The authors aimed to determine whether the Gait Deviation Index (GDI) could be feasible to characterize gait in patients with Parkinson's disease (PD) and evaluate outcomes of levodopa treatment. Twenty-two PD participants were evaluated with clinical examination and 3-D quantitative gait analysis (GDI was calculated from gait analysis) in 2 states (OFF and ON) after taking levodopa. Twenty age-matched healthy participants (CG) were included as controls. The GDI value in the OFF state was 83.4 ± 11.5 (statistically different from CG) while clinical scales demonstrated a moderate-severe gait impairment of these patients. Significant improvements are evident from clinical scores and by GDI values in the ON state. The mean GDI for the ON state (GDI(ON): 87.9 ± 10.4) was significantly higher than in for the OFF state (GDI(OFF): 83.4 ± 11.5), indicating a global gait improvement after the treatment. The results show that GDI has lower value as an indicator of pathology in PD patients than in quantifying the effects of levodopa treatment in PD state.