Objective Address the alarm problem by redesigning, reorganizing, and reprioritizing to better discriminate alarm sounds and displays in a hospital. Background Alarms in hospitals are frequently misunderstood, disregarded, and overridden. Method Discovery-oriented, intervention, and translational studies were conducted. Study objectives and measures varied, but had the shared goals of increasing positive predictive value (PPV) of critical alarms by reducing low-PPV alarms in the background, prioritizing alarms, redesigning alarm sounds to increase information content, and transparently conveying who initiated alarms. An alarm ontology was iteratively generated and refined until consensus was achieved. Results The ontology distinguishes five levels of urgency that incorporate likely PPV, three categories for who initiates the alarm (hospital staff, patient, or machine), whether it is clinical or technical, and clinical functions. Conclusion This unique collaboration allowed us to make progress on the alarm problem by making unintuitive leaps, avoiding common missteps, and refuting conventional healthcare approaches. Application Hospitals can consistently redesign, reorganize, reprioritize, and better discriminate alarms by priority, PPV, and content to reduce nurse response times.