23 terms

Biostats RCT


Terms in this set (...)

randomized clinical trial
2 most important aspects of RCTs
treatment artificially manipulated
subjects randomly assigned
What do you start with in a RCT?
a defined population that fits certain criteria
What do you do with your selected population in RCTs?
randomly divide them into groups such as new vs old drug or placebo vs drug
Why must you use a control group?
Control group necessary to determine whether an
observed outcome is a true effect of the treatment
or would have occurred in the absence of the
therapy under evaluation.
What can happen if you do not have a concurrent control group?
changes in patient characteristics
• changes in the natural history of disease
• changes in diagnostic criteria or therapy
• improvements in health care
What does randomization allow?
a higher likelihood that the groups are comparable in both known and unknown characteristics
- gives validity
What does randomization prevent?
selection bias
What biases does patient blinding prevent?
post randomization selection
post randomization selection bias
placebo group patients may eat healthier or exercise more
Attrition bias
those that know they are on a placebo are more likely to drop out
information bias
groups receiving tx are more likely to report AEs than placebo groups
Why blind the HCP?
prevents paying more attention to tx group
Why blind data analysts?
prevents the use of particular statistical methods in order to favor one group or the other
When would you use stratification?
when you know of a particular trait that will affect the outcome - gender, age, etc
What are the two types of crossover studies?
Planned crossover
randomized -> new treatment vs current treatment -> pts are then switched to the opposite treatment
washout period
the time for the effects of the first treatment option to subside before crossing over
unplanned crossover
randomized - tx option 1 vs tx option 2 -> pts them switch due to various factors
refuse to comply or just drop out
pts accidentally include themselves in a tx option group that they were not assigned
triple blinded
health care providers
Gold standard
triple blinded placebo controlled randomized clinical trial