Joint modeling of longitudinal biomarker trajectories and survival outcomes. Links repeated measurements over time to survival endpoints for dynamic risk prediction and personalized medicine applications.
Usage
jointmodeling(
data,
id,
time_longitudinal,
biomarker,
survival_time,
survival_status,
covariates,
functional_form = "linear",
random_effects = "intercept_slope",
error_structure = "independent",
survival_model = "cox",
association_structure = "current_value",
estimation_method = "bayesian",
mcmc_chains = 3,
mcmc_iterations = 12000,
mcmc_burnin = 2000,
mcmc_thin = 5,
dynamic_prediction = TRUE,
prediction_horizon = "0.5,1,2,3",
prediction_window = 2,
internal_validation = TRUE,
cv_folds = 5,
discrimination_metrics = TRUE,
plot_trajectories = TRUE,
plot_mean_trajectory = TRUE,
plot_survival_curves = TRUE,
plot_dynamic_auc = TRUE,
plot_residuals = FALSE,
competing_risks = FALSE,
left_truncation = FALSE,
time_varying_effects = FALSE,
multiple_biomarkers,
baseline_hazard = "unspecified",
prior_specification = "default",
convergence_diagnostics = TRUE,
parallel_computation = TRUE
)Arguments
- data
The data as a data frame in long format (multiple rows per patient).
- id
Patient ID for linking longitudinal measurements
- time_longitudinal
Time axis for biomarker trajectory
- biomarker
Longitudinal biomarker variable
- survival_time
Survival time variable
- survival_status
Survival event indicator
- covariates
Additional predictive variables
- functional_form
Longitudinal model specification
- random_effects
Random effects specification
- error_structure
Within-subject error correlation
- survival_model
Survival distribution choice
- association_structure
Association structure specification
- estimation_method
Estimation approach
- mcmc_chains
Parallel MCMC chains
- mcmc_iterations
MCMC sampling iterations
- mcmc_burnin
MCMC warm-up period
- mcmc_thin
MCMC thinning parameter
- dynamic_prediction
Enable dynamic prediction
- prediction_horizon
Prediction time windows
- prediction_window
Prediction look-ahead window
- internal_validation
Cross-validation assessment
- cv_folds
Cross-validation partitions
- discrimination_metrics
Compute discrimination measures
- plot_trajectories
Generate trajectory plots
- plot_mean_trajectory
Show mean trajectory
- plot_survival_curves
Generate survival plots
- plot_dynamic_auc
Show discrimination over time
- plot_residuals
Generate diagnostic plots
- competing_risks
Include competing events
- left_truncation
Handle delayed study entry
- time_varying_effects
Time-dependent association parameters
- multiple_biomarkers
Multiple longitudinal outcomes
- baseline_hazard
Baseline hazard modeling
- prior_specification
Bayesian prior choice
- convergence_diagnostics
MCMC quality assessment
- parallel_computation
Enable parallel processing
Value
A results object containing:
results$instructions | a html | ||||
results$progress | a html | ||||
results$errors | a html | ||||
results$warnings | a html | ||||
results$dataSummary | a html | ||||
results$longitudinalResults | a html | ||||
results$survivalResults | a html | ||||
results$jointModelResults | a html | ||||
results$diagnostics | a html | ||||
results$dynamicPredictions | a html | ||||
results$validation | a html | ||||
results$finalResults | a html | ||||
results$trajectoryPlot | an image | ||||
results$meanTrajectoryPlot | an image | ||||
results$survivalPlot | an image | ||||
results$dynamicAUCPlot | an image | ||||
results$residualPlot | an image |
Details
Note: This analysis can be computationally intensive. Expected runtime: 2-10 minutes depending on data size and model complexity.
Examples
# \donttest{
# Example: PSA trajectory and prostate cancer survival
jointmodeling(
data = psa_data,
id = patient_id,
time_longitudinal = visit_time,
biomarker = psa_level,
survival_time = followup_time,
survival_status = death_status,
covariates = c("age", "stage"),
functional_form = "linear"
)
# }