for injection/infusion Dispersion Infusion/injection Parenteral Conventional Unique PhPID Dispersion for infusion Dispersion Infusion Parenteral Conventional Unique PhPID Emulsion for infusion Emulsion Infusion Parenteral Conventional Unique PhPID Emulsion for injection/infusion Emulsion Infusion/injection Parenteral Conventional Unique PhPID Solution for infusion Solution Infusion Parenteral Conventional Unique PhPID Solution for injection/infusion Solution Infusion/injection Parenteral Conventional Unique PhPID Solution for injection Solution Injection Parenteral Conventional Unique PhPID Suspension for injection Suspension Injection Parenteral Conventional Unique PhPID Emulsion for injection Emulsion Injection Parenteral Conventional Unique PhPID Gel for injection Gel Injection Parenteral Conventional Unique PhPID Prolonged-release suspension for injection Suspension Injection Parenteral Prolonged Unique PhPID Intraperitoneal solution Solution Injection Intraperitoneal Conventional Unique PhPID Prolonged-release solution for injection Solution Injection Parenteral Prolonged Unique PhPID Prolonged-release dispersion for injection Dispersion Injection Parenteral Prolonged Unique PhPID Dispersion for injection Dispersion Injection Parenteral Conventional All dose forms result in unique PhPIDs. 25 SOLUTION – ORAL SUSPENSION/ORAL SOLUTION PhPID EDQM Administrable Dose Form Basic Dose Form Administration Method Intended Site Release Characteristics Unique PhPID Oral suspension Suspension Swallowing Oral Conventional Unique PhPID Gastro-resistant oral suspension Suspension Swallowing Oral Delayed Unique PhPID Prolonged-release oral suspension Suspension Swallowing Oral Prolonged Unique PhPID Modified-release oral suspension Suspension Swallowing Oral Modified Unique PhPID Oral drops, solution Solution Instillation/ Swallowing Oral Conventional Unique PhPID Oral solution Solution Swallowing Oral Conventional Unique PhPID Oral/rectal solution Solution Administration/ Swallowing Oral/Rectal Conventional Unique PhPID Oral solution/concentrate for nebuliser solution Solution Inhalation/ Swallowing Oral/Pulmonary Conventional Finding: In some cases, delivery devices represent the only way to understand if a product is a solution or a suspension according to EDQM. Detailed information on delivery device (syringe for suspension/spoon for solution) is not always described explicitly in SmPC. Clarification in EDQM could potentially improve understanding and assignment of these dose forms. CREAMS vs OINTMENTS PhPID EDQM Pharmaceutical Administrable Dose Form Basic Dose Form Administration Method Intended Site Release Characteristics Unique PhPID Ear ointment Ointment Application Auricular Conventional Unique PhPID Ear/eye ointment Ointment Application Auricular/Ocular Conventional Unique PhPID Oromucosal cream Cream Application Oromucosal Conventional Unique PhPID Cream Cream Application Cutaneous/Transdermal Conventional Unique PhPID Eye cream Cream Application Ocular Conventional Unique PhPID Ear cream Cream Application Auricular Conventional 26 Unique PhPID Nasal cream Cream Application Nasal Conventional Unique PhPID Vaginal cream Cream Application Vaginal Conventional Unique PhPID Rectal cream Cream Application Rectal Conventional Same PhPID Ointment Ointment Application Cutaneous/Transdermal Conventional Transdermal ointment Ointment Application Cutaneous/Transdermal Conventional Cutaneous/nasal ointment Ointment Application Cutaneous/Transdermal Conventional Unique PhPID Cutaneous spray, ointment Ointment Spraying Cutaneous/Transdermal Conventional Unique PhPID Urethral ointment Ointment Administration Intravesical/Urethral Conventional Unique PhPID Nasal ointment Ointment Application Nasal Conventional Unique PhPID Eye ointment Ointment Application Ocular Conventional Unique PhPID Oromucosal ointment Ointment Application Oromucosal Conventional Unique PhPID Inhalation vapour, ointment Ointment Inhalation Pulmonary Conventional Unique PhPID Rectal ointment Ointment Application Rectal Conventional Unique PhPID Vaginal ointment Ointment Application Vaginal Conventional Finding: SmPC does not always provide the details on administrable dose form. EDQM Administrable Dose Form mapping for Ointment, Transdermal ointment and Cutaneous/nasal ointment resulting in same PhPID. From pharmacovigilance perspective generating same PhPIDs for the above EDQM terms would be beneficial. From prescription perspective this approach would represent a potential risk for the patient receiving inappropriate pharmaceutical form PATCH PhPID EDQM Pharmaceutical Administrable Dose Form Basic Dose Form Administration Method Intended Site Release Characteristics Unique PhPID Cutaneous patch Patch Application Cutaneous/Transdermal Conventional Unique PhPID Transdermal patch Patch Application Cutaneous/Transdermal Prolonged Unique PhPID Oromucosal patch Patch Application Oromucosal Prolonged 27 Finding: Though EDQM definitions are quite straightforward on both Cutaneous and Transdermal - what is rationale behind to have dual intended site? The curent dwefinitions in EDQM: Cutaneous patch - Flexible single-dose preparation intended to be applied to the unbroken skin to obtain a local effect by penetration of the active substance(s) into the skin. Transdermal patch - Flexible single-dose preparation intended to be applied to the unbroken skin to obtain a systemic delivery over an extended period of time. Transdermal patches consist of a backing sheet supporting a reservoir or a matrix containing the active substance(s) and on the top a pressure-sensitive adhesive, which assures the adhesion of the preparation to the skin. The backing sheet is