Reply to: Reassessing the evidence linking clinical leadership to AI deployment outcomes

We thank Dr Bair for the interest in our article. Dr Bair raises three primary concerns: (1) the definition of “impact” in a field prone to publication bias, (2) the stability of our statistical modeling given the small number of non-significant results, and (3) the validity of using last authorship as a proxy for leadership. In this response to Dr Bair, we clarify the organizational science perspective underpinning our analysis of leadership in healthcare AI deployment. At the same time, we wish to reaffirm our view that the original findings represent a meaningful first step toward characterizing the role of team leadership in AI deployment in healthcare. In what follows, we (1) acknowledge explicitly that the causal language used in parts of our original paper should be interpreted with caution and clarify how we interpret the observed association; (2) engage with the mediation problem that Dr. Bair identified in relation to trial design and geography; and (3) describe the expanded contributor-level data collection currently underway in our group and outline what we believe this research needs next.