Imagine trying to find the right key for a lock, but you have a hundred similar-looking keys and no map. This project builds a smart digital map that looks at a patient's symptoms and history to figure out exactly which rheumatic disease they have much faster. Instead of a long process of trial and error, it uses computer patterns to match patients with the right treatment immediately.
The journey to diagnosing rheumatic diseases is currently long and inefficient, leading to persistent disability and economic loss due to a trial-and-error approach in initial treatment.
Three clinical models: a patient symptom checker, a provider decision support tool for referrals, and a patient-patient similarity network for treatment optimization.
If you are a health-tech app developer dealing with high user churn due to vague symptom checkers — this project developed a patient symptom checker that uses machine learning to better categorize rheumatic complaints. This improves user accuracy and guides them toward the correct medical specialty.
If you are a clinic network dealing with inefficient referral rates and long diagnostic delays — this project developed a decision support tool for providers. It provides guidance on additional examinations and referral decisions to reduce the trial-and-error approach in initial treatment.
If you are a pharma company dealing with broad patient groups in clinical trials — this project developed a patient-patient similarity network. This allows for the optimization of diagnostic groups, helping to identify which specific patient subgroups respond best to different therapies.
Based on available project data, specific pricing or licensing costs for the clinical models are not provided.
The project includes federated learning pipelines and a GDPR compliant digital research environment, which are designed to support data integration across different healthcare levels in Europe.
Based on available project data, the specific IP and licensing terms are not disclosed, though the project aims to publish analysis protocols.
It integrates data from primary care, secondary care, and online patient advice using federated learning to maintain privacy.
The project runs from 2023-05-01 to 2028-04-30, suggesting that final validated models will be available toward 2028.
The consortium is heavily weighted toward academic and research institutions (14 out of 24 partners), indicating a strong focus on scientific validation. However, there is a significant industrial presence with 4 companies (17% ratio), including 4 SMEs, which suggests a clear path toward commercializing the resulting decision support tools and symptom checkers.
Contact Academisch Ziekenhuis Leiden for inquiries regarding the SPIDeRR analysis protocols.
Contact us to track the development of these clinical models for your health-tech roadmap.