impermeable to the active substance(s) and normally impermeable to water. In reservoir systems the active substance may be dissolved or dispersed in a semi-solid basis or in a solid polymer matrix, which is separated from the skin by a rate-controlling membrane. The pressuresensitive adhesive may, in this case, be applied to some or all parts of the membrane, or only around the border of the membrane and the backing sheet. Matrix systems contain the active substance in a solid or semi-solid matrix, the properties of which control the diffusion pattern to the skin. The matrix system may also be a solution or dispersion of the active substance in the pressure-sensitive adhesive. The releasing surface of the patch is covered by a protective liner to be removed before applying the patch to the skin. OCULAR FORMULATIONS PhPID EDQM Pharmaceutical Administrable Dose Form Basic Dose Form Administration Method Intended Site Release Characteristics Unique PhPID Eye cream Cream Application Ocular Conventional Unique PhPID Eye gel Gel Application Ocular Conventional Unique PhPID Eye ointment Ointment Application Ocular Conventional Unique PhPID Eye drops, solution Solution Instillation/ Swallowing Ocular Conventional Unique PhPID Eye drops, emulsion Emulsion Instillation/ Swallowing Ocular Conventional Unique PhPID Eye drops, suspension Suspension Instillation/ Swallowing Ocular Conventional Unique PhPID Eye drops, prolonged-release Drops (unspecified) Instillation Ocular Prolonged Unique PhPID Eye lotion Solution Bathing Ocular Conventional Finding: SmPC does not always provide the details on administrable dose form if it is a solution/emulsion/suspension, which can cause PhPID inconsistency. 28 COVID-19 VACCINE DOSE FORMS Finding: COVID-19 vaccines dose form assignment varies within different authorities. EDQM definition of the different dose forms used could be improved to simplify a harmonized assignment. Furthermore, when dose forms are expressed differently within different jurisdictions, a clear base for decision of which dose form should be used to create a global PHPID. Please find example of dose forms assigned for the Covid-19 vaccine Comirnaty within different authorities: Authority of Approval Administrable Dose Form EDQM Definition EMA dispersion for injection Liquid sterile preparation consisting of two or more phases of which at least one is dispersed in the liquid phase, intended for administration by injection. To be used only when emulsion for injection is not appropriate. Solid sus pension preparations are excluded. FDA suspension for injection Liquid sterile single-dose or multidose preparation consisting of a suspension intended for administration by injection UK solution for injection Liquid sterile single-dose or multidose preparation consisting of a solution intended for administration by injection. Conclusion & Recommendations This section outlines the ‘Recommendations/Improvement Suggestions’ coming from the pilot results as well as the high-level ‘Action Plan’ activities proposed for process/system owners and/or key stakeholders. Pilot results confirmed that centralized core EDQM dose form characteristics (and their codes), Release Characteristics, Intended Site, Administration Method and Basic Administrable Dose Form, can be used as input in the generation of the global PhPID and solve the issues with mapping between different dose form terminologies. Using centralized EDQM core characteristics would harmonize the levels of granularity between regions and significantly increase the quality of one-to-one mapping between a regional terminology and a centrally controlled vocabulary. This approach would ensure a consistent PhPID construction and thereby allow gathering information into one global data source. Having a global data repository would facilitate a much faster and more efficient detection of drug safety signals, and substantially increase the probability of detecting rare adverse drug reactions (ADRs), proving basis for holistic pharmacovigilance support and enabling exchange of ICSR information between NCs/regulators, industry and other stakeholders globally. However, the pilot results have demonstrated also that there are challenges to overcome and there is no “perfect” solution. The resulting characteristics set is close to be definitional, except for a couple of PDF These exceptions must be worked on (based on the defintions) but could also disappear by obsoleting the original PDFs in EDQM, when the importance of the difference is debatable. For example is the difference between « Capsule, hard » and « Capsule, soft » relevant ? A first analysis did not find necessary difference in the definitions. No requirement to change the definitions is made here, it’s the responsibility of EDQM and these characteristics are used for important acts. The appearance of new non-definitional characteristics sets for PDFs could be avoided by screening them before adding them to STD, but this is also not a requirement. Creating a dose form ontology to add a level of granularity The global Dose Form terminology must be tailored to link to different existing Dose Form terminologies, some more high level, some more granular. The main steps consisted in a spreadsheet to : 1. Making groups of identical 4 characteristics (TRAC, RCA, AMEC, ISIC_SPLIT) 2. Classifying the groups alphabetically by ISIC_SPLIT 3. Putting the PDFs with « no transformation » first Then to concatenate or split the groups based on 2 main criterion : clinical relevance and impact on the business rules for the determination of the strength. These business rules will be determined by the WHO and the FDA, based on specifict requirements