We classified subjects as regular users of aspirin if they reported both using aspirin in the
month before baseline and regularly using pain medications
during the previous 6 months. We used information
about subjects' general use of pain medications
in classifying subjects as regular aspirin users because
aspirin accounted for about three-quarters of the adult,
nonsinus analgesics sold over the counter in the United
States at the time of the baseline e~amination.'~ To assess
the categorization of regular and occasional aspirin
use, we looked at the number of days before the baseline
examination that the subjects had taken aspirin:
82% of regular users and 56% of occasional users had
taken aspirin within the past 3 days.