Where does learning come from? An empirical study of performance in robot-assisted surgery teams

Combining existing theories, and original fieldwork and data on 1671 robotic surgery operations performed by more than 100 surgeons within a large university polyclinic during a four-year period, we study the relation between learning and team performance. We reconstruct performance in terms of effective surgery times recorded automatically and with considerable precision by a robot surgeon operated by two human surgeons. We decompose the overall experience of the members of surgery teams in four component elements: Experiential learning (individual learning from own past experience); vicarious learning (dyadic learning from others), embedded relational learning (triadic learning from shared alters), and ecological learning (diffuse learning from accumulated collective experience). After controlling for the initial conditions of patients elected for surgery, for the skills of the surgeons, and for the factors that are specific to the kind of clinical problem that motivates surgery, we find that embedded relational learning is the most reliable predictor of the operational performance of surgery teams. This happens because embeddedness in communities of shared partners allows team members to learn from a wider and more diverse pool of experiences; facilitates coordination during surgery, and increases trust among team members thus motivating knowledge sharing within the team. We conclude with a discussion on the value of networks of relations that connect team members with knowledge and experiences developed outside of the team.