Contracted by the Commission of
the European Communities
ETD/93/A02600/MI/10
Under the auspices of: Commission of the European Communities (DG III) Directorate General of Industry, P. Deboyser (In succession of F. Sauer) Project Supervisor : B. Hughes (in succession of P. Brunko) Contractor: Heymans Institute of Pharmacology, University of Gent Prof. M. Bogaert, PhD, MD, clinical pharmacologist Project leader : R. Vander Stichele, MD, Senior Researcher Co-contractor : Mercator College, Department of Applied Linguistics P. Van Hauwermeiren, PhD, director. Project managers : R. Coppieters, PhD (sept 93 - July 94) C. Meuleman, PhD (August 94 - April 95) Terminographic assistance : Mercator College, Centre for Terminology R. Eylenbosh, PhD., C. De Groote, R. Van Hove. Eurodicautom, H. Wellenstein. Informatics assistance : Mercator College, Centre for Informatics J. Onraet, R. Van Hove. Internet application managers : B. De Wispelaere, University of Gent T. Vander Stichele, University of Gent Secretarial assistance : A. De Smet, Heymans Institute C. Thienpondt, Mercator College End report assistance : N. Moeykens Translators English Technical : R. Coppieters, PhD> Popular : J. Riches French Technical : R. Godijns, PhD Popular : H. Pauwels Spanish Technical : P. Vanden Bulcke Popular : P. Vanden Bulcke Portuguese Technical : J. Coppens Popular : M.G. Soarez Italian Technical : K. Meert Popular : A. Muratori Dutch Technical : R. Coppieters, PhD Popular : E. Eylenbosch, PhD Greek Technical : K. Savaidis Popular : K. Savaidis Danish Technical : A. Sichlau Popular : A. Sichlau German Technical : B. Van den Heede, PhD Popular : I. Mayer-Falk Validators English Technical : R. Coppieters, PhD Popular : M. Hill French Technical : Prof. J. Giroud, MD, PhD Popular : D. Staquet Spanish Technical : P. Gomez, MD Popular : P. Gomez, MD Portuguese Technical : G. Lima C. MD Popular : M.G. Soarez Italian Technical : M. Miselli, MD Popular : M. Magris, PhD Dutch Technical : E. Eylenbosch, PhD Popular : E. Eylenbosch, PhD Greek Technical : E. Theologos, MD Popular : I. Ikonomopulos, PhD Danish Technical : M. Weber, MD Popular : B. Sichlau German Technical : U. Goerlich, MD Popular : U. Srinivasan
Belgium, Germany and most southern European countries have a long tradition in providing patients with written drug information enclosed in medication packages. In the Netherlands, Denmark, the United Kingdom and Ireland, on the other hand, such a tradition is far less widespread, not to say non-existent.
Sociological research has clearly indicated that patients read the inserts and believe them to be a useful and rich source of information on medication, in spite of the fact that those inserts are written in technical language and, for that reason, sometimes hard to understand. It needs pointing out, however, that the very same readers of the technical package inserts criticize the inserts on readability, legibility and graphic illustration and admit to having difficulties remembering their contents.
1.2 Motivation for the mandatory inclusion of understandable patient package inserts
Patients are entitled to receive information on the nature of their medications. Any such information should be drafted in plain language, meant to be read by lay persons.
The user package leaflet is not a do-it-yourself guide for the patient, but a tool in the relationship that exists between physicians, pharmacists and patients, designed to assist the patient in correctly taking the medication. In addition, the user package leaflet is meant to promote patient compliance and to provide him/her with detailed information on side effects so that he/she can respond adequately when confronted with them.
1.3 The terminology problem
The authors of user package leaflets are academically trained physicians and pharmacists. They have not benefited from any specific linguistic training and will consequently have to overcome considerable hurdles at the terminology level when attempting to write texts in plain language meant to be read by lay persons. Some centralized assistance is needed to provide the authors with popular equivalents of technical terms, if one is to avoid a babylonic multiplication of terms.
1.4 Choice of the cooperating institutions
The Commission wanted to take advantage of the Belgian experience in this field.
