![[Translate to English:]](/fileadmin/_processed_/5/6/csm_willkommen_d34eda0786.png "[Translate to English:]")
![[Translate to English:]](/fileadmin/_processed_/e/1/csm_ueberuns_c1c8cb27e1.png "[Translate to English:]")
![[Translate to English:]](/fileadmin/_processed_/8/5/csm_ACL_11_GrafikFingerTablet_LB8_c4ce258375.jpg "[Translate to English:]")
![[Translate to English:]](/fileadmin/_processed_/b/8/csm_studien_6aaab3c7bf.jpg "[Translate to English:]")
![[Translate to English:]](/fileadmin/_processed_/6/4/csm_publikaitonen_4abdce50fc.png "[Translate to English:]")
Scope
A comprehensive descriptive analysis precedes any predictive modeling. Endpoints of clinical interest as well as prognostically important and predictive effect-modulating variables have to be defined. The decision framework ("action space"), the weighting of endpoints and decision rules must also be defined in advance. Depending on the requirements, a suitable algorithm for estimating heterogeneous therapy effects is selected. Individualized treatment recommendations are ideally presented in an application (e.g. shiny app).
Mission
Predictive modeling makes it possible to predict which patient will respond how well to available therapeutic options and which drug can be (empirically) recommended in an individual case.
Background
An individual response to therapies can be determined in routine data such as statutory health insurance billing data or hospital data.