provided by the project. Examples : the Auricular group consists in all the PDFs that differ for some characteristics but not in a way that is clinically relevant and can hence be concatenated. One group had identical characteristics but was splitted nonetheless : Oral drops and Oral liquids. The dose form groups created were then named (e.g. Cutaneous dose form). ADF BDF SOM A D F _B D F A D F _S O M TRAC RCA IS IC_SPLIT AMEC SYS Count UC Doubles> 1 Sum Doubles 1x x x x xx x x x x 420 48 2x x x x x x x xx402 12 26 3xx x x x x x x 381 21 47 4xx x x x x x 380 21 48 5xx x x x x 378 22 50 6xx x x x 377 22 51 7 x xx x x x x x x 350 33 78 8 x xx x x x x x 340 37 88 9 x x x x x x x 306 51 122 10 x x x x x x 293 56 135 11 x x x x x 128 78 300 12 x x x x 113 79 315 UMC/FDA+ sys x x x x x 197 77 231 UMC/FDA x x x x 179 82 249 Figure 4: Evaluation of the potential of EDQM characteristics to be definitional In the resulting ontology the dose form groups are linked to the Intended Sites (e.g. Cutaneous) through the property hasDoseForm (with the corresponding property hasISI leading back to the Intended Site). The granular PDFs are all subclasses of the grouped Dose Forms. Figure 5: Naming the dose form groups Figure 6: Structure of the simple dose form ontology The ontology is available on WebProtégé, and is created automatically in OWL/XML from the source spreadsheet, based on a custom program in Java, OpenJena and OpenCSV. Remarks: • Some Oral and Rectal combinations still have to be solved. The combination should be kept as much as possible on the higher level. • The « to be transformed dose form » ontological class was created as a temporary class for the study only. It contains two interesting properties : ◦ hasADF links to the ADF obtained after transformation ◦ a label containing in a text form the course of the transformation and the resulting ADF. • Four isolated ADF arrive just below owl:thing, coming from faulty ADF not corresponding to existing PDF In addition, this simple ontology has been used by Natalie J Karapetian to link the RxNorm to the dose form groups, allowing for further analysis of its usefulness. Conclusion: The EDQM STD is the best candidate for a global terminology of dose forms. Analysing the structure and content of the standard terms by the means of spreadsheets and ontologies allowed to discover and propose precise, surgical modifications to the resource to improve its capacity to be a central resource for the characterisation of Pharmaceutical dose forms. An additional step towards this was made with the creation of a simple and small ontology. By adding a middle layer of granularity – dose form groups – to the EDQM STD which is more granular than RxNorm for example, the alignement of dose forms between EDQM and other descriptions of dose forms used widely like SNOMED-CT can be improved. The work reported here will be continued in collaboration with UNICOM partners. Figure 7: Visualising a PDF in the ontology (in WebProtégé) Document version D8.1_ANNEX 5_Comparison EDQM/RxNORM Authors Natalie Karapetian Yuri Quintana Robert Vander Stichele Changes Contribution of Harvard Medical School to UNICOM TITLE: Evaluating the interoperability 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. September 15, 2021 Contents INTRODUCTION ........................................................................................................................................... 2 METHODOLOGY .......................................................................................................................................... 3 RESULTS ...................................................................................................................................................... 3 RXNORM Dose Representation ............................................................................................................... 3 EDQM Dose Representation .................................................................................................................... 4 MAPPING RXNORM DOSE FORMS TO EDQM DESCRIPTORS .......................................................... 5 PROPOSED INTEGRATION OF RXNORM AND EDQM ........................................................................ 6 DISCUSSION............................................................................................................................................ 7 CONCLUSION .............................................................................................................................................. 8 REFERENCES .............................................................................................................................................. 9 INTRODUCTION Patients need to be able to safely refill prescriptions when in another country due to travel or relocation. The conventions for identifying therapeutic drugs vary greatly between common drug classification systems in different parts of the world, making it difficult for patients and pharmacists to be sure they are getting the correct drug and dose. The lack of a comprehensive mapping database creates significant obstacles to safe transnational medication prescribing practices and conducting large-scale international clinical trials. The increasing volume of medications on the market internationally, along with the expansion in global medical tourism1 and immigration, escalates the need for global infrastructures for drug interaction and clinical decision support to reduce medication errors and protect patient safety. A directive issued in the European Union in 2011 mandated cross-national recognition of prescriptions and provided measures to