http://www.sciencedaily.com/releases/2005/12/051221091232.htm Research by Dr Lisa Jackson and colleagues at Group Health Co-operative, Seattle, found evidence of serious bias in estimates of influenza vaccine effectiveness in seniors. Their study suggests that the association between vaccination and risk of death is influenced by the fact that relatively healthy seniors (that is, those already less at risk from dying) are more likely to get vaccinated.
However, the research does not indicate that there is NO benefit in vaccinating the elderly, just that the differences in health between seniors who receive vaccine and those who do not make it difficult or impossible to tell what benefit is being derived from the vaccine.