|Multiple regression models examining associations with a clinically significant HADS score|
|Model 1: Multiple logistic regression||Model 2: Multi level multiple logistic regression|
|Odds ratio||95% CI||P value||Odds ratio||95% CI||P value|
|Male sex||0.54||0.29, 1.01||0.054||0.52||0.26, 1.03||0.062|
|Married (vs single/ divorced)||0.42||0.19, 0.92||0.031||0.40||0.17, 0.96||0.040|
|Third level education or higher||0.31||0.14, 0.71||0.006||0.31||0.13, 0.74||0.009|
|First degree relative to proband||2.91||1.26, 6.73||0.013||3.18||1.22, 8.31||0.018|
Footnote. Multivariable models adjusted for age. Model 1 is a multivariable logistic regression, with a clinically significant HADS score as the outcome variable and adjusted for age. Model 2 is a multivariable logistic regression model with a random effects variable for family and adjusted for age. Log likelihood for model 1 is −130.45, and −130.05 for model 2. For the random effects model 2, the random intercept (ψ) is 0.253, and the estimated intraclass correlation of the latent responses (ρ) is 0.071.
McGorrian et al.
McGorrian et al. BMC Medical Genetics 2013 14:1 doi:10.1186/1471-2350-14-1