Skip to main content

Table 8 Comparison of prevalence ratios and respective confidence interval estimates (obtained by unadjusted Poisson/Cox regression, Poisson regression with scale factor adjusted by χ2 and deviance, Poisson/Cox regression with robust variances, log-binomial regression and logistic regression) and odds ratio with confidence interval estimate (obtained by logistic regression) with the Mantel-Haenszel prevalence ratio in the analysis of the original and modified data using asthma as the outcome, maternal smoking as the risk factor and social class as confounder (situation 2).

From: Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio

Original data

Point estimate

95% Confidence interval

 

value

% diff.

lower

upper

width

% diff.

PR Mantel-Haenszel

1.22

--

1.03

1.45

0.41

--

PR Poisson/Cox (unadj)

1.22

0.0%

0.99

1.51

0.51

23.9%

PR Poisson (χ2)

1.22

0.0%

1.03

1.45

0.42

2.7%

PR Poisson (deviance)

1.22

0.0%

1.03

1.46

0.43

3.9%

PR Poisson/Cox (robust)

1.22

0.0%

1.03

1.45

0.41

-0.2%

PR log-binomial

1.23

0.1%

1.04

1.45

0.41

-0.4%

OR logistic regression

1.36

11.0%

1.05

1.76

0.71

70.9%

Modified data

Point estimate

95% Confidence interval

 

value

% diff.

lower

upper

width

% diff.

PR Mantel-Haenszel

1.62

--

1.41

1.87

0.47

--

PR Poisson/Cox (unadj)

1.62

-0.4%

1.33

1.96

0.63

36.0%

PR Poisson (χ2)

1.62

-0.4%

1.38

1.90

0.52

10.9%

PR Poisson (deviance)

1.62

-0.4%

1.39

1.88

0.49

4.3%

PR Poisson/Cox (robust)

1.62

-0.4%

1.40

1.86

0.46

-2.0%

PR log-binomial

1.65

1.7%

1.44

1.90

0.46

-1.5%

OR logistic regression

2.49

53.5%

1.90

3.27

1.37

193.1%