Restricted mean models for transplant benefit and urgency

Area of Science:

• Medical Statistics
• Transplantation Science
• Health Services Research

Background:

• The Lung Allocation Score (LAS) is vital for prioritizing lung transplant candidates in the US.
• Accurate estimation of survival on the waitlist and post-transplant is essential for LAS calculations.
• Patient urgency and transplant benefit are key components of the LAS.

Purpose of the Study:

• To develop a statistical model for estimating restricted mean survival time that accounts for dependent censoring.
• To improve the accuracy of transplant benefit and urgency estimations within the LAS framework.
• To assess the impact of accounting for dependent censoring on lung transplant priority.

Main Methods:

• Developed a novel statistical model using pseudo-observations to handle dependent censoring.
• Incorporated longitudinal measures, such as LAS, into the survival modeling.
• Validated the model's performance through simulations comparable to the LAS setting.

Main Results:

• The proposed method accurately estimates restricted mean survival time, even with dependent censoring.
• Accounting for dependent censoring resulted in larger estimated transplant benefits for many high-urgency patients.
• Adjusting for dependent censoring significantly impacts LAS, affecting organ offer priority and timing.

Conclusions:

• The developed statistical approach provides a more accurate way to model survival in lung transplant candidates.
• Improved LAS estimations through this method can lead to more equitable organ allocation.
• This methodology has the potential to optimize lung transplant outcomes by refining patient prioritization.