Already in 1984, the Belgian health authorities decided to transform technical package inserts into full information patient package inserts, written in lay language and addressing adults with a schooling grade up to the age of 16. The transition from technical package inserts to patient package inserts for about 6,000 medicinal preparations started in mid-1988 and was completed in 1991.
The Heymans Institute of Pharmacology at the University of Gent had been entrusted with the evaluation of the Belgian package inserts project as early as 1988, so it could rightly claim a significant experience in this field. Moreover, the Heymans Institute contributes to the Belgian drug policy in different ways: it provides physicians and pharmacists with appropriate information on medical drugs, it acts as an advisory body to the Drug Registration Authorities for the Belgian market and the Institute participated in the validation of older medicines and in the revision of package inserts of older medicines. A contract with the European Commission was signed in September 1993. The Heymans Institute needed the co-operation of an institution specialized in translating disciplines. Preference was given to the Mercator Language College 1 At the time of the contract known as PHVT (Provinciale Hogeschool voor Vertalers en Tolken). in Gent, Belgium for various reasons. Firstly, at the Mercator College, the major European languages are taught, namely English, French, German, Spanish and Dutch. Secondly, the Mercator College had acquired some experience in medical terminography as it was its director, Dr. P. Van Hauwermeiren, who published in 1988 a trilingual medical glossary under the authority of the Belgian Ministry of Health.
Thirdly, the Mercator College had proved to be a dynamic institute whose staff was able to provide the scientific, pragmatic and logistical basis the project would require. A cooperation agreement was consequently concluded between the Heymans Institute and the Mercator College.
1.5 Aim of the project
As the then 12 Member States of the EC, using 9 different languages, were expected to comply with the EC Directive, a glossary of scientific and popular medical terms in nine languages (English, French, German, Dutch, Spanish, Portuguese, Italian, Greek and Danish) was required. The contract between the Heymans Institute and the European Commission stipulated the creation of 9 medical glossaries. For each of the nine official languages of the European Union, a list was to be selected of 1,400 scientific medical terms frequently used in the context of written drug information but difficult to understand for a lay person. In a next step, a popular equivalent for each scientific term was to be proposed in the nine languages.
2.1 Compilation of the basic list of technical terms in Dutch
Since no English or French computerized compendia of medical terms used in technical data sheets were available at the time of the start of the project, it was decided to select those terms from a compendium including the totality of Dutch scientific data sheets on medical products. This compendium was available from the Belgian Pharmaceutical Association under the form of a CD-ROM optical disk. A computer programme, scanning the technical data sheets, generated an alphabetical list of approximately 45,000 Dutch word forms. Each word form was supplied with numerical data on frequency of occurrence. Those 45,000 word forms included medical as well as non-medical terms, as both types of word forms occurred in the data sheets.
This preliminary list was reduced to a list of about 11,000 headwords through the elimination of irrelevant word forms (articles, prepositions, etc.), of terms with a frequency of occurrence of less than five, and of related forms often spelled in a variety of ways (i/k - ph/f - ae/e, etc.). A second reduction of this list of 11,000 headwords was achieved by leaving out the following terms:
2.2 Towards an English basic list
As explained above, the basic list of 1,830 medical terms stems from a Dutch frequency search. This basic list had to be translated into the nine community languages. It seemed advisable to provide the future translators with an English version, English being the lingua franca of the medical field. The Dutch list was translated into English by Dr. R. Coppieters, member of the English department of the Mercator College 2 The majority of the English terms were found in the APBI Data Sheet Compendium, Eurodicautom, Dorland's Medical Dictionary and the BMA Family Health Encyclopaedia. .
Next, a scientific and a popular medical equivalent had to be provided for each term in the English list in each of the nine community languages (Dutch, French, German, English, Spanish, Italian, Portuguese, Greek and Danish).
In the basic list, a number of non-medical terms were present (e.g. abrupt, activate, antecedent, elective, intrinsic, manifest, perennial, recurrent, replication). Those non-medical terms were included in the basic list because their frequency count was ten or more in the Belgian Dutch language CD-ROM compendium. It was decided to leave them in as the target reader had been defined as 'a mentally healthy adult with a formal education to the age of 16'. Any such reader might be expected to have problems understanding terms such as those indicated above. Consequently, medical as well as non-medical terms needed to be put in plain language in order to assist future authors of package inserts in writing fully intelligible texts for the general public.
