Fire fatalities among children: an analysis across Philadelphia's census tracts

OBJECTIVE: This study investigates the possible causes of high levels of residential fire deaths to children younger than 15 years of age in Philadelphia from 1989 to 2000. METHODS: The authors analyzed 246 deaths from 146 residential fires by census tract in Philadelphia using both individual level data and aggregate level data drawn from the records of the Fire Marshall's Office. Death rates by age and sex were calculated using the 1990 Census. Data on fires from official records were combined with aggregate level data by census tract from the 1990 Census and analyzed using logistic regression. Newspaper articles on the fires analyzed were used to identify residences with possible fire code violations. The authors used data from the Philadelphia Bureau of Licenses and Inspections to provide evidence of code violations. RESULTS: The statistically significant variables that resulted from the logistic regression were census tracts in the highest quartile for low income households, census tracts in the highest quartile for single-parent households with children younger than age 18, census tracts in the highest quartile for houses built before 1939, and the number of children younger than 15 years of age in a census tract. CONCLUSION: Population characteristics by census tract are useful in identifying risk factors for residential fire deaths of children. Census tracts identified as at highest risk can provide fire prevention units with opportunities to take preventative measures such as the distribution of smoke detectors, and the education of residents about the dangers of careless smoking and planning for the rescue of children in the event of a fire.