Fig. 2

Prediction of the PDQ-39 scales at baseline by cognitive, motor and affective baseline parameters. PDQ-39 total was predicted by the UPDRS-III PIGD subscore (β = 1.18, p = 0.003) and HADS depression (β = 0.59, p < 0.001), accounting for 30% of the variance [F(3,86) = 13.64, p < 0.001]. The PDQ-39 subscale communication was predicted by the HADS depression (β = 0.71, p = 0.008), accounting for 17% of the variance [F(2,87) = 10.05, p < 0.001]. The PDQ-39 subscale mobility was predicted by the UPDRS-III PIGD subscore (β = 3.40, p < 0.001) and HADS depression (β = 0.99, p = 0.001), which accounted for 36% of the variance [F(5,84) = 11.19, p < 0.001]. The PDQ-39 subscale ADL was predicted by the UPDRS-III total score (β = 0.68, p = 0.003) and accounted for 18% of the variance [F(4,85) = 5.93, p < 0.001]. The PDQ-39 subscale emotional well-being was predicted by the UPDRS-III PIGD subscore (β = 1.20, p = 0.02) and HADS depression (β = 0.80, p = 0.001), accounting for 25% of the variance [F(3,86) = 10.94, p < 0.001]. The PDQ-39 subscale social support was predicted by the HADS anxiety (β = 0.27, p = 0.03), explaining 4% of the variance [F(1,88) = 4.64, p = 0.03]. The PDQ-39 subscale cognition was predicted by the HADS depression (β = 0.60, p = 0.014), explaining 12% of the variance [F(2,87) = 7.26, p = 0.001]. For clarity, only significant predictors from the final regression model are displayed. Each colour of the arrows represents a separate multiple linear regression model