Skip to main content
. 2023 Feb 3;13:40. doi: 10.1038/s41398-023-02338-0

Table 2.

Hazard ratios (HRs) and 50th percentile differences (PDs, in years) of incident depression.

Time spent in outdoor light Events/total Incidence rate per 1000 person-year Model 1 Model 2 E Valuec
HRa /50th PDb (95% CI) P Value HRa /50th PDb (95% CI) P Value
Summer
2 h 2645/83466 2.59 (2.49, 2.69) 1 (Ref.) 1 (Ref.) NA
0 (Ref.) 0 (Ref.)
Below 2 h 1798/49841 2.95 (2.82, 3.09) 1.14 (1.08, 1.21)a <0.001 1.12 (1.05, 1.19)a <0.001 1.49
−0.65 (−0.95, −0.35)b <0.001 −0.54 (−0.84, −0.24)b <0.001
Above 2 h 9193/247669 3.06 (3.00, 3.12) 1.19 (1.14, 1.24)a <0.001 1.11 (1.06, 1.16)a <0.001 1.46
−0.87 (−1.08, −0.65)b <0.001 −0.49 (−0.71, −0.27)b <0.001
Winter
1 h 4133/127792 2.64 (2.56, 2.72) 1 (Ref.) 1 (Ref.) NA
0 (Ref.) 0 (Ref.)
Below 1 h 2676/72906 3.01 (2.89, 3.12) 1.14 (1.08, 1.20)a <0.001 1.07 (1.02, 1.13)a 0.005 1.34
−0.63 (−0.87, −0.38)b <0.001 −0.38 (−0.62, −0.14)b 0.002
Above 1 h 6827/180278 3.13 (3.05, 3.20) 1.19 (1.14, 1.23)a <0.001 1.09 (1.04, 1.13)a <0.001 1.40
−0.87 (−1.06, −0.67)b <0.001 −0.40 (−0.59, −0.20)b <0.001
Average
1.5 h 1539/50589 2.48 (2.36, 2.60) 1 (Ref.) 1 (Ref.) NA
0 (Ref.) 0 (Ref.)
Below 1.5 h 2469/71933 2.80 (2.69, 2.91) 1.13 (1.06, 1.21)a <0.001 1.09 (1.02, 1.16)a 0.012 1.40
−0.61 (−0.93, −0.29)b <0.001 −0.46 (−0.78, −0.14)b 0.005
Above 1.5 h 9628/258454 3.07 (3.01, 3.13) 1.24 (1.18, 1.31)a <0.001 1.13 (1.07, 1.20)a <0.001 1.51
−1.11 (−1.38, −0.84)b <0.001 −0.63 (−0.90, −0.35)b <0.001

Model 1 was unadjusted; Model 2 adjusted for age, gender, education, Townsend deprivation index, smoking status, drinking status, body mass index, total physical activity, sleep duration, fracture history, vitamin D supplement, hearing loss, use of sun/UV protection, PM2.5, hypertension, hyperglycemia, and genetic risk.

HR hazard ratio, 50th PD the difference in the median time (in years) until the first 50% of the participants, 95% CI 95% confidence interval.

aHazard ratios (HRs) and 95% confidence intervals (CIs) of incident depression by Cox models.

bThe 50th percentile difference (PDs, years) and 95% CIs of incident depression by Laplace regression models.

cThe E value is the minimum strength of association, on the risk ratio scale, that an unmeasured confounder would need to have with both the treatment and outcome, conditional on the measured covariat to explain away a treatment-outcome association.