2.3 Definition of each item
Soon after the start of the project, cooperating translators were confronted with problems of polysemy, which means that a term can have more than one meaning, with different translations according to the meaning. To maintain homogeneity, the translators requested definitions of the terms. An English definition of each of the 1,830 medical terms was therefore provided. This task was performed by the Centre for Terminology of the Mercator College.
2.4 Multilingual translation of the basic English list of technical terms
Senior staff member at the Mercator Language College were entrusted with the English, French, German, Dutch and Spanish translations. The translation of the 1,830 technical medical terms into Italian, Portuguese, Greek and Danish was carried out by external, professional translators, specialized in medical translations or collaborating with people from the medical world 3 For a full list of the translators, see appendix.. Assistance was provided by EURODICAUTOM, in the form of printouts of terms available in the European Terminology database. The translators were instructed to record their work in a structured text-oriented database.
Thus, each English scientific medical term was provided with a scientific equivalent in each of the nine languages.
2.5 Transformation of technical terms into lay language equivalents in each of the nine languages
In a second stage native professional translators, either with medical experience in the translation field, or collaborating with a physician, were selected for each of the nine languages. Their task consisted of finding a lay language equivalent (word, paraphrase) for the technical medical terms.
2.6 Validation
Finally, a random sample of some 250 terms was drawn from each translated list (scientific and popular medical terms) and submitted for validation to an academically trained linguist in the native country or a native speaker with a medical training 4 For the validation reports, see appendix.. Validation for both the medical and the popular versions was not required in every language since some translators closely cooperated with physicians (Denmark, Greece, Germany) or were physicians themselves (Portugal). Each validator was provided with validation instructions 5 See appendix. and a list of corresponding English definitions. The validation reports are included in a separate booklet. It should be noted that all validations proved to be satisfactory. Comments and propositions for adaptations were limited to 2 % of the proposed terms.
2.7 Project time-table
The following survey summarizes the project time-table :
September 1993 : Contract signed between the Heymans Institute and the Commission.Deadline : September 1994.
October 1993 : Contract signed between the Heymans Institute and the Mercator College.
December 1993 : A basic Dutch list of 1,830 medical terms is selected.
January 1994 : The English basic list is translated.
May 1994 : A dictionary is compiled.
June 1994 : The French, German and Spanish scientific medical versions
are ready.
September 1994 : A new deadline is negotiated with the Commission
: May 1, 1995.
January 1995 : The Danish and Greek scientific medical
versions are ready.
February 1995 : The popular versions are ready in the nine languages.
A partial validation process is run. A first software interface is created (and
abandonned because of amount of manual work needed to complete)
April 1995 : Completion of the project.
Note : The software version of the glossary was transformed into an internet application in the summer of 1995.
The translators had also been allowed or were even invited to add terms they believed to be relevant or to include synonyms in the commentary field of the database record.
Eventually nine glossaries have been produced, each exceeding the required 1,400 terms, in a structured format, with ample comments and/or synonyms in a majority of the languages.
First approach
In a preliminary approach, a limited number of terms from the glossary
were processed in a Rich Text File (RTF) language, in a Windows Help environment. This
approach was abandoned because of it would require too much manual input to complete the
processing of complex interactions between languages and terms.
In April 95, the first version of the multilingual glossary of technical and medical terms was ready. By that time, the World Wide Web became accessible on a large scale in Europe, and it seemed appropriate to develop the glossary into an Internet application. This would facilitate distribution and eliminate substantial copying costs. Furthermore, the structured content of the glossaries allows full exploitation of the Hypertext Mark Up Language (HTML), with computerized creation of the thousands of hyperlinks, necessary for this application.
