Skip to main content
. 2010 Apr 15;7(3):A59.

Table 4.

Multilevel Modeling Results for Individual and Area Characteristics as Predictors of Death Without Transport to a Hospital Among People Who Died of Cardiac Events (N = 2,570), Tampa, Florida, Metropolitan Statistical Area, 1998-2002

Characteristic Null (Random Effects Only) Model, OR (95% CI) Model Including Decedent Characteristics, OR (95% CI) Model Including Decedent Characteristics and Area of Residence, OR (95% CI)
Age, y
25-49   1.83 (1.42-2.35) 1.82 (1.42-2.35)
50-59   1.45 (1.17-1.80) 1.45 (1.17-1.80)
60-64   1.34 (1.06-1.69) 1.34 (1.06-1.68)
65-69   1 [Reference] 1 [Reference]
Sex
Men   1.11 (0.92-1.34) 1.11 (0.92-1.34)
Women   1 [Reference] 1 [Reference]
Race/ethnicity
White   1.43 (1.10-1.87) 1.49 (1.15-1.94)
Black or Hispanic   1 [Reference] 1 [Reference]
Marital status
Unmarried   2.16 (1.82-2.56) 2.13 (1.79-2.52)
Married   1 [Reference] 1 [Reference]
Education level
No college degree   0.99 (0.77-1.27) 1.01 (0.79-1.31)
College degree   1 [Reference] 1 [Reference]
Area of residencea
Low population density     0.81 (0.67-1.01)
High population density     1 [Reference]
PUMA variance 0.0283 0.0118 0.0062
Total variance at PUMA level, % 0.9% 0.4% 0.2%

Abbreviations: OR, odds ratio; CI, confidence interval; PUMA, public use microdata area.

a

Based on land area, in square miles, in the PUMA from US Census Bureau's public use microdata 5% sample (14). Low was defined as 23% or more housing units on 1 or more acres, and high was defined as less than 9% of housing units on 1 or more acres.