Stratistics: Risk stratification tools for complex chronic patients

Caring for chronic conditions and multimorbidity requires anticipating which patients will need more intensive intervention, something that is hard to determine without analytical tools that capture the overall complexity of each case. Stratistics responded to this need by providing medical and healthcare staff with risk stratification tools.

The project's objectives were to:

• Identify the weight and influence of the most common chronic diseases in multimorbidity situations and assess their impact on the cost and efficiency of healthcare services.
• Build, from the real data models used to store clinical histories, an ontology that would define the semantics of the information in a standardized way.
• Identify the medical and social variables that most directly affect chronic diseases and determine the correlations that would make it possible to segment and explain the most common morbidity situations.
• Predict the risk of developing chronic diseases, especially multimorbid ones, in specific population segments, indicating the variables that explain those probabilities.
• Develop a software system that would help medical institutions implement stratification models and exchange data through system interoperability, ensuring the privacy and security of information.
• Verify and validate the results of the stratification system in the setting of the Clínica Universidad de Navarra.

Stratistics was structured around the following areas:

• Analytical stratification models: descriptive and predictive mathematical models and algorithms applied to classifying patients according to their risk.
• Proven methodological basis: development drawing on internationally recognized stratification experiences, such as those of the Johns Hopkins Hospital in Baltimore or the UK's King's Fund, and national ones, such as those of Baix Llobregat, La Fe Hospital in Valencia, or Osakidetza.
• Alignment with healthcare standards: application of the recommendations of the Joint Commission International, the IEMAC, and the Ministry of Health's Strategy for Addressing Chronicity.

CTIC's participation focused mainly on developing the mathematical models and stratification algorithms that form the analytical core of the solution, together with the associated data processing workflows and the scripts for training and serializing those models, which were later integrated into the system's technology platform. The work was carried out within a consortium coordinated by SIVSA, together with the Clínica Universidad de Navarra and Tesis Medical Solutions, and validation was performed on anonymized real clinical data from the Clínica Universidad de Navarra.