Effectiveness of a Voice Smoke Alarm Using the Child's Name for Sleeping Children: A Randomized Trial

To test maternal voice alarm effectiveness under residential conditions and determine whether personalizing the maternal voice alarm message with the child's first name improves effectiveness.

Using a randomized, nonblinded, repeated measures design, we compared 3 maternal voice smoke alarms with respect to their ability to awaken 176 children 5-12 years old from stage 4 slow-wave sleep and prompt their performance of an escape procedure. A conventional residential high-frequency tone smoke alarm was used as a comparative reference. Children's sleep stage was monitored in a residence-like research setting.

Maternal voice alarms awakened 86%-91% of children and prompted 84%-86% to escape compared with 53% awakened and 51% escaped for the tone alarm. A sleeping child was 2.9-3.4 times more likely to be awakened by each of the 3 voice alarms than the tone alarm. The median time to awaken was 156 seconds for the tone alarm and 2 seconds for each voice alarm. The proportions of children who awakened and escaped differed significantly between the tone alarm and each voice alarm, but no significant differences were found between each pair of the voice alarms, regardless of whether the child's first name was included in the alarm message.

The maternal voice alarms significantly outperformed the tone alarm under residential conditions. Personalizing the alarm message with the child's first name did not increase alarm effectiveness. These findings have important implications for development of an effective and practical smoke alarm for children.

ClinicalTrials.gov: NCT01169155.