Skip to main content

Table 2 Sensitivity analyses identifying each case definition’s intrinsic specificity from its RERI corrected for disease misclassification

From: The effect of disease misclassification on the ability to detect a gene-environment interaction: implications of the specificity of case definitions for research on Gulf War illness

Se

Sp

Corrected RERI

By CDC definition

Se

Sp

Corrected RERI

By Modified Kansas without exclusions

1.000

1.000

6.36

(2.50–12.98)

1.000

1.000

6.80

(2.76–13.77)

1.000

0.950

5.69

(2.14–11.74)

1.000

0.950

6.09

(2.37–12.45)

1.000

0.900

5.10

(1.82–10.63)

1.000

0.900

5.45

(2.02–11.27)

1.000

0.850

4.56

(1.53- 9.64)

1.000

0.850

4.88

(1.71–10.22)

1.000

0.830

4.37

(1.42- 9.27)

1.000

0.845

4.83

(1.68–10.12)

1.000

0.820

4.27

(1.37- 9.10)

1.000

0.840

4.77

(1.65–10.02)

1.000

0.819

4.26

(1.36- 9.08)*

1.000

0.839

4.76

(1.64–10.00)*

1.000

0.800

4.09

(1.26- 8.75)

1.000

0.800

4.37

(1.42- 9.27)

1.000

0.700

3.26

(0.80- 7.23)

1.000

0.700

3.48

(0.91- 7.66)

Se

Sp

Corrected RERI

By Modified Kansas with exclusions

Se

Sp

Corrected RERI

By CDC Severe

1.000

1.000

6.46

(2.47–13.55)

1.000

1.000

7.65

(2.89–17.15)*

1.000

0.950

5.78

(2.11–12.25)

1.000

0.950

6.85

(2.46–15.47)

1.000

0.900

5.18

(1.78–11.09)

1.000

0.900

6.14

(2.06–13.99)

1.000

0.850

4.64

(1.49–10.06)

1.000

0.850

5.50

(1.70–12.67)

1.000

0.800

4.15

(1.22- 9.14)

1.000

0.800

4.92

(1.37–11.50)

1.000

0.790

4.06

(1.16- 8.96)

1.000

0.750

4.40

(1.06–10.44)

1.000

0.787

4.03

(1.15- 8.91)

1.000

0.700

3.92

(0.78- 9.49)

1.000

0.786

4.02

(1.14- 8.89)*

1.000

0.650

3.48

(0.52- 8.63)

1.000

0.785

4.01

(1.14- 8.88)

1.000

0.600

3.08

(0.28- 7.84)

Se

Sp

Corrected RERI

By Research

Se

Sp

Corrected RERI

By Research

1.000

1.000

8.88

(3.77–18.61)*

1.000

1.000

8.88

(3.77–18.61)*

1.000

0.950

7.95

(3.25–16.79)

0.950

1.000

8.88

(3.77–18.61)

1.000

0.900

7.12

(2.77–15.18)

0.900

1.000

8.88

(3.77–18.61)

1.000

0.850

6.38

(2.35–13.75)

0.850

1.000

8.88

(3.77–18.61)

1.000

0.800

5.71

(1.96–12.47)

0.800

1.000

8.88

(3.77–18.61)

1.000

0.750

5.10

(1.60–11.31)

0.750

1.000

8.88

(3.77–18.61)

1.000

0.700

4.55

(1.26–10.27)

0.700

1.000

8.88

(3.77–18.61)

1.000

0.650

4.04

(0.95- 9.32)

0.650

1.000

8.88

(3.77–18.61)

1.000

0.600

3.58

(0.67- 8.46)

0.600

1.000

8.88

(3.77–18.61)

  1. Note that the corrected RERI could not be adjusted for confounding because the corrections for disease misclassification are ecologic, not amenable to multivariable analysis at the subject level; confounding adjustment would reduce these RERIs by 1–2 points
  2. Abbreviations: Se sensitivity, Sp specificity, RERI relative excess risk due to interaction
  3. *Asterisk indicates the levels of Se and Sp where the corrected RERI equals the biased RERI. Our reference values of biased RERI by case definition are: CDC 4.26 (3.77–18.61); Modified Kansas without exclusions 4.76 (1.64–10.00); Modified Kansas with exclusions 4.02 (1.26–8.69); CDC Severe 7.80 (2.67–18.13); Research 8.88 (3.77–18.61)
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us