facilitate health professionals verification prescriptions issued by other member states2. While this directive requires that prescriptions be written using a “common name2,” there are significant variations in medication names and formulations between countries. A global mapping system of pharmaceutical equivalents does not currently exist. Accordingly, a survey of 3,307 German citizens who have prescribed medication while in another EU member state revealed that 12% experienced medication problems related to being prescribed medications different from their normal prescriptions, and 80% reported a lack of health information exchange between countries.3 Standardizing the naming system of medications would facilitate developing an international drug information database supportive of clinical decision support aiding clinicians in efficiently determining medication equivalency and interactions. Creating a drug database capable of integrating drug information across countries will support greater interoperability of prescription information, facilitate improved clinical decision support, improve the safety of cross-border medication dispensing, and expand the accessibility of health data for research efforts. To establish this global drug database, a standardized drug identification system must be established. The Identification of Medicinal Products (IDMP) standards was created in 2012 by the International Organization for Standards (ISO) to outline key characteristics of medicinal product classification systems for international harmonization4. One of the key components of IDMP is ISO 11616, which defines the elements required to construct universal Pharmaceutical Product Identifiers (PhPID). PhPIDs are internationally recognizable identifiers for each drug product. They are composed of three main attributes: the active pharmaceutical substance, the dosage form, and the strength,5 allowing pharmaceutically equivalent drug products to be recognized as synonymous regardless of regional differences in branding and packaging. To enhance prescribing interoperability across Europe and better adhere to IDMP standards, the European Union commissioned a project called Up-Scaling the Global Univocal Identification of Medicines (UNICOM) to establish a unified drug database for all medications prescribed across the EU and internationally6. As the representation of medication dosage forms varies greatly between regional terminologies, a significant barrier to the generation of universal PhPIDs has been the absence of mappings between dosage form representations. The European Directorate for the Quality of Medicines and Healthcare (EDQM) has created a controlled vocabulary relating dose form descriptions to key dosage form characteristics7, upon which the UNICOM project has structured their dosage form descriptions. To facilitate harmonization between European and United States drug databases, the World Health Organization, United States Food and Drug Administration (FDA), Uppsala Monitoring Centre, and UNICOM completed a pilot project seeking to align dosage form descriptions between the FDA’s Terminology for Structured Product Labeling (SPL) and EDQM8. The pilot results were discussed in a June 11th, 2021 webcast and identified difficulties in harmonization between FDA SPL, EDQM, and ISO standards, indicating the necessity for ISO standard revision8. The drug terminology system operated by the United States National Library of Medicine is known as RxNorm and is of great significance within the United States and abroad. RxNorm terminology is the basis for the U.S. Department of Veterans Affairs National Drug File-Reference Terminology (NDF-RT9), for the Drug Ontology (DrOn)10, and for the Observational and Medical Outcomes Partnerships (OMOP) opensource common data model, which is used ubiquitously across the United States and Europe to conduct multinational drug studies11. Interoperability between the Canadian drug ontology OCRx and RxNorm has also been demonstrated, further contributing to the international significance of RxNorm serving as a standard drug terminology. While studies have been conducted to evaluate interoperability between the FDA SPL and SNOMED to EDQM8, there have been no studies to our knowledge that have evaluated interoperability of dosage form descriptions between RxNorm and EDQM. Given the tremendous national and international significance of RxNorm, the objective of this paper is to assess a mapping of RxNorm dosage forms to EDQM-based descriptors. METHODOLOGY A description of the dosage form representations in RxNorm and EDQM will be provided, followed by a proposal of how the RxNorm and EDQM data models could be used to generate universal pharmaceutical product identifiers (PhPID). A description of the derivation of EDQM characteristics to the RxNorm dosage forms will be provided. Using the attributed EDQM characteristics to RxNorm dose forms, each RxNorm Dose Form will be a collection of EDQM dose Forms with an identical (or) similar combination of characteristics, based on a simple ontology, developed in the UNICOM Project. This ontology has two levels: a first level based on the ISI characteristic; and a second level based on functional grouping of identical combinations of characteristics, taking into account differences in representing strength. This ontology was operationalized in Web Protégé. In one application, all the EDQM dose forms were integrated into the ontology, and in another application, all the RxNorm Dose forms. RESULTS RXNORM Dose Representation RxNorm is a normalized naming system established by the United States National Library of medicine for branded and generic