Fifty-one sounds were presented to sixty participants in an online sound categorization study. The 4-second stimuli consisted in everyday sounds which were strongly or remotely associated with the seven clinical alarm risk categories described in IEC 6060-1-1-8 (Drug Administration, Perfusion, Oxygen, Temperature, Cardiac, Power Down and Ventilation). One additional category was added for the ?Blood Pressure? sounds. After a headphone screening test, participants had to categorize all sounds into one of the eight categories (or a no-category group) and indicate the quality of fit of their evaluation (Poor, Fair, Perfect). Half of the participants were healthcare professionals with experience with medical devices, and the other half were not healthcare professionals and had never interacted with medical devices. The categorization exercise was adapted from a classic usability methodology called Card Sorting, generally used for information architecture. With the purpose of designing usable sounds, we have successfully implemented the same procedure for the auditory modality. The results demonstrated a strong consistency within and between groups for the categories of Blood Pressure, Cardiac, Power Down and Ventilation. Lower consistency was observed in the other categories, revealing the difficulty of sound design for specific contexts. Having evidence of agreement among participants is an important output for the general goal of designing a library of informative sounds for medical devices.