Research Design, Biostatistics, and Literature Evaluation
Pharmacokinetic and Pharmacodynamic trials
Studies designed to evaluate effects of unique disease states on clinically relevant pharmacokinetic and
pharmacodynamic parameters
routine clinical care laboratory samples (Clin Pharmacokinet 2015;7:783-95)
Large study size not needed to define properties in unique patient populations
Population based pharmacokinetic and pharamacodynamic studies can help define novel clinical dosing
strategies in unique patient populations
Noninferiority trials seek to answer whether a competing treatment is no worse than the established
Example: Quetiapine is inferior to (worse than) haloperidol for hyperactive ICU delirium versus
quetiapine is no worse than haloperidol by X margin.
burden (e.g., costs, inconvenience, labor)
Analytic approach to a noninferiority trial
What is an acceptable threshold for βnoninferiorityβ?
The FDA provides guidance for non-inferiority thresholds:
| (a) | Establish the smallest plausible benefit of the existing standard therapy. |
|---|---|
| (b) | Insist on some preservation of the treatment effect of the standard therapy. |
| (c) | The FDA recommends that 50% of the standard treatment effect be preserved when |
evaluating mortality for thrombolytic trials.
ii.
Maximum allowable negative outcome
Threats to meaningful comparisons
Effect of standard treatment was not preserved. (e.g., heparin infusion does not achieve
therapeutic anticoagulation)
ii.
Intention-to-treat analysis: Suboptimal administration of the standard treatment
Incidence versus prevalence
Incidence: Measures the occurrence of a disease (or event) over a time interval
Prevalence: Measures the occurrence of a disease (or event) at a fixed point in time
Internal
Does the study design adequately test the hypothesis?
ii.
Are the study methods sound?
External
Is the study population representative of the clinical setting?
ii.
Are the study findings generalizable outside the study setting?
iii.
Can the study be replicated in clinical practice?
Bias: Systematic error leading to an estimate of association in the study population that varies from the
Selection bias: Systematic selection of subjects that leads to an imbalance, or an advantage, in favor
of one cohort over the other