Functional analysis
The first version of the mulitlingual glossary is but a starting point for a
broad consultation process of relevant interested parties in the
different member states. Nine
working documents now exist, constructed with the same methodology, around a common
basic list. This first version is a starting-point for discussion and study, and not a finished
product. A broad consultation process of relevant interested parties in the different member
states should now develop.. Each member state has one or more official languages, and
should receive the appropriate glossaries. Pharmacists, physicians, regulatory affair
managers, consumer representatives in each of the member states can study and comment
the glossary.
The results of this process of refinement must be collected in a systematic way. Therefore, communication with all the interested constituencies is necessary.
There is of course the traditional way of publishing the multilingual glossary in print, and distribute hard copies to interested readers. In addition, however, it would be interesting to present the multilingual glossary in an interactive way on the information highway. This would turn the glossary into an easily accessible and always up to date working tool. It would furthermore facilitate the process of refining the glossary by organising electronic mail between the users of the glossary and the authoring team.
This would help to finalise the multilingual glossary into a useful tool for enhancing the quality of communication with patients about medicines.
The glossary will be of particular interest for the application of the European Directive 92/27/EEC on user package leaflets. Therefore, it needs to be accessible to the writers of user leaflets within the European Pharmaceutical industry, working in environments with easy access to the Internet.
User interface of the Internet application.
The multilingual glossary is accessible as a site
of the information highway. Anonymous visits to the site are possible, with optional possibility
to comment through E-mail.
Viewers will enter the application through a welcome screen, where they can choose
These documents have been broken down in smaller files to allow faster response to user requests.
The basic information has been made in WordPerfect 5.1 for DOS, which is intensively used in translator schools and by linguistic professionals. It provided easy and standard exchange between translators and the project team.
The WordPerfect files were then transferred to Hyper Text Markup Language ASCII files (through a beta-version of WordPerfect HTML editor for Windows).
Hyperlinks have been laid between all documents (glossary, multilingual lemma collection, dictionary and lists). The creation of hyperlinks was programmed in the WordPerfect macro Language.
Access to the application is platform independent.
The interface with the user is English, and not yet adaptable to the user preference.
It was not possible to integrate the Greek documents into the Internet application. The problems with the representation of Greek letterset together with the full Western European letterset in Internet applications have not been solved in a satisfactory way, yet, as Internet browsers, network servers and user screens need special adaptations. The Greek glossary is however available by File Transfer Protocol (FTP) within the application.
A number of translation problems were identified by the various translators and need to be brought to the reader's attention : overlap between scientific and popular terms, adjectives, terms acting both as an adjective and a noun, word class, synonyms, polysemy, non-medical terms and terms left untranslated.
4.1 Overlap between scientific and popular terms
As indicated above, a scientific and a popular equivalent had to be provided for each of the 1,830 terms of the English basic list in each of the nine EC-languages. The translators were allowed to consider the scientific term as equivalent to its popular version or to propose a popular equivalent. It has already been indicated that differences between Germanic and Romance languages cropped up. Romance languages being closely related to Latin, many scientific medical terms of a Latin origin pose no problems to native speakers of Romance languages. Therefore those terms did not require a popular equivalent. This explains why the French, Spanish, Italian and Portuguese vocabularies contain a number of popular terms which are identical to the original scientific term.
However, the number of terms for which the scientific and popular versions are identical is considerably smaller in the Spanish vocabulary than in the other Romance language vocabularies. In fact, it would appear that the Spanish translator went at great lengths to propose a popular equivalent for nearly every scientific term. In this respect, the Spanish lay language list might conceivably stand as a reference list for any further popular versions.
As in the Germanic languages scientific medical terms appeared to be less accessible to laymen, more alternative popular terms or paraphrases had to be supplied. English, which has a relatively large Romance component in its vocabulary anyway, was a relative exception, i.e. it had more cases of overlap between scientific and popular terms than the other Germanic languages. The Greek vocabulary, for obvious reasons, has also a large overlap ratio.
The translation of scientific terms into popular equivalents can be based upon either of two principles : familiarity or accuracy. The French and Spanish translators clearly aimed at accuracy, whereas the other translators mostly gave preference to using terms familiar to the general public. As the translation instructions did not indicate which principle had to be used, the choice was the translators'. It is obvious that, whenever accuracy is the first aim, familiarity is liable to suffer or that whenever translators decide to opt for familiarity, the accuracy of their translations may be at risk. Future research will have to decide upon measures of efficiency.
