Traditional auditory evacuation cues, such as fire alarms, frequently fail to capture individuals’ attention, particularly when they are using handheld devices. This study investigates the effectiveness of attention shift interventions delivered via emergency text messages displayed on these devices to improve evacuation responses. Participants in a virtual reality simulation of a metro station were exposed to emergency cues through auditory broadcasts or visual text messages. The findings show that visual attention shift interventions reduce the likelihood of re-engaging in pre-evacuation activities while improving evacuation performance when compared to auditory cues. Continuous visual cues were more effective than auditory cues in keeping attention during evacuation. These results highlight the potential for personalized visual communication to improve emergency response strategies in environments where handheld devices are widely used, providing critical insights into developing more effective emergency communication systems for improved evacuation behavior and overall safety during emergencies.