Clinical Collabsheets: 53 Questions to Guide a Clinical Collaboration

Shems Saleh, William Boag, Lauren Erdman, Tristan Naumann

Abstract: Clinical Machine Learning (ML) is a rapidly-growing field due to the digitization of hos- pital records, recent advances in ML techniques, and the ability to leverage increasing computational power for large and complex models. The high stakes and often unintuitive nature of clinical data make effective collaboration between clinicians and ML researchers one of the most important aspects of working in this interdisciplinary space. However, there are few resources codifying best practices for collaboration on Clinical ML projects. In this paper, we interviewed 18 experts in the Clinical ML field and distilled their ad- vice and experiences into a list of questions (a Helathcare Collabsheet) ML scientists and clinicians can use to promote effective discussion when working on a new project. We in- tend this for a broad audience as checklist of discussion points to hit at a kickoff meeting, even for experienced researchers. This resource will enable more successful partnerships in Clinical ML with improved interdisciplinary communication and organization.