• Carlee Lehna, Amanda Speller, Carol Hanchette, Erin Fahey, and Mary-Beth Coty.
• From the *School of Nursing and †Department of Geography and Geosciences, University of Louisville, Kentucky.
• J Burn Care Res. 2016 Jan 1; 37 (1): 12-9.

AbstractThe purposes of this study were to use geographic information systems to create a cartographic risk model predicting areas of increased potential for fire occurrences and to validate the model. Seven literature-identified risk factors associated with burn injury were older than 65 years, non-white race, below high school education, low socioeconomic status, rented housing, year home built, and home value. Geographic information system methods were used in risk factor model development. Model validation occurred using residential county fire dispatch data and statistical analysis. Areas of high and severe risk were primarily located in the northwestern and central county regions. A strong correlation (r = .66) was found between risk model scores and fire incidence rates. Significant differences in mean fire rates by risk category (F (187,3) = 87.58, P < .0001) were found, with the exception of the low and medium risk categories. Fire incidence rates among census tracts showed positive spatial autocorrelation (Moran's I = 0.542, P < .0001) producing a map showing a significant cluster of high fire incidence in the northwestern region. The risk model has potential to lead to more targeted and effective fire prevention education programs. Such models would allow fire departments to focus limited resources in areas of highest fire risk.

11 keywords...

hide…