Healthcare-associated infections are often associated with poor hand hygiene (HH) by healthcare workers (HCWs). The objective of this cross-sectional study at the Umberto I teaching hospital in Rome was to quantify compliance with HH by direct observation following a multimodal strategy devised by the World Health Organisation and to map critical areas for improvement. Predictors of HH compliance were identified using a multivariable logistic regression model. Adjusted odds ratios (aORs) and 95% confidence intervals (CIs) were calculated. Eighty-four trained observers from 50 wards collected 4,081 observations showing that overall HH compliance was 71.9%. The multivariable analysis found a positive association with the outcome for midwives compared to physicians (aOR=2.5, 95% CI: 1.5-4.1), and a negative association for healthcare assistants (aOR=0.5, 95% CI: 0.3-0.8). There was greater compliance during public holidays and weekends (aOR=1.5, 95% CI: 1.1-2.1), but compliance was lower for external staff (aOR=0.7, 95% CI: 0.5-0.9). We found a positive association with all HH indications after interaction with a patient or with patient surroundings compared with the indication "before touching a patient" (all p<0.001); the highest association was with the indication "after contact with biological fluids" (aOR=7.7, 95% CI: 4.7-12.5). Overall, we observed reasonable compliance levels, but it is important to increase adherence to HH practice and monitor any behaviour change.