pharmaceutical products. RxNorm was created to support interoperability between medical-related terminologies and related knowledge bases across medical applications used in the United States12. RxNorm standardizes representation of pharmaceutical ingredients, strength, dose form, and brand name information. For each unique drug product, an array of codes is generated to represent different identifiers of that drug product. Codes of interest to this investigation include the Ingredient (IN) code derived from the United States Adopted Name (USAN), the Dose Form (DF) code selected from a controlled list of dose forms provided by RxNorm, and the Semantic Clinical Drug (SCD) code, which aggregates descriptions of the product’s ingredient, strength, and dose form. RxNorm dosage forms are aggregated into Dose Form Groups (DFG) based on the route of administration, release characteristics or product type. Each dose form is attributed to at least one Dose Form Group, with dose forms often belonging to several Dose Form Groups. For this reason, RxNorm does not provide an ontology for classifying dosage forms but instead a list of defined dosage forms contained in overlapping groups. A sample from the RxNorm representation of dosage forms for sublingual tablets is as follows, illustrating the redundancies of dosage form representations between Dosage Form Groups: Oral Product (Dosage Form Group) Sublingual Tablet Capsule Tablet Pill Sublingual Tablet Buccal Tablet Chewable Tablet Sublingual Product Sublingual Tablet Sublingual Film EDQM Dose Representation EDQM maintains a set of controlled vocabularies to describe six key characteristics of pharmaceutical dosage forms7. These six characteristics include State of Matter, Basic Dose Form, Transformation (TRA), Release Characteristics (RCA), Intended Site (ISI), and Administration Method (AME). Lists of defined terms that characterize dosage forms in each of these six areas are maintained by EDQM. The Basic Dose Form and the State of Matter refer to the drug’s form (such as a cream, tablet, implant) and the associated physical state of matter (solid, semi-solid, liquid, or gas). Transformation refers to whether the product requires alteration before administration, such as through dilution or reconstitution. Release Characteristic refers to any alteration of the drug release timing (such as prolonged or delayed-release), Intended Site refers to the anatomical site of drug administration (such as oral, parenteral, or ocular), and Administration Method refers to the method of drug administration (such as via swallowing, chewing, or inhalation). A sample from the EDQM representation of dosage forms for sublingual tablets is as follows: Oral (Site of Administration) Sublingual Tablet State of Matter: Solid Basic Dose Form: Tablet Transformation: No Transformation Release Characteristics: Conventional Intended Site: Oromucosal Administration Method: Orodispersion MAPPING RXNORM DOSE FORMS TO EDQM DESCRIPTORS A list of all dose forms recognized by RxNorm was downloaded from Appendix 2 of the RxNorm Technical Documentation (version reviewed July 6, 2020) located on the United States National Library of Medicine website (https://www.nlm.nih.gov). This list of dose forms was uploaded into a Microsoft Excel spreadsheet, and columns were created representing the six characteristics to describe pharmaceutical dose forms under EDQM: state of matter, basic dose form, release characteristics, intended site, transformation, and administration method (Figure 1). Figure 1:Mapping of RxNorm Dosage Forms to EDQM dosage form descriptors. EDQM definitions for drug basic form, transformation, state of matter, release characteristics, intended site, and administration method were used to assign EDQM terms to each RxN orm dosage f The definition for each RxNorm dose form provided by Appendix 2 of the RxNorm Technical Documentation was used to assign descriptive EDQM characteristics for each dosage form. The EDQM Standard Terms and Internal Controlled Vocabularies for Pharmaceutical Dose Forms (Version 1.2.0) were consulted to manually fill the six characteristics for each RxNorm dosage form based on the provided definitions for each characteristic. To accommodate for the lack of delineation between manufactured dose forms and administrable dose forms, columns were made to describe the state of matter and basic dose form of the drug as it was supplied by the manufacturer (known as the Manufactured Dose Form) and after any indicated transformation (known as the Administrable Dose Form). For dose forms not requiring transformation, the Manufactured Dose Form and the Administrable Dose Form state of matter and basic dose forms were equivalent. The release characteristics, intended site, and administration method were subsequently assigned in reference to the administrable dose form. PROPOSED INTEGRATION OF RXNORM AND EDQM We propose applying the EDQM framework of drug dosage form descriptions to RxNorm dosage forms. For each drug product in RxNorm, the DF code can be mapped to RCA, ISI, TRA, and AME codes. The mapping of RxNorm dosage forms to these characteristics is described in the previous section. Additionally, the IN and SCD codes can be mapped to globally accepted ingredient and strength codes. Accordingly, the decision of which ingredient and strength codes should be used universally for this purpose is outside the scope of this paper. The combination of ingredient, strength, and four dosage form descriptor codes can then be used to generate a universal pharmaceutical product identifier (PhPID) to identify pharmaceutical products independent of regional naming conventions (Figure 2). Figure 3: