Estimating Smoke Alarm Effectiveness in Homes

Abstract

This paper sets out to answer two questions. What effect does installing smoke alarms have on reported fires and casualties for the “average” household? How much of an effect would increasing smoke alarm utilization have on the total number of fires and casualties in the United States? The installation of smoke alarms in homes without them will reduce the expected number of fires reported from a (formerly) non-smoke-alarm residence by a factor of 3.5 to 5 (a 70 to 80% reduction) and reduces the number of expected casualties by a factor of 2.5 to 3.5 (a 60 to 70% reduction). Unexpectedly, the number of casualties per reported fire is lower for non-smoke-alarm households compared to smoke-alarm households. This could be due to changes in people’s behavior when they have a smoke alarm in the home or because the less dangerous fires are preferentially extinguished when smoke alarms are present. If smoke alarms were installed in all residences, the number of fires reported to the fire department could be reduced by 25% or more. More realistically, each percent increase in smoke alarm penetration reduces reported fires by more than 2.6%. There is evidence that the number of homes with smoke alarms are lower than current estimates that rely on telephone-only survey interviews. The actual smoke alarm usage is probably below 92%. It seems likely that this is due to a correlation between phone presence and smoke alarm utilization.