CHAPTER 13 Taking a Closer Look at Fourfold Tables 181

LISTING 13-1:

R output from risk ratio calculation on data from Figure 13-2

> riskratio.wald(obese_HTN)

$data

Outcome

Predictor

Disease1 Disease2 Total

Exposed1

14

7

21

Exposed2

12

27

39

Total

26

34

60

$measure

risk ratio with 95% C.I.

Predictor

estimate

lower

upper

Exposed1

1.000000

NA

NA

Exposed2

2.076923

1.09512

3.938939

$p.value

two-sided

Predictor

midp.exact

fisher.exact

chi.square

Exposed1

NA

NA

NA

Exposed2

0.009518722

0.01318013

0.00744125

$correction

[1] FALSE

attr(,”method”)

[1] "Unconditional MLE & normal approximation (Wald) CI"

>

Notice that the output is organized under the following headings: $data, $measure,

$p.value, and $correction. Under the $measure section is a centered title that says

risk ratio with 95% C.I. — which is more than a hint! Under that is a table with the

following column headings: Predictor, estimate, lower, and upper. The estimate col-

umn has the risk ratio estimate (which you already calculated by hand and rounded

off to 2.17). The lower and upper columns have the confidence limits, which R

calculated as 1.09512 (round to 1.10) and 3.938939 (round to 3.94), respectively.

You may notice that because R used a slightly different SE formula than our man-

ual calculation, R’s CI was slightly wider.

Odds ratio

The odds of an event occurring is the probability of it happening divided by the

probability of it not happening. Assuming you use p to represent a probability, you

could write the odds equation this way: p

p

/ (

)

1

. In a fourfold table, you would

represent the odds of the outcome in the exposed as a/b. You would also represent

the odds of the outcome in the unexposed as c/d.