The proposed common data model for generating pharmaceutical product identifiers from EDQM 2 and RxNorm. The mapping algorithm between the RxNorm DF code and the release characteristics (RCA), intended site (ISI), transformation (TRA), and administration method (AME) are defined in this paper. COMPARING RxNORM AND EDQM WITH A COMMON ONTOLOGY An ontology was created based on the EDQM site of administration, with the site of administration as the primary grouping mechanism. Each RxNorm dosage form was fitted into this EDQM-based dosage form ontology, with preservation of the transformation, release, administration method, and intended site of administration codes previously derived. This ontology was uploaded to the online ontology-builder Webprotege and is accessible here: https://webprotege.stanford.edu/#projects/34af25bd-e27b-4b36-bbc14498b0706971/edit/Classes?selection=Class(%3Chttp://unicomproject.eu/doseforms.owl%23DoseForm%3E). give a screenshot splitscreen of the ontology opened on the dose form group in both terminologies. DISCUSSION The mapping of RxNorm dose forms to EDQM descriptors revealed several issues requiring attention before reliable mapping is established. There is a significant disparity in granularity of dosage form descriptions between RxNorm and EDQM, with RxNorm recognizing 179 dose forms and EDQM recognizing 428 dose forms (not including the veterinary dose forms). While there is considerable overlap between the dosage forms in both systems, some dosage forms are exclusive to RxNorm (such as urethral suppository) and many exclusive to EDQM. RxNorm Dose Form Groups vary significantly in the rationale behind granularity and grouping, resulting in Dose Forms often belonging to several Dose Form Groups. Due to the inconsistent granularity of RxNorm dosage form groups and redundant dose form Figure 4: RxNorm anchored in the dose form groups Figure 5: EDQM anchored in the dose form groups 5 The substance should be present in at least one authorized medicinal product in at least one jurisdiction or an internationally recognized experimental product. For each substance at the chose granularity level, the identification numbers in different systems should be identified and listed in a mapping file (WHO Drug Dictionary, INN Modified, SNOMED, UNII, CAS, EU-SRS, SPOR, US-SRS, …). The identification number of the substance to be submitted to the HASH function must be chosen and the order of the substances if >1. An investigation of the requirements for the identification number will be conducted and a proposal will be made by WHO UMC. There are several issues regarding substances that needs clarification: • Should the reference substance only be expressed when the active ingredient is identified as a salt or ester? • For some substances (e.g. biologicals) further specifications are possible, but do not necessarily need to be taken into account for the PHPID production. • If a substance belongs to a collection of substances with the same therapeutic moiety, all relevant salts and esters in that collection should also be defined and submitted to the above procedures. It should also be established how to refer to INN in for the salt and ester variations1. • In Adverse Event Reports it is possible that brand names or unspecified substances are mentioned. Hence, it may be necessary to include in the collection of possible specific substances class of “substance unspecified”. It is hoped that the need for this will diminish, as Companies and Marketing authorization Authorities adopt IDMP in the future. These unspecified substances should however not be used to generate PhPID calculations on level 2-4. Dosage Form EDQM is the system of reference for dosage form that follow ISO 11239, is recognized by EMA, and will therefore be used to express dosage forms in this project, also for numerical presentation. It is proposed to choose the most granular administrable dose form for PHPID production, not the basic dose form. For those dosage forms that do not undergo transformation processes from distribution to administration phase the pharmaceutical dose form will be used if an administrable dose form is not available within EDQM. Some issues related to dose forms needs further elaboration: • When dose forms are expressed differently within different jurisdictions as seen for some of the Covid-19 vaccines in table 1, a clear base for decision of which dose form should be used to create a global PHPID is necessary. Authority of approval Administrable dose form EMA dispersion for injection FDA suspension for injection UK solution for injection Table 1. Example of dose forms assigned for the Covid-19 vaccine Comirnaty within different authorities 1 An INN is usually designated for the active part of the molecule only, to avoid the multiplication of entries in cases where several salts, esters, etc. are actually used. In such cases, the user of the INN (pharmacopoeia commissions, regulatory bodies, pharmaceutical manufacturers) has to create a modified INN (INNM) himself; mepyramine maleate (a salt of mepyramine with maleic acid) is an example of an INNM. When the creation of an INNM would require the use of a long or inconvenient name for the radical part of the INNM, the INN programme will select a short name for such a radical (for example, mesilate for methanesulfonate). 