|
Activity Number:
|
29
- Statistical Issues Specific to Therapeutic Areas, Power and Sample Size Calculations, and Trial Monitoring
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 8, 2021 : 1:30 PM to 3:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #319087
|
|
|
Title:
|
Do unadjusted pilot analyses lead to biased decision in proceeding to a larger trial?
|
|
Author(s):
|
Robert Montgomery*
|
|
Companies:
|
University of Kansas Medical Center
|
|
Keywords:
|
Pilot studies;
MCID;
Feasibility
|
|
Abstract:
|
Pilot studies provide invaluable information about the feasibility of a proposed clinical trial and help researchers better plan for recruitment, management of attrition and other design characteristics. Pilot studies are also frequently used to determine if continuing to a larger trial is warranted, typically by estimating if it is plausible that a minimum clinically important difference (MCID) can be reached. A typical approach is to use a confidence interval for the endpoint of interest, and if it contains (or exceeds the MCID) to determine that continuing to a larger trial is warranted. However, pilot studies are typically unadjusted for covariates while follow-up larger trials routinely adjust for demographic and other variables. Through simulation we show that this approach has no impact when potential covariates are balanced between groups, however, this is unlikely in pilot studies with small sample sizes. Failing to adjust for covariates can lead to biased confidence intervals and thus incorrect decisions about the plausibility of an MCID. Caution must be used in deciding whether pilot data provide evidence to continue into a larger trial based on an MCID.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2021 program
|
