Abstract
<sec> <title>UNSTRUCTURED</title> <p>Documentation burden for clinicians is persistent despite initiatives to reduce it. We present the concept of information infrastructures as a way of understanding this issue. Recognizing the electronic patient record (EPR) as an information infrastructure identifies key factors underlying documentation burden: the multiplicity of documentation requirements from non-clinical EPR users, including managers, regulators, and artificial intelligence (AI) tools; evolution of requirements over time and requirements due to long-term data uses; clarity required when sharing documentation across organizational boundaries; and limited electronic efficiencies in EPR design due to paper-era logics. Reframing documentation burden as an issue of information infrastructure suggests designers should challenge inherited assumptions within the EPR and develop innovative ways of reconciling requirements of clinical and non-clinical users. Policymakers should recognize documentation burden risks associated with AI. Researchers wanting to understand causes of documentation burden should undertake ethnographies that extend beyond the hospital ward.</p> </sec>