In some cases it was impossible to explain the scientific terms in understandable lay language. This was particularly the case with terms regarding blood chemistry. In such cases it was decided to maintain the scientific term as a valid popular equivalent since a description or paraphrase might be just as inaccessible to the layman as the scientific term.
4.2 Adjectives
Besides nouns and verbs, the English basic list also contains a relatively large number of adjectives. Those adjectives were included in the list on the basis of the frequency criterion (count of ten or more).
The translation of some adjectives into popular language posed a problem in Romance as well as in Germanic languages. Some adjectives could easily be translated by a single word; others, however, could only be translated into lay language by means of a description. In some cases the description of the adjective inevitably contains a scientific medical noun term. Term 161 'antithrombotic' is a good example. The popular version of this term uses a description in all languages except Greek. The Danish, French and Italian popular equivalents of this term contain the scientific term (respectively trombedanelse, thrombose and trombi). It stands to reason that the layman will have difficulty in understanding the scientific term 'trombosis'. Therefore, one might have expected an explanation of the scientific term, 'trombosis' in the present case. Some translators expounded on the scientific term and put their comment between brackets. The French translator, for instance, included a paraphrase of 'thrombose' in the commentary field, whereas the Danish translator did not provide an explanation for 'trombodanelse'. The Spanish translator on the other hand, solved the problem by incorporating the explanation of 'trombo' in the description, i.e. 'que impide la formacion de tapones de sangre y los disuelve'. This is, of course, only possible when the scientific term can be easily explained in lay language.
It needs reemphasizing that scientific noun terms could not always be provided with a single word popular equivalent, nor could those nouns always be provided with an intelligible popular equivalent description 6 Examples of scientific terms (nouns and adjectives) translated into popular equivalents by means of a description are to be found in the appendix. .
In some languages, more specifically in Spanish and Italian, a considerable number of terms can be interpreted both as adjectives and as nouns, according to their linguistic environment. The substantive form of those terms mostly refers to medicines or to the substance of a medicine. When the English term of the basic list had to be interpreted as a noun, the Spanish and Italian translators also translated the term by a noun 7 See appendix..
4.3 Word class
The French, Spanish and German translators indicated the word class of the translated terms, although the translation instructions did not explicitly require them to do so. Any such information is undoubtedly of great interest to both the translator and the user. It would be interesting to have the word class indicated in all the languages. There was, unfortunately, no time to launch this operation in the other six languages. Therefore, it was decided to leave out word class indications in the user interface, whilst maintaining them in the data base entry.
4.4 Synonymy
The translation instructions stipulated that all the terms of the basic English list were to be provided with one single scientific and one single popular equivalent and to list synonyms in the commentary field of the database. However, this instruction was not consistently respected by the translators. The original Dutch and English vocabularies did not include synonyms. The Portuguese and Italian translators proposed synonyms for a very limited number of terms. The Danish, Greek and German translators provided a large number of the scientific as well as the popular terms with synonyms. Finally, the largest number of synonyms was found in the French and Spanish vocabularies.
There is no doubt that the inclusion of synonyms will assist the authors of user package leaflets in their writing task. Therefore, the suggested synonyms were kept in the database. However, the translators were required to single out one term which they deemed to be the most suitable, always within the context of written drug information and with respect to the proposed definition. This reference term will appear on the interface screen. When the term has one or more synonyms, this will be indicated by an asterisk, inviting the reader to consult the data bank.
Putting forward just one reference term clearly proved to be an arduous task. As no scientific selection criteria were available, the translator's individual intuition had to be relied upon. In this respect the exhaustive work done by the French and Spanish translators might stand as an example
4.5 Non-medical terms
Many translators questioned the presence of non-medical terms in a medical glossary. As explained before, this was the direct result of the frequency count identifying terms with a frequency of occurrence of ten or more.
