They're either prepared or they're not! Prevalence and risk factors for the presence of carbon monoxide alarms in Maine households

Introduction: The proper use of CO alarms can dramatically reduce morbidity and mortality from carbon monoxide (CO) exposure. However, little is published about either the prevalence or predictors for not having CO alarms in US households. Methods: In 2004, the Maine Behavioral Risk Factor Surveillance System (BRFSS) survey included 3 questions about CO alarms in households. We estimated the household prevalence of CO alarms and assessed personal, household and emergency preparedness factors associated with the absence of an alarm using bivariate analysis (Rao-Scott Chi-square tests, 95% confidence intervals and odds ratios). For logistic regression we accommodated both personal and household-level factors by constructing combined variables such as marital status and household number of children. Results: There were 3,304 respondents; 64.7% (95% CI: 63.0-66.4) of Maine households had no CO alarm. Characteristics statistically significantly associated with no household alarm included: Household: > 1 child in the household (69.2% vs. 54.9%), Female head of household with household income <$0,000 compared > 1 adult with household income > $50,000 (72.6% vs. <66.4%); Personal: Hispanic ethnicity (80% vs. 64.4% ); Age > 65 years (72.0% vs. 62.5%), Having a high school education or less (67.9% vs. 62.3%); Emergency preparedness: Not owning a generator (67.1% vs. 56.0%), Having none vs. all of a 3-day food and water supply and battery-powered radio (80.4% vs. 58.2%). Conclusion: The prevalence of CO alarms is strikingly low -- widespread education, prevention and distribution programs should be initiated with an emphasis on the vulnerable populations identified as most at risk.