The plot of the age distribution, grouped by 10-year
intervals, into postoperative bleeding and nonpostoperative
bleeding groups shows that the proportion
of elderly patients is larger in the postoperative
bleeding groups (Figure 2). This observation was
tested in a new regression permitting patients up to
and including 35 years of age to have a different risk
of postoperative bleeding from patients above 35:
P-OB ~ 1+(age > 35).