Sepsis is a common and deadly condition, but diagnosis in not always knowable in real-time. The optimal treatment during times of diagnostic uncertainty differs across patients. Despite this reality, sepsis performance is uniformly assessed and reported for a population knowable only in retrospect—the patients ultimately judged to have sepsis at hospital discharge. This limits effective audit and feedback to incentivise clinician behaviour. Personalised, real-time assessments of a patient’s risk of death and likelihood of infection could instead be used to guide treatment recommendation and performance assessment. Clinicians and health systems could be judged on whether their responses are appropriately calibrated given the urgency of the situation. Were antibiotics prescribed at an appropriate time given the urgency of the patient’s clinical status? With the information available, were the best treatment decisions made? Did treatment plans change as new data became available? Organising treatment recommendations and performance assessment by risk of death and likelihood of infection could optimise sepsis care. For more head to: codachange.org/podcast