Skip to main content
. 2022 Dec 6;13:7129. doi: 10.1038/s41467-022-34504-1

Table 2.

Results of linear regression models with depressive symptom scores as the dependent variable

A. Model summary
R2 ΔR2 F change df1 df2 p
Unadjusted 0.0144 44.31 1 3033 <0.0001
Model 1 0.0754 0.0610 24.95 8 3025 <0.0001
Model 2 0.0841 0.0087 7.19 4 3021 <0.0001
Model 3 0.1109 0.0267 18.13 5 3016 <0.0001
B. Overview of coefficients
Dependent: Unadjusted Model 1 Model 2 Model 3
PHQ-9 sum score B t p B t p B t p B t p
Shannon −0.1200 −6.66 0.0000 −0.0729 −3.84 0.0001 −0.0602 −3.15 0.0016 −0.0422 −2.22 0.0263
Gender (Women = 1) 0.1154 6.52 0.0000 0.1266 6.82 0.0000 0.1186 6.45 0.0000
Age −0.0388 −2.06 0.0397 −0.0406 −2.13 0.0333 −0.0656 −3.38 0.0007
Education −0.0426 −2.10 0.0360 −0.0346 −1.68 0.0922 −0.0227 −1.12 0.2627
Smoker (Yes = 1) 0.0380 2.04 0.0410 0.0395 2.15 0.0314
Physical activity (min/wk*intensity) −0.0440 −2.49 0.0127 −0.0371 −2.13 0.0333
Audit alcohol (sum) 0.0614 2.99 0.0028 0.0588 2.90 0.0037
BMI (kg/m2) 0.0519 2.74 0.0062 0.0444 2.35 0.0188
GI disorder (yes = 1) 0.1027 5.85 0.0000
Diabetes (yes = 1) 0.0624 3.36 0.0008
Diarrhea (yes = 1) 0.0799 4.57 0.0000
PPI (yes = 1) 0.0394 2.19 0.0284
Antibiotics (yes = 1) 0.0264 1.53 0.1267

Tables present a model summary (A) and an overview of standardized regression coefficients (β) of predictor and covariates. Unstandardized values and constants are omitted for clarity. Models 1 to 3 also include ethnicity. Coefficients for ethnicity are reported separately in the text.