Estimates show that 80 per cent to 99 per cent of alarms set off in hospital units are false or clinically insignificant, representing a cacophony of sounds that do not present a real danger to patients. These false alarms can lead to an alert overload that causes a health care provider to miss important events that could be harmful or even life-threatening. As health care units become more dependent on monitoring devices for patient care purposes, the alarm fatigue issue has to be addressed as a major concern in order to prevent healthcare providers from undergoing alarm burden, as well as to increase patient safety. The main goal of this thesis is to propose a solution for the alarm fatigue problem by using an automatic reasoning mechanism to decide how to notify members of the health care team. Our specific goals are: to reduce the number of notifications sent to caregivers; to detect false alarms based on alarm-context information; to decide the best caregiver to whom a notification should be assigned. This thesis describes: a model to support reasoning algorithms that decide how to notify caregivers in order to avoid alarm fatigue; an architecture for health systems that supports patient monitoring, reasoning and notification capabilities; and three reasoning algorithms that decide: (i) how to notify caregivers by deciding whether to aggregate a group of alarms; (ii) whether, or not, to notify caregivers with an indication of a false alarm probability; (iii) who is the best caregiver to notify considering a group of caregivers. Experiments were used to demonstrate that by providing a reasoning system that aggregates alarms we can reduce the total of notifications received by the caregivers by up to 99.3 per cent of the total alarms generated. These experiments were evaluated through the use of a dataset comprising real patient monitoring data and vital signs recorded during 32 surgical cases where patients underwent anesthesia at the Royal Adelaide Hospital. We present the results of this algorithm by using graphs generated with the R language, which show whether the algorithm decided to deliver an alarm immediately or after a given delay. We also achieved the expected results for our reasoning algorithm that handles the notifications assignment task, since the algorithm prioritized the caregiver that was available and was the most experienced and capable of attending to the notification. The experimental results strongly suggest that our reasoning algorithms are a useful strategy to avoid alarm fatigue. Although we evaluated our algorithms in an experimental environment, we tried to reproduce the context of a clinical environment by using real-world patient data. As future work, we aim to evaluate our algorithms using more realistic clinical conditions by increasing, for example, the number of patients, monitoring parameters, and types of alarm.