Patients are able to report respiratory discomfort (dyspnoea) on a 0 to 10 scale. Patients reporting dyspnoea at admission have a 4-fold risk of dying in hospital. We asked whether assessing dyspnoea throughout hospitalisation identifies additional patients at risk of death or other poor outcomes. We conducted a retrospective cohort study of non-ICU patients at a tertiary care hospital. Bedside nurses documented patients’ pain and dyspnoea ratings each shift. We tested associations with inpatient mortality, 2-year mortality, and other outcomes. We evaluated 9785 admissions; 18% of patients reported dyspnoea at admission and 10% developed dyspnoea after admission. Patients with post-admission-onset dyspnoea had six-fold greater odds of death during the hospitalisation (OR 6.0, CI 4.2–8.5, p<0.0001) versus patients without dyspnoea. Compared with those reporting no dyspnoea, patients with dyspnoea had 50% greater mortality during the following 2 years, those with dyspnoea on the day of discharge had even greater mortality (HR 2.6 CI 2.1–3.2, p<0.0001). Rating predictions were quantitative: higher patient dyspnoea ratings predicted higher in-hospital and 2-year mortality. Pain (reported by 72% of patients) was not significantly associated with mortality. Dyspnoea was also related to greater length of stay, rapid response team activation, transfer to ICU, and discharge to extended care. In contrast to pain, both admission dyspnoea and post-admission-onset dyspnoea were associated with substantially increased odds of poor patient outcomes during the hospital stay and following discharge. Documenting dyspnoea throughout the hospital stay provides a powerful alarm of clinical compromise that may warrant additional attention and outpatient follow up.