Client: Fraunhofer Cluster of Excellence for Immune-Mediated Diseases CIMD
Background:
When referring patients to a specialist via the primary care provider (general practitioner / orthopedic doctor), suspected immunological diseases such as psoriatic arthritis and axial spondylarthritis often lead to delays and a high workload for medical professionals because of the incorrect allocation of patients. There is no strategy to evaluate unstructured findings in a targeted manner and to increase the probability of correct allocation by recognizing a specific pattern of findings.
Project Description:
Methods focused on machine learning and data sciences are used in close cooperation with medical expertise in order to reduce the workload:
- Unstructured data packets are structured using machine learning methods such as text mining
- Findings are recognized and linked after digitization
- An indication-specific algorithm is developed that evaluates the characterization of disease phenotypes from various sources (text, laboratory and imaging).
A demonstrator of the algorithm will be developed using historical data packages. With the help of a prospective clinical study at three locations, it will then be validated and adapted in Germany
Project Tasks:
As a socio-economic partner in the project, the Fraunhofer IMW team focuses on the aspects of user-centering and exploitation. In workshops with specialists, they analyze the prototype with a focus on usability. The use of artificial intelligence indicates possible savings potential and increases in efficiency.
Project Duration:
1.6.2020–31.3.2021