Some translators felt that those terms should have been combined with medical terms, thus complying with the terminographic definition of the concept of 'term'. 'Juvenile', for instance, ought to have been linked up with 'acne' if the presence of both terms is to be justified. Undoubtedly 'juvenile acne' appears as such in medical package inserts. However, the frequency research broke up the word pair identifying single words instead of 'terms'. It would admittedly have been better to take the frequency of semantically linked terms as a starting point instead of the frequency of isolated words 8 For examples see appendix..
4.6 Polysemy
A close look at the definitions list reveals the presence of a number of polysemous terms 9 For examples, see appendix., i.e. terms that have more than one meaning depending upon the context in which they are used. The Spanish, French and, to a lesser extent, German vocabularies provide a number of scientific terms with several popular equivalents with different meanings. The Dutch, Danish, English, Greek and Italian translators restricted themselves to including only one 'meaning' (as they had been required to do) even when faced with a polysemous term.
Here again, it might be advisable to follow the example of the French and Spanish translators who went to a great length to provide as much information as possible for the future authors of patient package inserts, thus actually going far beyond what was required under the project terms.
4.7 Terms left untranslated
In each of the nine vocabulary lists there are a number of terms for which either a scientific or a popular equivalent or even both, have not been provided. The absence of translations may be due to any of several reasons:
The basic list of items for the multilingual vocabulary was compiled from a frequency search of terms in scientific data sheets on medical products. This search was performed on a Dutch language CD-ROM optical disk available from the Belgian Pharmaceutical Association.
Due to limited funding and the tight time schedule, thorough academic research and testing had to make way for a more pragmatic approach calling upon a network of (native) professional translators, experienced in medical translations or working closely together with physicians. Assistance was provided by EURODICAUTOM, in the form of printouts of terms available in the European Terminology database.
Besides a list of English definitions and translation instructions, the translators were provided with a disk and a free text database highly facilitating data input. Specific instructions and frequent individual monitoring of the translation tasks proved to be a satisfactory approach.
As it appears from the project time-table the results of the projects were due at an earlier date. However, as the translators confronted the project leaders with the justifiable request for a definition or context for each of the 1,830 terms. The provision of those definitions proved to be a time-consuming task and resulted in a revision of the deadline from September, 94 to April, 95.
The glossary was in the first place meant for the authors of user package leaflets. They will
now have at their disposal:
- an English dictionary of 1.830 medical terms, frequently used in written drug
information
- nine glossaries of 1830 scientific and popular medical terms, one in each of the nine
official languages of the European Union.
- an electronic database of the dictionary and glossaries
- a user-friendly software interface as an Internet application, installed at the
World-Wide Web server of the University of Gent, Belgium.
The accessibility of medical scientific terms is linguistically and culturally determined. There are fundamental differences between the Romance and Germanic languages in this respect. Hence, it seems advisable to repeat the frequency count of terms in each of the nine EC-languages.
This project is open for extension to other languages such as Swedish, Finnish and American English.
In future projects of multilingual glossaries, it seems advisable to set up a pilot project per language with initial translation of a limited list of terms. Such a pilot project might reveal a number of translation problems from the start and avoid some of them in the later phases of the project. .
Glossaries such as those produced by this project can provide a starting-point for computerized readability formulas, another potential useful tool for authors of user package leaflets.
The medical glossary project has undoubtedly brought about an interesting and useful tool, which, to all intents and purposes, will need further improvement.
It would be of interest to submit the medical vocabulary to longitudinal research.
Field work could provide indications on the appropriateness of the proposed scientific and popular translations.
The actual use of the glossaries should be monitored. Indeed, the glossaries will probably not only be used by the authors of user package leaflets but also by physicians, pharmacists, educational health workers, etc.
Criticisms and comments from the users will have to be collected.
The glossary could also be improved by means of academic testing. Encyclopaedia and recently compiled compendia on medical products in various languages can be consulted in order to improve each glossary. Observations put forward by both the translators and the users should be submitted to rigorous academic research and suggestions should be explored for further improvements and research (comparative lexicographical research, cross-language studies, socio-cultural analysis and research, etc.).
Finally, the methodology developed for this project in the field of medication information might be applicable to similar projects in other fields of consumer information.