To assess the level of alarm fatigue among intensive care unit (ICU), cardiac care unit (CCU) and emergency room (ER) nurses, identify associated demographic and occupational factors, determine the most frequent sources of alarms and evaluate nurses’ psychological reactions to alarms.

A cross-sectional, descriptive–analytical study.

ICUs, CCUs and ERs of six public teaching hospitals affiliated with Tehran and Kashan Universities of Medical Sciences in Iran.

Using a multistage stratified random sampling method, 285 nurses were approached, of whom 260 completed and returned the questionnaires (response rate: 91%). Participants were registered nurses with at least a bachelor’s degree or higher and 3 months of experience in ICUs, CCUs or ERs.

The primary outcome was the level of alarm fatigue measured using the validated Nurses’ Alarm Fatigue Questionnaire. Secondary outcomes included factors associated with alarm fatigue and nurses’ reported psychological responses to frequent alarms.

The mean score of alarm fatigue was 26.4±7.9, indicating a moderate level. After adjusting for confounders and hospital-level clustering using multivariable mixed-effects regression, higher monthly income was significantly associated with lower alarm fatigue (β=–0.15, p=0.03), and nurses working rotational shifts reported significantly higher fatigue compared with those with fixed shifts (β=0.18, p=0.02). Other demographic and occupational factors were not significant. Reported psychological reactions to alarms included indifference (14%), irritability (18%) and anxiety/stress (15%).

ICU, CCU and ER nurses experience a moderate level of alarm fatigue, with income and shift type as independent associated factors. The association between income and alarm fatigue may reflect the role of financial stress as an additional job demand that compounds the burden of frequent alarms, particularly in contexts where low base salaries lead nurses to rely on overtime and multiple shifts. These findings underscore the need for targeted managerial and educational interventions, including shift schedule optimisation and attention to workload-related stressors, alongside alarm prioritisation strategies. Due to the cross-sectional design, causal inferences cannot be drawn.