What question did this study set out to answer?

To develop and evaluate a generalized Cox regression framework that simultaneously estimates non-linear covariate shapes and time-dependent hazard effects on survival.

February 14, 2006Statistics in Medicine

Joint estimation of time‐dependent and non‐linear effects of continuous covariates on survival

View Full Paper

Ask AI

Bookmark

To develop and evaluate a generalized Cox regression framework that simultaneously estimates non-linear covariate shapes and time-dependent hazard effects on survival.
Formulated a product hazard model, lambda(t;x) = lambda(0)(t)exp[beta(t)r(x)], representing time-varying beta(t) and non-linear r(x) components via regression splines.
Estimated parameters simultaneously using maximum partial likelihood and constructed likelihood ratio tests for model comparisons.
Assessed functional shape recovery and error rates through simulation studies and demonstrated practical utility in cancer epidemiology data.
Simulations demonstrated reasonably accurate recovery of both functional forms and well-calibrated likelihood ratio test sizes when the model is correctly specified.
Model misspecification generated substantially biased risk estimates and highly inflated type I error rates.
Application to cancer epidemiology demonstrated that joint estimation uncovers nuanced, time-varying prognostic relationships missed by conventional models.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Statistical modeling study reveals accurate joint recovery of time-dependent and non-linear risks in survival analysis, indicating standard single-effect models cause substantial estimation bias.

Cite This Study

Abrahamowicz et al. (2006) studied this question.

synapsesocial.com/papers/6a7152adc1a24a6142dbae78 https://doi.org/10.1002/sim.2519

View Full Paper

Ask AI

Bookmark

Implication

Statistical modeling study reveals accurate joint recovery of time-dependent and non-linear risks in survival analysis, indicating standard single-effect models cause substantial estimation bias.

Joint estimation of time‐dependent and non‐linear effects of continuous covariates on survival

Key Points

Discussion

Implication

Cite This Study

Joint estimation of time‐dependent and non‐linear effects of continuous covariates on survival

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider