Skip to main content
. 2009 Feb 2;338:a3079. doi: 10.1136/bmj.a3079

Table 2.

 Univariable and multivariable Cox survival analyses for potential predictors of no recovery from major depressive disorder (n=204), measured at baseline, with follow-up for three years

Characteristics Univariable Multivariable
Hazard ratio (95% CI) P value Hazard ratio (95% CI) P value
Age (years)* 1.02 (1.00 to 1.04) 0.143
Sex (women v men) 1.06 (0.75 to 1.52) 0.729
Education: 0.325
 Middle v low 1.28 (0.88 to 1.89) 0.185
 High v low 0.95 (0.49 to 1.85) 0.878
Living alone (yes v no)* 1.32 (0.91 to 1.89) 0.145
Comorbidity (somatic or psychiatric):
 Chronic somatic diseases 0.408
 1 v 0 1.22 (0.72 to 2.04) 0.463
 ≥1 v 0 1.39 (0.85 to 2.27) 0.191
 Anxiety (yes v no)* 1.28 (0.88 to 1.85) 0.184
Cognitive dysfunction, MMSE score ≥24 v <24 1.35 (0.79 to 2.27) 0.270
Depression:
 MÅDRS score (0-60)*† 1.05 (1.03 to 1.08) <0.001
 PRIME-MD score (0-9)* 1.22 (1.06 to 1.41) <0.001 1.25 (1.08 to 1.45) 0.003
 Family history (yes v no)* 1.30 (0.88 to 1.92) 0.185 1.45 (0.97 to 2.17) 0.070
 Early onset (<55 years)* 1.30 (0.90 to 1.89) 0.158
 Treatment (yes v no) 1.10 (0.77 to 1.57) 0.617
Quality of life, PCS (0-100)* 0.99 (0.98 to 1.00) 0.055 0.98 (0.97 to 0.99) 0.008

MMSE=mini-mental state examination score; MÅDRS=Montgomery Åsberg depression rating scale; PRIME-MD=primary care evaluation of mental disorders; PCS=physical component scale of SF-36.

*P<0.20. These eight variables were associated with poor outcome in univariable analyses and therefore selected for multivariable analysis.

†MÅDRS and PRIME-MD were highly correlated. Therefore only PRIME-MD was used for the multivariable model.