Two-stage model for multivariate longitudinal and survival data with application to nephrology research

Area of Science:

• Biostatistics
• Nephrology Research
• Longitudinal Data Analysis

Background:

• Follow-up studies often collect diverse outcomes like longitudinal measurements and time-to-event data.
• Investigating associations between multiple longitudinal biomarkers and survival is crucial but complex.
• Existing joint modeling often focuses on a single longitudinal outcome and survival.

Purpose of the Study:

• To propose a two-stage model-based approach for analyzing multivariate longitudinal and survival data.
• To investigate the complex associations between multiple biomarkers (calcium, phosphate, parathormone, creatinine) and patient survival.
• To provide new insights for nephrology research, specifically within the Peritoneal Dialysis Programme.

Main Methods:

• Developed a two-stage model-based approach for multivariate longitudinal and survival data.
• Applied the model to data from the Peritoneal Dialysis Programme at CHP, Portugal.
• Compared the multivariate approach with models analyzing each biomarker individually.

Main Results:

• The multivariate model successfully studied the complex association structure between multiple biomarkers and patient survival.
• The proposed approach yielded valid results, offering new insights into nephrology research.
• Demonstrated the advantage of analyzing multiple biomarkers simultaneously over separate analyses.

Conclusions:

• The multivariate longitudinal and survival model is effective for complex association structures in nephrology.
• This approach provides a more comprehensive understanding of biomarker-survival relationships than traditional methods.
• The findings offer valuable insights for patient survival analysis in peritoneal dialysis programs.