Background Many home fires begin in the kitchen. Kitchen smoke alarms are more likely to produce nuisance alarms, but few previous studies have examined the role of alarm sensor and battery types on the functionality of smoke alarms located nearest to the kitchen.

Methods Data were analysed from a 2×2 factorial randomised controlled trial conducted in rural Iowa homes (n=628). Enrolled households were randomly assigned into one of four smoke alarm/battery combinations: ionisation/zinc, ionisation/lithium, photoelectric/zinc and photoelectric/lithium. Alarm functionality was determined using a smoke test. Alarm type and battery type were compared using an intent-to-treat analysis. Logistic regression was used to identify factors that might impact the functionality of smoke alarms located nearest to the kitchen 42 months after installation.

Results Photoelectric alarms with lithium batteries had the highest rate of functionality (90.2%), whereas ionisation alarms with carbon/zinc batteries had the lowest (76.5%). Forty-two months following installation, 6.4% more of photoelectric alarms were functional than ionisation alarms, and 7.9% more of alarms with lithium batteries were functional than those with carbon/zinc batteries. Logistic regression revealed that when the indicator of nuisance alarms was included, the effect of alarm type became statistically insignificant and ionisation alarms were less likely to be functional at 42 months, partly due to increased nuisance alarms.

Conclusions Alarm type is an important consideration for certain locations. Photoelectric alarms may be more appropriate for installation nearest to the kitchen despite their increased cost. These findings can help guide consumer choices to increase protection against home fire-related injuries and deaths.