In January 2005, Australia became the first country to introduce a publicly funded pneumococcal vaccination program for persons 65 years and older which is free at point of service, although the vaccine cost had previously been partially subsidized. Hospitalization in this age group is an important indicator of risk of invasive pneumococcal disease but vaccine uptake has been suboptimal. To determine vaccination rates and predictors of vaccination in the elderly hospitalised patients before and after January 2005. We validated vaccination status against general practitioner (GP) records for patients aged > or = 65 years admitted to a large teaching hospital in Sydney between 16th of May 2005 and the 20th of February 2006 and examined predictors of vaccination. Commencement of the new program resulted in a significant increase in vaccination uptake from 39% of inpatients prior to the free program to 73% in the same cohort of inpatients post January 2005. We found that patient recall of vaccination status was not reliable. Self-report of pneumococcal vaccination had a sensitivity of 0.53 and a specificity of 0.55, highlighting that validation of vaccination status is required. Age over 80 years and dementia significantly predicted under-vaccination. This highlights the importance of integrating free vaccine supply and delivery in primary care to achieve high vaccination coverage. However, demented patients and the very elderly remain under-vaccinated, despite being admitted to hospital for active management of acute conditions.