d.hip Projekt – digital health twin
A main focus of our platform is dedicated to clinical prediction models for prevention, diagnosis, therapy and aftercare, which are based on structured analyses and interpretations of various types of data. Among many others, these data include medical images, genetic information, laboratory evaluations, vital signs information from mobile apps or wearable devices. Through the individual assessment of a patient’s situation not only in direct comparison with existing, relevant and documented data but additionally with each patient’s own data collected over the course of his or her life, clinical statements become much more accurate, precise and personalized.
This platform-based application is our so-called Digital Health Twin. We are thus replicating and building in the medium term what will successively be a holistic individual digital model or the Digital Patient (Digital Health Twin).
The digital likeness of an individual patient updates itself with each new exam, comparing it both with relevant clinical data and with its own younger self. Therefore, the journey of our patient data begins with documentation not only as a “sick person”, but from several previous screening and prevention appointments that allow us to capture and recording the “healthy or symptom-free person”.
Responsible handling of clinical data is an absolute priority. This is why the implementation of our project requires close cooperation and collaboration between physicians, clinicians, IT and cybersecurity specialists and product developers, as well as fundamental and intensive coordination with data protection officers and ethics committees.
As part of the framework of this development, we are primarily focusing on two selected clinical application areas: rheumatic diseases and breast health.
The greatest potential of our Digital Health Twin is to detect pathological developments earlier and counteract them accordingly before they manifest when therapy would become less promising but still more costly. Multimodal data types will not only be available in a more structured way but will also be interpreted across sectoral boundaries and incorporated into patient-specific models. As a result, therapeutic strategies will also be tailored to personalized care. However, many insights will only become apparent when different factors influencing disease progression can be recorded and their effects on clinically measurable data can be correlated. This requires high-quality processed data on the one hand and strong machine learning algorithms on the other. These are structures such as those available at our d.hip!