In the U.S., donor lungs are allocated to recipients based on a lung allocation score (LAS). While the statistical models used to construct the LAS control for patients’ demographic and clinical values, they do not account for selection bias, which arises because: (1) individuals are removed from the waitlist once they receive transplant (dependent censoring), and (2) in order to receive transplant, individuals must survive on the waitlist long enough for a suitable lung to become available (survivor bias). Failure to account for selection bias can lead to inaccurate predicted probabilities and suboptimal organ allocation. The goal of this dissertation is to improve the predictive accuracy of the LAS by mitigating selection bias so that lungs are allocated to the appropriate patients in the appropriate order. This goal was accomplished via three aims. First, we proposed a weighted estimation strategy to mitigate selection bias in the pre- and post-transplant LAS models, constructed a modified LAS score using these weights, and compared its performance to that of the existing LAS. Second, we examined the clinical impact of our modified LAS in both observed data and through simulations. Third, we conducted qualitative semi-structured interviews with lung transplant surgeons and pulmonologists throughout the U.S. to examine respondents’ understanding of selection bias and how it may affect the LAS and organ distribution. We found that our modified LAS exhibited better discrimination and calibration than the existing LAS and led to changes in patient prioritization. Diagnosis group, six-minute walk distance, continuous mechanical ventilation, functional status, and age exhibited the largest impact on prioritization changes. Simulations suggest that one-year waitlist survival may improve under the modified LAS, while one-year post-transplant and overall survival remain comparable to that under the existing LAS. Finally, our qualitative study demonstrates that selection bias can arise at several points along the transplantation pathway. To address such bias, transplant centers must consider both patient health and program health within constraints imposed by donor organ scarcity. We hope that this work can inform future revisions of the LAS and other prediction models in organ transplantation to ensure more equitable allocation of donor organs.