Randomized Controlled Trials Flashcards
What is the goal of a Randomized Controlled Trial (RCT)?
Test whether an intervention works by comparing it to a control condition
What quality of RCTs make them the gold standard for drug studies?
Randomization as it reduces the effects of confounding
How are patients separated into groups in parallel randomization?
The study sample is divided into two groups and different intervention/control are applied
How are patients divided into groups in a cluster randomized RCT design?
Instead of individuals randomly separated between control/intervention groups, entire clusters of individuals are randomized
Ex. All patients at Pharmacy A receive intervention, while all patients at Pharmacy B receive placebo
How are patients separated into groups when using cross-over randomization?
Study sample is randomly divided into groups and outcomes are evaluated. A period of time without interventions passes, and the groups switch intervention/control and a second set of outcomes are evaluated
Now researchers have data on intervention vs. Control in the same patient, the patient serves as their own control which reduces confounding in the study
Why should a study be well-designed before it begins?
To limit bias in the study and outcomes. Pre-defined process reduces the ability to change the results to fulfill a preferred outcome. Also hold the researcher accountable for the integrity of the study
What is power in regards to sample size in a research study?
This value is the minimum sample size that allows the study to reliably detect differences between groups
What are the consequences of a study that is overpowered in terms of sample size?
There could be ethical concerns with unnecessary exposure to intervention
Cost considerations with a larger sample size
What is the benefit of increasing sample size in a research study?
It can reduce chance in the findings of a given study
What is the intervention compared against in RCTs?
- Placebo/usual care
- A different dose
- A different drug/procedure
- Standard treatment (most common control in drug studies)
What are hard endpoints in a study?
These are the final events that the study is hoping to prevent
Ex. Death, stroke, MI
What are the two types of research outcomes?
- Objective (measurable): blood pressure, lipid levels
- Subjective (patient interpretation): back pain
What are surrogate endpoints in research studies?
These are outcomes that our study hopes to see change. Surrogate endpoints are correlated with hard endpoints
Ex. Blood pressure (surrogate endpoint) —> Stroke (hard endpoint)
What are the limitations of surrogate endpoints?
The extent to which surrogate endpoints and hard endpoints are correlated is not always known
What is the primary endpoint?
This is the main result measured at the end of the study to see the effect of the intervention
What are secondary endpoints?
These are additional results of interest, but not the main focus of the study
These results can serve as the basis for future studies where they are the primary endpoints
What are composite endpoints?
This is a primary endpoint that contains several events
Ex. MACE (Major Adverse Cardiovascular Events) looks at MI, stroke and death together instead of death (hard endpoint) alone
What threat to internal validity does randomization reduce?
Confounding is reduced
What are the qualities of well randomized groups?
- Similar in things we can measure (ex. Age, sex, BP)
- Similar in things that are unknown (ex. Genetics)
What are the three methods of randomization?
- Complete
- Block
- Stratified
What are some qualities of complete randomization?
- There are no limitations to randomization
- won’t necessarily get equal numbers in each group (ex. Sample A: 52, Sample B: 48)
What are some characteristics of block randomization?
Used to force balance in the number of subjects in each group
Ex. Block of 6 (6 subjects randomized (3 into invention, and 3 into control), followed by block of 10 (5 into intervention, and 5 into control), continue once all participants have been randomized
What are some characteristics of stratified randomization?
Used to achieve similarities in certain baseline characteristics between all groups
Ex. Everyone under 65 is randomized, then everyone over 65 is randomized. This ensures both groups have the same age composition
What is IIT in RCT study outcome analysis?
IIT (intention to treat)
When a study follows IIT, they analyze the data for all patients (even those that dropped out, died, etc.) and according to the group they were originally randomized to
IIT preserves the value of randomization and reasons for why patients left study
What is blinding in a study?
This is when parties to a study are unaware of what study group a subject has been randomized into
What are the consequences of known group assignment?
- Subjects may report outcomes or adverse effects differently or behave differently
- Physicians/investigators may report outcomes or events differently or treat subjects differently
- Researchers may collect data, classify outcomes or events and interpret subject response differently
If any of the above occur, it affects the internal validity of the study (bias)
What is an unblinded (open-label) study?
Participants, researchers, and study evaluators all know what group subjects have been placed into
What is a single blinded study?
Study evaluators do not know group assignment, while researchers and participants know what group they fall under
What is a double blind study?
This is the minimum standard
Either researcher of study evaluator do not know group assignement, but subjects do not know their group assignment
What is a triple blinded study?
Neither study evaluators, researchers, and participants do not know group assignment
Used by drug studies hoping for drug approval (very rigorous blinding method)
What is placebo effect?
This is the perceived or actual effect from an ineffectual or inactive treatment