6 • If the administrable dose form can be expressed in differently in ml, but the posology in drops, the relationship between these two forms of expression should be made explicit (eg. 20 drops per ml). • For solid oral forms, galenic features of controlled release are important sometimes be specified (rapid release, prolonged release BID or OD), but may share the same strength. To summarize, the ontology of dosage forms in EDQM should be carefully studied to understand the implications with regard to for example local or systemic action, and to traditional gastro-intestinal absorption and avoidance of the liver circulation, also in relation to the characteristics of EDQM. The investigation should lead to suggestions for improvements to EDQM. The new alternative proposal that is under investigation in ISO workgroup 6, using the characteristics of the PDF instead of the described system that maps to EDQM, is not in scope of this project but will be further investigated within ISO TC 215 workingroup 6. Strength For the unit of strength, the UCUM standard should be used and the numerical values of UCUM should be expressed for the following elements of strength has associated with the substance or specified substance: • The presentation strength, also called basis of strength, is the strength of a substance described as a qualitative term describing the discrete unit in which a Pharmaceutical Product is presented, such as weight per tablet. • The concentration strength is the strength of a substance expressed as the amount of substance per unit of measurement, such as millilitre or gram • The reference strength shall be expressed based on anhydrous free acid, anhydrous free base or a substance created to express activity, ie active moiety Rules and roles should be made explicit to express the above concepts for strength and their extent of integration in the PHPID. If the reference substance is identified as an active moiety, will the reference strength be required? If there is no exact strength related to the substance, the strength interval will be expressed as RTO<PQ,PQ>data type according to ISO11616. There are several issues regarding strength that needs to be investigated: • The denominator should be made explicit and can be of a different nature for oral solid forms and liquids (100mg per tablet for oral forms, 250mg/ml for syrups, 1000mg per sack for soluble powder). • For liquids, the strength shall be expressed per total volume per container and strength (concentration) per unit volume. The strength concentration per unit volume shall be calculated from the strength per total volume of the container. There is a difference between weight per volume and concentration (e.g 5mg/2ml vial) or 2.5mg/ml in a 2ml vial). (in case of concentration the value of the denominator is an implicit 1)2. • For injectables, the difference between unidose vials and multidose vials has numerous implications for expression of strength, pack size and units of administration in the signatura. • For a patch, strength shall be expressed as per time unit or per each patch according to approval. If not rate, use quantity per each/contained by each. • EDQM has ‘transdermal patch’ as dose form for a patch with systemic action and ‘cutaneous patch’ is meant for a local effect. The difference between the site of action (systemic or local) should be expressed as such, not ‘hidden’ in the way of expression of strength. 2For PhPID and liquid preparations, the strength shall be expressed per total volume per container and strength (concentration) per unit volume, at every instance of PhPID level 2 and 4. 7 • For topical products (creams, ointments), the establishment of the unit of administration in the posology is not as self-evident as it is in solid oral dosage forms. To determine the nature the strength expressions and the above-described issues for different types of product, the EU IG described patterns for expressions of Pharmaceutical Product, will be evaluated3. Using the EU IG patters would generate data for PHPID generation for Covid-19 vaccines according to table 2. Product Substance Strength by presentation Strength by concentration Administrable dose form Covid-19 vaccine Pfizer (Comirnaty) Tozinameran NA 100 µg/ml Dispersion for injection Covid-19 vaccine Moderna COVID-19 vaccine mRNA (mRNA 1273) NA 200 µg/ml Dispersion for injection Covid-19 vaccine AstraZeneca COVID-19 vaccine NRVV Ad (ChAdOx1 nCoV19) NA 2.5 × 108 infectious units Suspension for injection Table 2. Using pattern 3A from the EU IG, strength presentation would not be applicable for PhPID generation, only strength concentration. There are also some discrepancies seen for which units are used to express strength within different jurisdictions for the Covid-19 vaccines, see table 3. Like the dose forms, a clear base for decision of expression of strength for Global PhPIDs are necessary. Authority of approval Strength per dose (0.5 ml) EMA 2.5 × 108 infectious units UK 5 × 1010 viral particles Australia 5 × 1010 viral particles Table 3. Expression of strength for COvid-19 vaccine from AstraZeneca. The procedures of choice for the numerical representations Substance A choice needs to be made between the available candidates; EUTCT, UNII, UMC Substance ID etc. This choice should be acceptable at an international and global level and should fulfil the requirements in the ISO/IEC 15459, specifying the common rules applicable for unique identification. Any chosen system must be able to reflect the chosen level of granularity for substances. Dosage form The EDQM number will be used for this pilot. 