Isabella Jørgensen: Population-Wide Disease Trajectories for Systems Medicine

Speaker:
Isabella Jørgensen (Copenhagen University, Denmark)

Title:
Population-Wide Disease Trajectories for Systems Medicine

Lecture Series:
SysBio Lecture Series: AI for Systems Medicine

Abstract

Chronic diseases often exhibit complex co-occurrence patterns driven by shared exposures, molecular mechanisms, and treatments. Temporal disease trajectories provide a data-driven approach to model these health state transitions over time. Disease trajectory modeling using population-scale data has demonstrated that the order, timing, and evolution of comorbidities carry discriminative power across various conditions. 

The Danish National Patient Registry (DNPR) include diagnoses, in the International Classification of Diseases (ICD-8 and ICD-10), given at Danish hospitals since 1977. The disease trajectory method first identifies comorbidities that occur more together than expected. Subsequently, when a significant direction is present between disease pairs, we piece them together to create longer seqeunces of diagnoses, so-called disease trajectories. Here, I will discuss several use cases of disease trajectories. Among others, we created common disease trajectories across the entire Danish population and released them as a browser in which everyone can investigate their disease of interest and its comorbidities in the Danish population, dtb.cpr.ku.dk. 

In another more specific case,  we identified 5752 kidney transplant recipients using procedure codes in DNPR. We identified disease trajectories that clustered into three major pre-transplantation disease groups: glomerulonephritis, hypertension, and diabetes. This trajectory-based stratification successfully reproduced known associations of diabetes and low albumin levels with worse survival outcomes. Importantly, our method uncovered novel associations within specific subgroups that were invisible in global population analyses. For example, among recipients with glomerulonephritis, higher basophil levels and bacterial infections were linked to poorer survival, representing previously undetected risk factors. Consequently, this approach offers enhanced precision for clinical decision-making and therapeutic stratification by identifying subgroup-specific biomarkers and risk factors relevant for systems medicine.

Read more about her research here.