R Script for the Statistical Analysis of the Study: 'Clinical Trajectories and Determinants of One-Year Survival in Cardiogenic Shock: A Multistate Analysis of the ALTSHOCK-ENIGMA Registry'

This repository contains the complete, reproducible R code used for the statistical analyses in the ENIGMA Project (2026). The script implements a comprehensive analytical pipeline to map the complex clinical trajectories of patients, from acute hospital admission to post-discharge outcomes.

The analysis relies on advanced survival analysis, specifically:

• Data Wrangling & Time-to-Event Alignment: Precise calculation of time-dependent variables to track the onset and duration of Mechanical Circulatory Support (IABP, Impella, ECMO) and Heart Replacement Therapies (HRT).

• Multiple Imputation (mice): Handling of missing clinical data at discharge (e.g., LVEF, TAPSE, OMT score) across 10 imputed datasets, with custom conditional matrices to exclude imputation for in-hospital deaths.

• Multi-State Modeling (mstate): Expansion of patient timelines into a robust 13-state transition model, tracking transitions between acute therapies, various discharge settings (Home, Rehab, Hospice), and mortality.

• Convolution & State Occupation Probabilities (SOP): Integration of in-hospital multinomial regression models (nnet) with post-discharge generalized linear models to simulate and plot dynamic State Occupation Probabilities over a 365-day follow-up.

In strict compliance with EU General Data Protection Regulation (GDPR) and ethical committee requirements, the original dataset containing sensitive patient health information is not included in this public repository.