Skip to main content

Table 1 Eight scenarios depicting cross-sectional studies carried out in different underlying conditions, yet all uncovering prevalence ratios between exposed and non-exposed groups of 2.0.

From: Measures and models for causal inference in cross-sectional studies: arguments for the appropriateness of the prevalence odds ratio and related logistic regression

Scenario

Type of outcome event

Duration of outcome ( T ¯ i )

IDR*

CIR**

1

Rare

Long and equal

2.053

(+2.6)

2.025

(+1.2)

2

Rare

Long and unequal

0.205

(-89.7)

0.227

(-88.6)

3

Rare

Short and equal

2.053

(+2.6)

1.809

(-9.5)

4

Rare

Short and unequal

0.205

(-89.7)

0.443

(-77.8)

5

Common

Long and equal

3.000

(+50.0)

2.230

(+11.5)

6

Common

Long and unequal

0.300

(-85.0)

0.656

(-67.2)

7

Common

Short and equal

3.000

(+50.0)

1.037

(-48.1)

8

Common

Short and unequal

0.300

(-85.0)

1.000

(-50.0)

  1. * In brackets: % bias (IDR) calculated as [-(PR - IDR )/PR]*100, where PR = 2.0 is fixed.
  2. ** In brackets: % bias (CIR) calculated as [-(PR - CIR )/PR]*100, where PR = 2.0 is fixed.
  3. Time units used in the scenarios
  4. Average duration of outcome long and equal → T 1 = T 0 = 1
  5. Average duration of outcome long and unequal → T 1 = 1; T 0 = 0.1
  6. Average duration of outcome short and equal → T 1 = T 0 = 0.1
  7. Average duration of outcome short and unequal → T 1 = 0.1; T 0 = 0.01
  8. Time interval for projecting CIR → Δt = 1
  9. Prevalences according to exposure group and outcome frequency
  10. Rare outcome → P 1 = 0.05; P 0 = 0.025
  11. Common outcome → P 1 = 0.5; P 0 = 0.25