KConnect · Project

Cloud Services That Search and Translate Medical Records Across Languages

digitalPilotedTRL 7

Imagine a hospital in Hungary gets a patient's medical records from Germany — but nobody on staff reads German well enough to catch every detail. KConnect built cloud-based tools that can automatically read, tag, translate, and search medical texts in multiple languages. Think of it as Google Translate meets medical search, but built specifically to handle the messy shorthand and abbreviations doctors actually use. The project turned earlier research prototypes into commercial services and trained a community of companies to build solutions on top of them.

By the numbers

1 million+

European citizens reached through partner search portals

30+

companies targeted for the Professional Services Community

EUR 3,083,083

EU contribution to the project

consortium partners across 7 countries

SMEs driving commercial exploitation

The business problem

What needed solving

Hospitals, insurers, and health platforms across Europe struggle with medical records and publications written in languages their staff cannot read fluently. Clinical notes are especially hard to process automatically because doctors use shorthand, abbreviations, misspellings, and organization-specific jargon that standard translation tools cannot handle. This language barrier slows down cross-border patient care and limits the reach of medical search portals.

The solution

What was built

The project delivered commercial cloud-based services for multilingual semantic annotation, medical search, and machine translation of health records and publications. Concrete deliverables include a Log Analysis and Document Classification Toolkit (classifying medical texts by readability, trustworthiness, and target audience), a Semantic Annotation Toolkit for configuring multilingual annotation pipelines, a Translator Adaptation Toolkit for adding new languages with standard software skills, and demonstrated translations for Hungarian, Swedish, and Polish.

Audience

Who needs this

Hospital IT vendors managing cross-border electronic health recordsMedical search portal operators serving multilingual user basesHealth insurance companies processing foreign medical documentationClinical research organizations analyzing multilingual medical literatureLanguage technology companies expanding into the healthcare vertical

Business applications

Who can put this to work

Hospital IT & Health Informatics

enterprise

Target: Hospital networks and health IT vendors managing electronic health records across multiple countries or language regions

If you are a hospital IT vendor dealing with patients whose medical records arrive in foreign languages — this project developed commercial cloud-based services for multilingual semantic annotation, search, and machine translation of Electronic Health Records. The tools handle the specific challenges of medical text including misspellings, acronyms, and negation. The project trained at least 30 companies to build solutions using these services.

Medical Publishing & Search Portals

any

Target: Operators of medical search portals and health information platforms serving clinicians or patients

If you are a medical search portal provider looking to serve users across language barriers — this project built multilingual medical search and text classification services already integrated into highly-used medical search portals. The KConnect services reached over 1 million European citizens through partner search portals. The toolkit classifies documents by readability, trustworthiness, and target audience.

Language Technology & Localization

SME

Target: Language service providers and localization companies looking to expand into the medical domain

If you are a language technology company wanting to add medical translation capabilities without deep language engineering expertise — this project created toolkits that let people with standard software engineering skills adapt machine translation and semantic annotation to new languages. The project demonstrated adaptation to Hungarian, Swedish, and Polish using publicly available resources and web-harvested data.

Frequently asked

Quick answers

What would it cost to use these multilingual medical text services?

The KConnect services were designed as commercial cloud-based offerings built by 4 SME partners with complementary business models. Pricing details are not published in the project data, but the commercial model was central to the project design. Contact the coordinator or SME partners for current licensing terms.

Can these tools scale to handle a large hospital network's volume of records?

The project was designed for wide adoption — cloud-based architecture supports scaling, and the services were already reaching over 1 million European citizens through partner search portals by the end of the project. The Professional Services Community of at least 30 trained companies was built specifically to multiply deployment capacity.

What is the IP and licensing situation?

The 4 SME partners drove the commercial exploitation strategy with complementary business perspectives. The tools were productised from earlier Khresmoi FP7 research into commercial cloud services. Specific licensing terms would need to be discussed with the individual SME partners who own the commercial services.

Which languages are supported?

The project covered Czech, German, English, French, Hungarian, Swedish, and Polish, with specific deliverables demonstrating adaptation to Hungarian, Swedish, and Polish. The toolkits were designed to make adding new languages possible by people with software engineering skills rather than requiring specialized language engineering expertise.

How does this handle the messy reality of clinical notes?

The project specifically tackled challenges unique to Electronic Health Records: misspellings, neologisms, organization-specific acronyms, and heavy use of negation and hedging. The semantic annotation toolkit provides configurable pipelines and terminology integration to handle these real-world text quality issues.

Is this still active and supported?

The project ended in July 2017. The commercial services were built by SME partners who intended ongoing commercial exploitation. Based on available project data, the current status of the services would need to be verified directly with the partner companies or the coordinator at Technische Universitaet Wien.

Consortium

Who built it

The consortium of 10 partners across 7 countries (AT, BG, CH, CZ, HU, SE, UK) is well-balanced for commercialization: 4 SMEs and 4 universities, with a 40% industry ratio. The SME-heavy composition is unusual and signals genuine commercial intent — these are not research-only partners but companies with existing customers and business models in medical text processing. The geographic spread across Central and Western Europe mirrors the multilingual challenge the project addresses. With EUR 3,083,083 in EU funding and an Innovation Action funding scheme, this was designed from the start as a market-oriented project building on proven FP7 research.

TECHNISCHE UNIVERSITAET WIENCoordinator · AT
ONTOTEXT ADparticipant · BG
THE UNIVERSITY OF SHEFFIELDparticipant · UK
UNIVERZITA KARLOVAparticipant · CZ
KING'S COLLEGE LONDONparticipant · UK

How to reach the team

Technische Universitaet Wien (Austria) — reach out to the project coordinator for current status of commercial services and partner introductions

Order a report on this consortium See TECHNISCHE UNIVERSITAET WIEN profile →

Next steps

Talk to the team behind this work.

Want to connect with the KConnect team or explore how multilingual medical text services could solve your cross-border health data challenges? SciTransfer can facilitate introductions and provide a detailed technology brief.

Order a report on this topic More in Digital & ICT