Subjects may have changed their aspirin use after
the baseline interview; however, this is unlikely to be a
problem in the present study for the following reasons.
Half of the patients were diagnosed within 8 years after
the baseline interview, three-fourths within 11 years,
and because the latency period for esophageal cancer is
likely to exceed 11 years, the majority of patients in this
study should have had tumors that were induced before
the baseline interview. Thus, even if the aspirin use of
the subjects changed over time, this study is based on
those measures of aspirin use most relevant to the etiology
of esophageal cancer.