3 https://www.ema.europa.eu/en/human-regulatory/research-development/data-medicines-iso-idmpstandards/spor-master-data/substance-product-data-management-services#eu-idmp-implementation-guide--- version-2.0-section 8 Strength Value+UCUM number should be used. Since the UCUM nomenclature do not provide a list with fixed values, the UMC value list will be used for this project. There should also be a consensus around how strengths are expressed,for example 0.1 g or 100 mg. Example of numerical representations Product Substance ID Strength by presentation Strength by concentration Administrable dose form Covid-19 vaccine Pfizer (Comirnaty) 36667 NA 100 µg/ml 50077000 Covid-19 vaccine Moderna 35854 NA 200 µg/ml 50077000 Covid-19 vaccine AstraZeneca 35853 NA 2.5 × 108 infectious units 11202000 Table 4. Numerical representation of 3 Covid -19 vaccine using the UMC simplified substance ID and EDQM dose form code. The reference substance and reference strength is identical to substance and strength in this case and have been excluded from the table. The procedure for presentation to the chosen MD5 HASH function Order of basic characteristics in single products There is a choice to be made between two possible sequences, when substance comes first: • Substance / dose form / strength • Substance / strength / dose form UMC have used the order of substance/ strength/ dose form for this pilot to keep the substance information with its relating strength but have not made any further investigations in relation to other conceptual systems. Combination products (combination of substances) A proposal needs to be made for the order of substances, in case of combination products and how their respective strength will be presented, after each substance or gathered consecutively? There is also a decision to be made for other cases: • Complex medicinal products with many substances (e.g. multivitamins). How and when is a PhPID useful and possible to create in a harmonized way (since countries sometimes define the number of active ingredients differently for these product)? • The policy for outdated FDC (fixed dose combinations) e.g. combinations of antipyretics, cough products and antibiotics. • What will the policy be for border line cases (adjuvants4) • Are complex parenteral hospital products for infusion in scope? Multiple products packaged as a kit with intent to being administered as one medical product shall be assigned one overarching PhPID according to ISO. 4ISO 11616: If an adjuvant is applicable for eg vaccine, the adjuvant term and term ID shall be displyed with the active substance (s) and specified substance(s) terms for the product on all applicable PhPID levels. This association shall be made by directly associating the assigned PhPID to MPID and PCID. 9 Hybrid products Both combinations of pharmaceutical products5 and combinations with medical devices or diagnostics6 needs to be considered. The hash function A few different hash functions were investigated: • MD5 (128 bit) o pros: fast o cons: considered broken, known to have hash collision • SHA2 (256 bit and 512 bit) o pros: considered secure (have vulnerabilities but are not considered to be too serious) o cons: slower then MD5 • SHA3 o It is yet to gain widespread support and implementation and will be unlikely to do so until significant flaws in SHA2 are found that necessitate an update. The MD5 algorithm is a much faster hashing algorithm then SHA2 but it is not cryptographically secure. Its main application is data integrity verification. It is possible to force a collision between two MD5 hashes if you control the input of both hashes. The procedure for making the resulting PHPID publicly available The procedures for producing the PHPID (the current document) should be made publicly available as a versioned, living document. There should be a publicly available Linked Open Data PHPID database with: • the PHPIDs • the basic concepts and their numerical representations (enabling checks of Hash functions) Additional linked data may be beneficial (for example other identifiers (e.g. INN, UNII, SPOR, EU-SRS, CAS, SNOMED and the link to ATC/ROA/DDD methodology) Other instances could govern links to standardized indications, contra-indications, side-effects, etc. but also to drug classifications (multi-axial SNOMED drug classes, WHO Standardized Drug Groupings, table of content of medicinal product dictionaries, simplified classifications for patients and medical students 5 From EU Implementation guide v 2.0: A medicinal product may contain one or more "pharmaceutical product(s)" (e.g. a kit containing vaginal tablets 500 mg and a vaginal cream 10% or a kit containing a combination of norethindrone acetate and ethinyl estradiol tablets and ferrous fumarate tablets). In these instances, a pharmaceutical product section is to be completed for each "pharmaceutical product". 6 From EU implementation guide v 2.0. Where applicable, the technical concept of a "pharmaceutical product" can also include information on a medical device if it is an "integral part" of the medicinal product and supports the pharmacological/metabolic/immunological action of the medicinal product, for example the scaffolding or net for a cell therapy medicinal product in accordance with Regulation (EC) No 1394/2007. Any other device copackaged (e.g. spoon, syringe) or integral (e.g. pre-filled pen) with the medicinal product must be recorded as part of the packaged medicinal product. Strength is not applicable for devices 10 The quality assurance process. The following should be investigated for the quality assurance process: • WHO UMC would be the executive responsible organization. • A suitable business and funding plan should be made • Appeal and feedback procedures should be in place • A multi-stakeholder steering committee should govern a quality assurance process. In a first phase, the procedures could be tested on the UNICOM Pilot Product List, and submitted to a round of comments inside UNICOM. Ultimately it is for EMA, FDA and other regulators to decide on the suitability of this approach.