Skip to main content

Table 3 Binary logistic regression with LS vs Non-LS as the dependent variable

From: Identifying risk factors for L’Hermitte’s sign after IMRT for head and neck cancer

Independent Variable

Logistic regression coefficient

Regression p value

Odds Ratioa (95% CI)

Original model

Age

−0.034

0.1

0.97 (0.9–1.0)

Laterality

−1.037

0.05

0.35 (0.13–1.0)

Diabetes

2.166

0.06

8.7 (0.9–83)

Hypertension

0.384

0.5

1.5 (0.5–4.3)

Cisplatin

−0.384

0.4

0.68 (0.3–1.8)

V20%

−0.007

0.7

0.99 (0.96–1.0)

V40%

0.059

0.05

1.06 (1.0–1.1)

Homogeneity Index

1.151

0.3

3.2 (0.3–30)

Constant

−0.129

  

Refined model

Age

−0.041

0.06

0.96 (0.9–1.0)

Laterality

−1.048

0.019

0.35 (0.2–0.8)

Diabetes

2.301

0.036

10 (1–85)

V40%

0.056

0.047

1.1(1.0–1.1)

Constant

0.200

  
  1. Abbreviations: CI – confidence interval Hosmer and Lemeshow test refined model χ2 (7) = 6.056, p = 0.553. Pseudo-R2 = 0.14 to 0.18. aOdds ratio per unit increase in variable or for bilateral radiation and being non-diabetic