The 2014-2015 influenza season was marked by circulation of antigenically drifted A/H3N2 strains, raising the possibility of low seasonal influenza Vaccine Effectiveness (VE). We assessed VE against hospitalization with laboratory-confirmed influenza for the 2014-2015 season, using routine surveillance data. Non-sentinel swab samples from Greek hospital inpatients were tested for influenza by RT-PCR in three laboratories, covering the entire country. We estimated VE using a test-negative design. Out of 883 patients with known vaccination status, 161 (18.2%) were vaccinated, and 392/883 patients (44.4%) tested positive for influenza, of whom 162 (41.3%) had type B and 151 (38.5%) had A/H3N2. Adjusted VE was 31.6% (95%CI: 2.9-51.8%) against any influenza, 46.8%, 95%CI: 12.5-67.6%) against type B and -1.9%, 95%CI: -69.5 to 38.7%) against A/H3N2. VE against non-ICU hospitalization appeared to be higher, but the difference did not reach statistical significance. Circulating A/H3N2 viruses showed substantial antigenic drift, while about half of the type B strains were similar to the vaccine strain. Despite the antigenic drift of the A/H3N2 strains, the vaccine still offered substantial protection against hospitalization with laboratory-confirmed influenza, mostly due to a surge in type B influenza late in the season. Vaccine coverage was low, even among groups targeted for vaccination, and considerable effort should be made to improve it. J. Med. Virol. 88:1896-1904, 2016. © 2016 Wiley Periodicals, Inc.