An introduction to mixed models and joint modeling: analysis of valve function over time

Area of Science:

• Biostatistics
• Cardiovascular Research
• Medical Informatics

Background:

• Clinical studies often seek prognostic biomarkers, such as risk scores, to predict disease progression.
• Current methods for evaluating biomarkers typically focus on single time points, limiting their prognostic utility.
• There is a need for robust statistical tools to assess the prognostic value of serial biomarker data.

Observation:

• This paper demonstrates statistical methodology for evaluating the predictive capability of serial echocardiographic measurements of allograft aortic valve function.
• The study introduces joint modeling of longitudinal and survival data to analyze the relationship between repeated valve function measurements and clinical outcomes.
• A prospective cohort of patients undergoing aortic valve or root replacement with an allograft valve was used for illustration.

Findings:

• The proposed joint modeling approach effectively utilizes serial echocardiographic data to predict time-to-death or time-to-reoperation.
• The methodology provides a flexible and appropriate framework for assessing the prognostic value of longitudinal biomarker measurements.
• Both optimal and suboptimal statistical methods were illustrated using real-world patient data.

Implications:

• This work offers improved statistical tools for biomarker evaluation in clinical research, particularly for longitudinal data.
• The findings can enhance the prediction of patient outcomes in cardiovascular surgery and other fields utilizing serial measurements.
• Accurate prognostic assessment using serial biomarkers can lead to more personalized and effective patient management strategies.