PPCP TERMCHECKER

Bridging the language gap

There is a strong push in all Nordic countries, the EU and elsewhere for patients to become engaged in the management of their own health by giving them access to their own Electronic Health Records (EHR). The health benefits of this engagement are well recognized, but so are certain problems. Health records are traditionally written by professionals for themselves and other professionals, and the language used is – or has typically been – a mixture of professional terms, jargon, intra-clinic abbreviations and similar. Understanding what the EHR says is not even limited to the patient and his/her immediate caretaker; it is often equally difficult for any other non-specialist health care provider.

Examples of communication challenges:

  • Will patients understand their medical record online?
  • Do health information sites understand the patient?
  • Do Watson and other decision support systems speak any language beyond English?
  • Is there a way to help health care professional to use standard medical language without effort?

To solve these challenges, there is a need to link professional expressions with everyday language. Bridging the gap between professional and non-professional terminology is in fact essential in ensuring that e-health applications are user-oriented and contribute to a high quality of care. Helping patients better understand their medical records involves expanding acronyms and abbreviations, transforming idiolectic structures to standardized terms, and providing colloquial explanations without loss or distortion of information.

The Patient-Professional Communication Plugin

The Patient-Professional Communication Plugin (PPCP) project has created a technical infrastructure linking multilingual professional and non-professional language use. The platform is built through the integration of existing commercial tools for multilingual terminology management and text management. To provide a proof-of- concept for the technical solution, a pilot database specifically related to ‘heart attack’ has been developed. Heart attack is a potentially life-threatening diagnosis which is wide-spread and requires rapid and appropriate medical attention. Research shows that large patient categories may be underserved due to various reasons where communication is likely to play an important role.

The PPCP technical platform is intended for written and spoken digital exchanges. The fundamental approach of PPCP is to ensure accuracy and validity in the translation between professional and non-professional language. Until recently, machine translation systems such as Google Translate have primarily relied on statistical frequency, which – as we know – often distorts results in unpredictable ways. The PPCP platform differs from this situation since all translations are pre-validated by professionals. It means PPCP may serve as the backbone for a very large number of uses, including the improvement of machine translation accuracy and validity.

How it works

A logged-in user first chooses his/her EHR base language, in which the medical record is written. The User interface adapts its format to a variety of devices such as computer screens, tablets and smartphones.

Once logged in, the user has a dropdown menu for choosing in which languages the explanations should be provided. The termbase behind the application contains professional and everyday language related to ́heart attack ́ in Danish, Finnish, Norwegian Bokmål, Swedish and English. Some concepts and terms are also explained in Arabic, Icelandic, Northern Sami, Somali and Tigrinya.

The difficult terms, where an explanation is available in the termbase, are highlighted in blue, and the user can see them one by one by pointing the curser to one of the highlighted words.

Since the content of the medical record is never altered by direct translation or interpretation, the intentions and tone of voice of its authors are never changed. We have learned this is an important point from a clinical point of view. All terminology and explanations in the database have been developed and validated by specialists. Anyone with a web browser and login information can work in the interface.

The solution is a proof-of-concept to show the principle of how a free-text health record can be made understandable. In addition, multimedia functions can be added to the EHR interface, so that images such as x-rays and magnetic resonance tests related to the patient’s status, videos and sound files can be used to further explain concepts and terms. The underlying termbase tool, TermWeb, allows this kind of multimedia support.

Potential fields of application: health sector

  • Proactively supporting clinicians in using standardized terminology. By integrating PPCP directly with Electronic Health Records (EHR) or writing tools, a physician could continue using his/her regular documentation routines (such as clinic-specific terms) while being prompted for the corresponding standardized terms. It implies dynamic interactive replacement of non- recommended terms. This would, in turn, facilitate automation of reporting requirements for national databases (such as heart or cancer registers) and improve safety aspects of the record. The need for costly terminology training and relearning for active health care providers would be reduced. Reduced costs of training would also be possible for professionals with a different native language.
  • When electronic health records use standardized terminology, information in the records may more easily be integrated with clinical decision support systems. These range from regional or national recommendations regarding therapeutic action to the use of broad-based predictive systems such as IBM’s Dr. Watson.
  • Support for automation of communications between patients and health care providers by text or speech through contact points such as 1177.se, helsenorge.no and sundhed.dk. This communication could be urgent and very specific, where correct triage and documentation is needed.
  • Powerful Dr. Watson-type systems that handle Big Data are typically monolingual. PPCP provides a potential avenue for broadening the scope of languages handled.
  • Many excellent smartphone and tablet applications for the interactive management of medical conditions lack the ‘language plugin’ that would facilitate their use among the less educated, old, chronically ill or non-native language speakers. A language plugin would also open new geographic markets for these applications. PPCP can provide the necessary backbone for this multilingual use.

Potential fields of application: eCommerce

Using the termbase for controlled language purposes

In the PPCP interface, terms are recognized and explained in a “free text” context. This function has the potential for wide generic use, for example in e-commerce. PPCP has been demonstrated to individuals involved with Product Information Systems, where product descriptions for e-commerce platforms are automatically generated from a database. Different e-commerce platforms (Amazon, Ebay etc.) often require different terms for their respective product descriptions. This means “free text” analysis and subsequent conversion of certain terms to what is “correct” for the particular e-commerce platform could be of great interest.