Biomedical Engineering Reference
In-Depth Information
FIGURE 12.2: Illustration of possible shapes of (s;c;q) as a function of s
for xed (c;q) = (0:5; 0:1):
a and b by a^b and = ( 0 ; ; 4 ); the model for this probability is given by
logf1 T (Y E ;c;q;)g = 0 + 2 c 1 q+ 3 c 1 (1q)(Y E ^1)+ 4 1(Y E > 1): (12.6)
The bolus effect is 2 c 1 q, the ci effect is 3 c 1 (1 q)(Y E ^ 1); the eect of failing
to dissolve the clot is 4 ; and 1e 0 is the probability of SICH if no tPA is given.
Because 3 > 0; the faster the clot is dissolved, the smaller the probability of SICH.
Figure 12.3 illustrates possible forms of T (Y E ;c;q;):
The model parameter vector is = (;); with dim() = 11.
If no bolus were given (q = 0), the effective delivered dose at s would be
d(s;c 1 ; 0) = c 1 s, so 2 would be dropped, the hazard function (12:4) would become
(s;c; 0;) = 3 + 4 5 (c 1 s) 5 1
1 + 4 (c 1 s) 5
for s > 0;
(12.7)
and the cumulative hazard function (12:5) would become
(s;c; 0;) = 3 s + c 1 logf1 + 4 (c 1 s) 5 g for s > 0:
(12.8)
The model for T would drop 2 and the linear component (12:6) would be reduced
to 0 + 3 c 1 (Y E ^ 1) + 4 1(Y E > 1); so dim() = 9.
The joint distribution of Y = (Y E ;Y T ) is
for yT T = 0; 1 and y E
f E;T (y E ;y T jc;q;) = f E (y E jc;q;) Pr(Y T = y T jy E ;c;q;)
0:
 
Search WWH ::




Custom Search