Table 2

Associations of probable depression (PHQ-9 ≥10) with stressors estimated using multivariable Poisson regression (n=2106)

Model 1Model 2Model 3Model 4Model 5Model 6
Age0.97 (0.96 to 0.98)0.97 (0.96 to 0.99)0.97 (0.96 to 0.98)0.97 (0.96 to 0.98)0.97 (0.96 to 0.98)0.97 (0.96 to 0.98)
GenderMale0.98 (0.77 to 1.23)0.97 (0.77 to 1.22)0.98 (0.78 to 1.23)1.01 (0.80 to 1.28)0.95 (0.76 to 1.20)0.97 (0.77 to 1.22)
Ethnic groupWhite(Ref.)(Ref.)(Ref.)(Ref.)(Ref.)(Ref.)
Black1.27 (0.74 to 2.18)1.13 (0.66 to 1.92)1.23 (0.73 to 2.08)1.27 (0.74 to 2.17)1.27 (0.74 to 2.20)1.28 (0.75 to 2.20)
Asian1.74 (1.28 to 2.37)1.75 (1.30 to 2.37)1.73 (1.28 to 2.34)1.73 (1.28 to 2.34)1.73 (1.27 to 2.35)1.73 (1.28 to 2.35)
Mixed0.83 (0.49 to 1.42)0.82 (0.48 to 1.41)0.84 (0.49 to 1.45)0.84 (0.49 to 1.43)0.83 (0.49 to 1.43)0.83 (0.49 to 1.41)
Other1.26 (0.75 to 2.13)1.22 (0.74 to 2.03)1.19 (0.71 to 2.02)1.24 (0.73 to 2.10)1.26 (0.75 to 2.12)1.26 (0.75 to 2.13)
Previous depression2.29 (1.79 to 2.93)2.21 (1.73 to 2.84)2.25 (1.77 to 2.86)2.25 (1.76 to 2.88)2.29 (1.79 to 2.93)2.28 (1.78 to 2.93)
Previous anxiety1.46 (1.14 to 1.88)1.45 (1.13 to 1.86)1.45 (1.14 to 1.85)1.44 (1.13 to 1.84)1.46 (1.14 to 1.87)1.46 (1.14 to 1.88)
Lives alone1.40 (1.05 to 1.86)
Housing tenureOther tenure(Ref.)
Renting1.24 (0.97 to 1.57)
Number of children0(Ref.)
10.80 (0.53 to 1.19)
20.78 (0.54 to 1.15)
3+1.41 (0.91 to 2.18)
Other caring responsibilities1.46 (1.09 to 1.95)
Access to healthcareNot needed(Ref.)
Needed, could access1.16 (0.92 to 1.48)
Needed, but could not access1.59 (1.14 to 2.22)
Long-term health condition1.18 (0.92 to 1.51)
Currently shielding/isolating1.74 (0.90 to 3.38)
Key worker0.91 (0.73 to 1.13)
Started working remotely during pandemic0.92 (0.76 to 1.13)
Job roleAcademic, specialist and management(Ref.)
Research, clerical and technical1.00 (0.78 to 1.27)
Teaching, facilities and clinical0.85 (0.56 to 1.29)
Contract typeOpen-ended/permanent(Ref.)
Fixed term/casual1.04 (0.82 to 1.32)
Thinks had COVID-19, past 2 months1.05 (0.83 to 1.33)
  • Estimates are reported as relative risks (95% CIs).

  • All estimates weighted to account for differences between sample and population in terms of age group, gender and ethnic group. Missing outcome and covariate information imputed using multiple imputations.

  • PHQ, Patient Health Questionnaire.