MSMpred

MSMpred models and predicts disease progression using multistate models to estimate transition probabilities and outcome risks.

Key Features:

• Model fitting from specific data: Fits multistate models from custom datasets provided in a predefined format, requiring user definition of states, transitions, and covariates such as age and gender.
• Data visualization: Generates histograms or barplots for covariate distributions and boxplots of patient length of stay in each state for uncensored data.
• Predictive capabilities: Accepts baseline covariate values for a new subject to predict probabilities including 30-day mortality and the most probable state at a specified future time.
• Visual representations: Produces stacked transition probability plots and other graphical outputs to represent transition probabilities over time.

Scientific Applications:

• Disease progression analysis: Models transitions between disease states to analyze dynamics of diseases with stages of increasing severity that may culminate in death.
• Clinical decision support: Provides patient-specific probabilistic outcomes to inform clinical prognostication and decision-making.

Methodology:

Uses multistate models to describe temporal evolution through defined states and transitions, fits models integrating user-defined covariates, and computes transition and outcome probabilities with complexity varying by the number of states and transitions considered.