OBJECTIVES: Liver diseases (LDs) can reduce health-related quality-of-life (HRQoL), with an important impact on the burden of LDs. Our aim was to analyze the impact of the major LDs on EQ-5D Visual Analogue Scale (VAS) and utility-index (UI) through different regression models, using HRQoL of the general population as a reference. METHODS: HRQoL data were measured using the EQ-5D-3L in a sample of patients with 9 major LDs enrolled during 2011-2012 in a multicenter study conducted in the most populated region of Italy, Lombardy. Such data were added to those recorded in 2013 on a representative sample of Lombardy general population. Relationships between the outcomes of interest (VAS and UI) and LDs were explored through ordinary least squares (OLS) and Tobit regression, that accounts for ceiling effect, adjusting for age and gender. Goodness-of-fit was assessed through R (OLS) and pseudo-R(Tobit). RESULTS: The sample included 9,817 subjects (3,017 with LDs). OLS and Tobit regressions performed similarly on VAS (R:0.13,pseudo-R:0.12) and estimated the largest and significant HRQoL reduction in patients with decompensated cirrhosis (DC), followed by those with autoimmune hepatitis (AIH) and hepatocellular carcinoma (HCC). As for UI, the Tobit model performed better than OLS regression (R:0.11,pseudo-R:0.15), with the highest decrease estimated in patients with DC, followed by those with HCC and compensated cirrhosis. CONCLUSIONS: The Tobit model performed slightly better than OLS regression on the UI, but not on VAS, maybe due to a stronger ceiling effect in UI. This distributional difference mirrors different data-generation mechanisms: UI is derived from the EQ-5D-3L domains, while VAS is reported by patients. Therefore, VAS and UI might capture different aspects of HRQoL, as supported by our results, that show how the same LD can be differently associated with VAS and UI (e.g.: AIH seemed to have an impact on VAS only, while HCC on UI only).