Stratification of Patients using advanced Integrative modeling of Data Routinely acquired for diagnosing Rheumatic complaints

Project Information

SPIDeRR

Grant agreement ID: 101080711

DOI

10.3030/101080711

EC signature date 20 March 2023

Start date 1 May 2023

End date 30 April 2028

Funded under

• Health

Total cost

€ 5 182 424,97

EU contribution

€ 5 182 424,97

Coordinated by

ACADEMISCH ZIEKENHUIS LEIDEN
Netherlands

Periodic Reporting for period 2 - SPIDeRR (Stratification of Patients using advanced Integrative modeling of Data Routinely acquired for diagnosing Rheumatic complaints)

Reporting period: 2024-11-01 to 2026-04-30

Summary of the context and overall objectives of the project

Globally 1.71 billion people have musculoskeletal symptoms, the leading contributor to disability. Early disease stratification is important to ensure appropriate care (most suited healthcare provider and best treatment choice). Currently the patient journey to diagnosis and effective treatment is long and inefficient, resulting in persistent disease burden and economical loss. This is due to insufficiently understood relations disease causes and similarities in symptoms between diseases, insufficiently distinguishing tests, trial and error approach in initial treatment. SPIDeRR aims to disentangle the real-life complexity of early diagnosis of rheumatic diseases by considering the complete web of factors influencing patients’ symptoms. SPIDeRR’s approach will go well beyond the state-of-the-art in the following ways: - By identifying different disease groups, requiring different therapies, amongst patients with similar symptoms in contrast to the traditional approach aiming to only capture one disease early. - By integrating all relevant data dimensions from every healthcare level (primary and secondary care and patients seeking advice online). - By translating and applying machine learning techniques from the “omics” field to clinical patient data, which will result in new pipelines for translational data science SPIDERR will deliver three clinical models -a symptom checker for patients -a decision support tool for (primary) care providers providing guiding additional examination and referral decisions -a patient-patient similarity network to optimise diagnostic groups in rheumatology and support treatment decision To achieve this we additionally deliver solutions for data integration and shared analyses though GDPR compliant digital research environment and federated learning pipelines. Finally we will test the acceptability of the models through stakeholders studies and provide an implementation scene tailored to current healthcare in Europe.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

During the second reporting period, SPIDeRR made strong progress towards improving stratification of patients with musculoskeletal complaints through secure data integration, digital triage tools, advanced modelling and implementation research. The consortium further established its legal, governance and technical framework for cross-border health-data research, centred on the myDRE trusted research environment and supported by data access agreements, DPIAs and a dedicated data governance framework. Primary- and secondary-care datasets from several European healthcare systems were obtained, structured and harmonised, enabling the first large-scale cross-country analyses of musculoskeletal complaint coding, referral patterns and care trajectories, covering more than 4 million adults in primary care and more than 600,000 patients in secondary rheumatology care. Rheumatic? progressed from model training to external validation and usability testing, with the immune-mediated rheumatic disease model showing strong rule-out performance and additional models retained for fibromyalgia and confirmation of osteoarthritis/fibromyalgia. The tool was further translated and culturally adapted across consortium languages, while validation and recruitment continued in general population, primary-care and rheumatology settings. Modular-SPIDeRR advanced towards operational model implementation through common data model development, a reusable Kedro-based analytical pipeline and refined validation strategies. The Rheumatic Digital Twin framework was finalised and published, with clinically relevant prototypes developed for joint-involvement patterns, referral-letter triage and longitudinal disease trajectories. WP5 progressed genetic and biological risk-factor analyses through harmonisation of genotyping datasets, polygenic risk score modelling and causal modelling of flare-risk data. Implementation work also advanced through completed patient journey studies, five SPIDeRR personas, interactive patient journey maps, and usability, acceptability and feasibility studies of Rheumatic?. Dissemination increased through publications, abstracts, oral presentations, newsletters, tutorials, workshops and patient-facing communication. Overall, SPIDeRR remains on track, with major achievements in secure data governance, harmonised multi-country analyses, validated digital triage tools, reusable AI pipelines, Digital Twin methodology and implementation evidence. The main challenges were ELSA Science AB’s bankruptcy and remaining SERMAS data-sharing restrictions under Spanish ENS requirements, both mitigated through task redistribution, platform continuity measures and local-analysis workflows.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

not applicable yet