A robust tool for complex time-to-event data analysis in the context of myeloproliferative neoplasms

A new thesis from Karolinska Institutet shows the effectiveness of flexible parametric survival models in modeling multiple time-scales, providing a robust tool for complex time-to-event data analysis. The models were tested in the context of myeloproliferative neoplasms, a group of chronic hematologic malignancies in which the bone marrow makes too many red blood cells, white blood cells, or platelets.

This overproduction can lead to various complications, including blood clots (thrombosis), bleeding problems, transformation to acute myeloid leukemia and myelodysplastic syndromes.

In her thesis, doctoral student Nurgul Batyrbekova at the Department of Medical Epidemiology and Biostatistics, describes a novel method to modeling multiple time-scales in time-to-event analysis by using flexible parametric survival models (FPM) that can seamlessly incorporate multiple time-scales without requiring data splitting. She also explores whether traditional survival models that rely on a single time-scale lead to bias and inaccuracies in certain scenarios.

In two of her studies, using the novel method, Nürgul investigates clinically relevant questions for MPN, specifically, how the rate of thrombosis and the rate of transformation to AML/MDS are affected by age and duration of MPN disease.

Overall, her thesis could show that the new way to use multiple time-scales in survival models was better and provides answers to clinically relevant research questions in the field of MPN.