Description:

The University Hospital Halle seeks an IT service provider to support the development and implementation of a Large Language Model for clinical data integration, using open-source solutions and providing training for local staff.

The University Hospital Halle (Saale) is carrying out the following procurement on behalf of the Medical Faculty of the Martin Luther University Halle-Wittenberg:

The University Medical Center Halle (UMH) is constantly interested in new, innovative solutions. In light of recent advances in AI, especially large language models, we would like to introduce this technology in the context of registry databases to provide staff and researchers with tools to support data entry. For this step, we are looking for a company that has experience with the operation of large-language models.

In a current research project running from 01.10.2024 to 31.03.2025, the goal is to develop a Large Language Model (LLM) for the reading of semi-structured clinical findings into a locally hosted secondary database. The databases already created by the Data Integration Center (DIZ) at the hospital and the established LLM infrastructure provide an optimal basis for this project.

Specifically, the aim is to implement two application scenarios for a proof-of-concept:

a) Information defined from semi-structured pathology findings is to be recognized, evaluated and recorded in a database in a structured manner in a predefined format.

b) In a second part, the programs and workflows established in point a) are to be transferred to semi-structured radiology findings.

The UMH is thus intended to expand the showcase role for the integration of AI tools into clinical and scientific workflows, from which both other hospitals and (e.g. public) administrations can benefit. In order to avoid dependencies on commercial providers, open source models are to be used.

The intention is to establish an internal AI solution for the above-mentioned application scenarios together with an IT service provider with strong AI knowledge. To this end, a support service is to be provided in the establishment of an internal AI in medicine. The service is to be provided on demand in the period until 31.03.2025. The aim is to enable local research and development employees to use LLMs and to build up or deepen subject-specific knowledge. In addition, independence from external companies (e.g. license fees for software), but especially from large market leaders (OpenAI, Microsoft, Google, etc.) is to be excluded by using open-source models and open-source tools. The planned applications are to be operated on local servers of the UMH.

The aim of the award is to contractually bind an IT service provider to support the planning and implementation of LLMs and their integration into the local IT infrastructure, as well as training for IT staff in the period from 01.01.2025 to 31.03.2025.

The planned award will take place within the framework of a negotiated award without a competitive bidding process in accordance with §12 UVgO. Interested companies that see themselves in a position to provide the service are called upon to submit a bid by 16.12.2024, 10:00 a.m. The procedure will be carried out exclusively via the federal government's e-procurement platform (electronic procedure / electronic submission of bids), interested companies are asked to register by 12.12.2024; 11:00 a.m. by e-mail to vergabe@uk-halle.de to receive the tender documents.

The documents to be submitted with the tender and the requirements for assessing the suitability of the candidates are set out in the tender documents.

The grounds for exclusion pursuant to Sections 123 and 124 GWB apply mutatis mutandis.

Binding period: 31.01.2025

Client:Medical Faculty of the Martin Luther University Halle-Wittenberg

Procurer:University Hospital Halle (Saale) Public Law Institute [Ernst-Grube-Straße 30,06120 Halle (Saale)]

Notice type: UVgO/NA

The Buyer:

Universitätsklinikum Halle (Saale); Geschäftsbereich I - Versorgungsmanagement; Vergabestelle

Additional information:

Link:

Notice Link

CPV Code(s):

72200000 - Software programming and consultancy services