Conclusion
Liver transplant centers learn by doing, but only shortly after entry. The importance of learning-by-doing varies among centers with some centers starting at a low survival rate and improving, while others enter with high survival rates. The effect of learning-by-doing increases in contexts with greater technical uncertainty, as measured by differences in the timing of entry, across survival durations, and across types of organ transplants.
These results highlight the importance of focusing on learning-by-doing shortly after entry rather than seeking to identify a constant return to experience, regardless of how experienced an organization is when it first appears in the data. The pattern of early learning-by-doing I identify follows what has been found in many studies in manufacturing, indicating that organizational learning-by-doing is not a phenomenon unique to reducing unit costs or production defects. Organizational learning-by-doing seems to operate quite similarly in complex industries with team-based production processes, such as organ transplants, shipbuilding (Thornton and Thompson 2001), semi-conductors (Irwin and Klenow 1994), air craft (Benkard 2000), auto manufacturing (Levitt et al. 2013), and in healthcare, for minimally invasive cardiac surgery (Pisano et al. 2001) despite vastly different information dissemination mechanisms from proprietary manufacturing practices to medical research published in a highly competitive academic environment.