FREM_ML has been developed using Danish National Registry data and shows strong predictive performance for one-year fracture risk.

This step focuses on validating and refining the model, including defining optimal risk thresholds for diagnostic DXA scans, ensuring the best alignment between risk estimates and Bone Mineral Density (BMD) measurements.

Validation of FREM_ML will be strengthened by incorporating Danish BMD data and ensuring consistency across the three collaborating countries through a common data model. The model will be tested and calibrated using registry data from Denmark, Sweden and Norway to ensure accuracy and adaptability across the Nordic region.

Service users are actively and continuously involved in the development and testing of FREM_ML to ensure that the tool is clinically relevant, easy to use, and patient-centred.

A Danish User Panel has been established, with around 20 participants reflecting the general community at risk. The User Panel contributes throughout the project by sharing their perspectives and experiences with osteoporosis or osteopenia, helping shape content, design, and overall relevance.

Insights from the User Panel, members of the public, and healthcare professionals are used to identify barriers and facilitators for implementation, and to guide the design, communication, and integration of FREM_ML into clinical practice.

FREM_ML will be integrated into XMO, a Danish GP system, building on experience from previous clinical decision support tools, and guided by user needs.

Implementation and integration will be conducted in three stages, that will focus on usability and a seamless integration of FREM_ML:

First, FREM_ML is adapted to data available in XMO systems. FREM_ML will be tested and validated in selected Danish general practices.

Second, an intuitive and interpretable user interface is developed through a co-design approach involving stakeholders, ensuring that FREM_ML fits into existing workflows

Third, FREM_ML is integrated into the XMO system, then implemented and evaluated in selected Danish general practices, to ensure usability and accuracy.

A randomised controlled trial will be performed, to assess the effectiveness of FREM_ML in general practice in Denmark. 

The study aims to include 44 clinics, that will be divided into two groups: one group will use FREM_ML (intervention), another will provide the usual standard of care (control group). 
The trial will follow individuals for 18 months using Danish National Health registries.

The primary outcome is the incidence of major osteoporotic fractures, with secondary outcomes including mortality, frequency of DXA scans, and treatment persistence.

The intervention is expected to reduce the yearly risk of a major osteoporotic fracture in high-risk individuals.

FREM_ML will be adapted and tested in primary care settings in Sweden and Norway, building on experiences and results from Danish feasibility studies. FREM_ML will be integrated into local GP systems and tailored to available data.

Feasibility studies will be conducted in selected clinics to assess use, clinical impact, and GP experiences.

Outcomes include changes in osteoporosis assessment and treatment before and after implementation of FREM_ML.