Biomedical Engineering Reference
In-Depth Information
TABLE 12.1: Elicited Utilities of the Joint Outcomes (The category \>
120"-corresponds to the clot not being dissolved by the end of the 120-minute
infusion. SICH = symptomatic intracranial hemmorhage.)
Timetodissolvethebloodclot,inminutes
0 1{1516{3031{4546{6061{7576{9091{105106{120>120
NoSICH100 95 90 85 80 75 70 60 50 30
SICH 7 6.5 6 5 4.5 4 2 1 0 0
The following rules protect patients in case all (c;q) pairs are too toxic or ine-
cacious. Let
T
be the maximum allowed
T
(1;c;q;) and
E
the minimum allowed
probability of dissolving the clot within 120 minutes, F
E
(1;c;q;): A pair (c;q) is
unacceptable if it is likely to be too toxic, Prf
T
(1;c;q;) >
T
jD
n
g > p
T
; or it is
likely to be inecacious, PrfF
E
(1;c;q;) <
E
jD
n
g > p
E
: The cut-os p
T
and p
E
are generally values between 0.80 and 0.99. These criteria are similar to established
Bayesian phase I/II dose acceptability rules (Thall and Cook, 2004; Thall et al.,
2008).
Once the design parameters and model are established, given a set of (c;q) pairs,
maximum sample size N; and cohort size, the trial is conducted as follows. The
rst cohort is treated at a starting (c;q) combination chosen by the physicians. For
each cohort after the rst, if no (c;q) is acceptable, then the trial is stopped early.
Otherwise, the next cohort is treated at (c;q)
opt
(D
n
); subject to the safety rule that
no untried concentration may be skipped when escalating. At the end of the trial,
the nal (c;q)
opt
(data) is selected.
12.4
Application
Because Y
E
is interval-censored, as a covariate in the linear term of
T
it can take
on only the values f0g, or an interval endpoint 1/6, ..., 5/6, 1, or 1(Y
E
> 1): In the
Search WWH ::
Custom Search