False physiologic monitor alarms are extremely common in the hospital environment. High false alarm rates have the potential to lead to alarm fatigue, leading nurses to delay their responses to alarms, ignore alarms, or disable them entirely. Recent evidence from the U.S. Food and Drug Administration (FDA) and The Joint Commission has demonstrated a link between alarm fatigue and patient deaths. Yet, very little scientific effort has focused on the rigorous quantitative measurement of alarms and responses in the hospital setting. We developed a system using multiple temporarily mounted, minimally obtrusive video cameras in hospitalized patients' rooms to characterize physiologic monitor alarms and nurse responses as a proxy for alarm fatigue. This allowed us to efficiently categorize each alarm's cause, technical validity, actionable characteristics, and determine the nurse's response time. We describe and illustrate the methods we used to acquire the video, synchronize and process the video, manage the large digital files, integrate the video with data from the physiologic monitor alarm network, archive the video to secure servers, and perform expert review and annotation using alarm "bookmarks." We discuss the technical and logistical challenges we encountered, including the root causes of hardware failures as well as issues with consent, confidentiality, protection of the video from litigation, and Hawthorne-like effects. The description of this video method may be useful to multidisciplinary teams interested in evaluating physiologic monitor alarms and alarm responses to better characterize alarm fatigue and other patient safety issues in clinical settings.