Alarms in ambulatory monitoring of electrocardiography (ECG) sound in our cardiology ward too frequently. To investigate the incidence and causes of inadequate alarms, we examined 3,699 alarms in a total of 100 inpatients during a period of 7 days retrospectively. Of the alarms, 49% was related to extreme bradycardia, 31% to extreme tachycardia and 10% to trained ventricular premature beats. Alarms were set off because of correct detection of cardiac abnormalities (50%), poor electrode-to- patient contact (17%) and wrong diagnosis (12%), and body movement of patients (11%). To reduce false alarms in ambulatory ECG monitoring, we though it would be necessary to (1) change electrodes once a day, (2) wipe the sweat from the patient’s chest surface, (3) stick electrodes away from movement sites, (4) tape down lead wires to prevent them from pulling on the electrodes, (5) select sticking sites to get enough QRS voltage (›1mV), and (6) set adequate threshold of heart rate to make alarms.

　　We previously reported that only 49% of alarms in ambulatory electrocardiography (ECG) correctly indicated clinically abnormal electrical conditions in patients admitted to the cardiology ward and that 6 types of precautionary measures including changing the electrodes once a day and wiping sweat off the patient&rsqo;s chest skin might help to reduce false alarms. We sought to investigate the efficacy of these 6 types of measures to reduce false alarms in ambulatory ECG monitoring by comparing the number, variations, and causes of alarms before and after taking these measures. After taking these actions, the number of total alarms reduced from 3,699 to 1,109. The number of alarms indicating poor electrode-to-patient contact reduced from 629 to 30 and alarms indicating wrong diagnosis reduced from 432 to 114. However, the number of alarms indicating body movement of patients did not change (415 versus 418). Furthermore, we sought to investigate whether lowering the sensitivity of ambulatory ECG monitoring system to detect arrhythmia was useful for reducing the impact of inadequate alarms. We evaluated the incidence of missing true arrhythmias in the system and concluded that lowering the sensitivity was not a good option to reduce false alarms because of the increased risk of missing true arrhythmias.