Deceased donor organs are an absolutely scare healthcare resource, meaning demand vastly exceeds a fixed supply with a hard limit. Allocation of donor organs requires difficult choices between thousands of waiting candidates who would benefit from transplantation. But allocation does not occur in a vacuum- transplant programs link donor organs and needy patients. Understanding how allocation rules shape their treatment and selection practices is key to designing effective systems. This three manuscript dissertation explores the relationship between transplant center practices, policy, and lives saved in heart and kidney allocation. In the introductory Chapter 1, we outline the broader issue of the allocation of absolutely scarce healthcare resources and discuss the current ethical framework for deceased donor organ allocation in the U.S. In Chapter 2, we create a novel mixed-effects model of the survival benefit of heart allocation, the key empirical outcome for the assessment of a transplant allocation system. We find that 5-year survival benefit associated with heart transplant varied across transplant centers, and high survival benefit centers performed heart transplant for patients with lower estimated waiting list survival without transplant. In Chapter 3, we describe how a new heart allocation policy was associated with widespread shifts in transplant center practice that threaten to undermine the effectiveness of the new system. Finally, in Chapter 4 we apply and extend the model from Chapter 2 to deceased donor kidney allocation. We develop a tool that can 1) improve shared decision-making between transplant programs and patients when making accept-reject decisions and 2) identify medically urgent transplantation candidates to align kidney allocation policy with federal law. Incorporating survival benefit into deceased donor organ allocation could save more lives and reduce racial/ethnic disparities in transplantation.