BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.cam.ac.uk//v3//EN
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:DAYLIGHT
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
TZNAME:BST
DTSTART:19700329T010000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
TZNAME:GMT
DTSTART:19701025T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:MRC Biostatistics Unit Seminars
SUMMARY:Virtual Seminar: ‘The DURATIONS design: a practica
l trial design to optimise treatment duration’ - D
r Matteo Quartagno\, MRC Clinical Trials Unit\, UC
L
DTSTART;TZID=Europe/London:20201215T140000
DTEND;TZID=Europe/London:20201215T150000
UID:TALK153958AThttp://talks.cam.ac.uk
URL:http://talks.cam.ac.uk/talk/index/153958
DESCRIPTION:Introduction: Trials to identify the minimal effec
tive treatment duration are needed in different th
erapeutic areas\, including antibiotics\, TB and H
epatitis C. However\, standard non-inferiority des
igns have several limitations\, including arbitrar
iness of choice of shorter duration arms and large
sample sizes.\n\nMethods: We propose a multi-arm
trial design\, allocating patients to different tr
eatment durations. We recast the problem of testin
g for non-inferiority of a single shorter treatmen
t duration in terms of modelling the entire durati
on-response curve within a pre-specified range. We
use fractional polynomials to flexibly model the
duration-response curve. We compare different meth
ods to draw inference from the estimated curve\, i
.e. to choose the “optimal” duration. We explore b
oth different estimands and estimation methods\, i
n terms of newly developed operating characteristi
cs.\n\nResults: Through simulations we show that\,
in a variety of scenarios\, when using fractional
polynomials 5-7 equidistant duration arms are gen
erally enough to get a good estimate of the durati
on response curve. In order to draw inference from
this curve\, the simplest estimand to target is t
he shortest duration non-inferior to the control o
ne within a certain non-inferiority margin. The ma
rgin can itself be dependent on duration. In term
s of estimation method\, using bootstrap is prefer
able\, as it accounts for model selection variabil
ity\, and leads to type-1 error rates within the n
ominal level in all scenarios considered.\n\nDiscu
ssion: Our proposed practical trial design is an a
lternative to standard non-inferiority designs\, a
voiding many of their limitations\, and yet being
fairly robust to different possible duration-respo
nse curves.
LOCATION:Virtual Seminar
CONTACT:Alison Quenault
END:VEVENT
END:VCALENDAR