Study Designs Flashcards
What are the types of study designs? And how are they different?
Interventional- places subjects in forced group allocation and randomizes groups and can prove causation.
Observational- observing the patient and giving no interventions. Applies no exposure or group allocation
What are the phases of an interventional study? And describe them?
Pre-clinical- prior to human investigation-bench and animal research.
Phase 1- small N, healthy volunteers, short duration, checking for safety, and pharmacokinetics.
Phase 2- larger N, medium duration, commonly use patients with disease of interest, safety a concern.
Phase 3- right before FDA approval- efficacy of drug is primary focus- superiority, equivalency, and non-inferiority.
Phase 4- post-marketing- checking for long-term effects in a large population of diseased patients. Expanded patient diversity.
What are the advantages of interventional study?
They prove causation.
What are the disadvantages of interventional studies.
Money
Cost/complexity
Ethical reasons
Generalizability-external validity
What are the interventional study types?
Explanatory- not flexible-no dose changes an less realistic to actual treatment.
Pragmatic- more realistic to patient treatment- doesn’t use placebo- allows co morbidities
What are the two types of study designs? And what are their sub groups?
Simple- randomized once and only tests one hypothesis.
Factorial- randomized more than once and tests 2 or more hypotheses.
The subgroups are parallel and cross over
What phase must happen before groups switch and a cross-over takes place?
Wash-out
What is the lead-in
The practice phase that can include a wash out. Can account for placebo effect and Hawthorne effect.
What’s an advantage of crossover?
Allows for a lower sample population
What are the disadvantages of crossover?
Only good for long term, not curable diseases.
Takes longer
Treatment by period
Smaller N but only if variance within groups is smaller than between groups
Complexity
What are the types of outcomes/end points?
Primary, secondary, tertiary etc and composite.
What are patient oriented outcomes vs surrogate markers?
Patient ortiented outcomes are life changing such as death stroke MI.
Surrogate markers are usually risk factors like blood pressure and cholesterol.
What the difference between non random and random sample selection?
Non-random doesn’t give everyone an equal choice of being allocated into any group.
Random- everyone has an equal chance at being allocated into any group.
What is the purpose of randomization?
To make the groups equal
Which table gives group characteristics and what value determine if the groups are equal?
Table 1 and the p value
What are the three types of randomization?
Simple- equal chance for being in any group.
Blocked- forced equality
Stratified- same level of smokers in each group.
What are the different ways of masking?
Single-blind: only subjects are masked
Double blind- both researcher and subjects are masked.
Open-label: neither subject nor researcher is blinded. Only good for objective readings like BP readings.
What are dummy therapy and double dummy studies?
dummy therapy- placebo used
Double dummy- more than one placebo used.
What are the placebo affect and the Hawthorne affect?
Placebo- feeling better because of attention (psychological)
Hawthorne- reporting positive results to please investigators.
What is post-hoc sub group analysis?
Fishing for data, must be prospectively planned.
What are ways to manage drop outs?
Keep them via
intent to treat: either use their last assessment or their baseline
* conservative-saves randomization-preserves baseline-maintains statistical power.
Ignore them via:
Per protocol - they must complete a certain percentage of the study to be considered.
Compliance must be predetermined.
* overestimates and reduces generalizability.
What is as treated?
Ignores group assignments, allows patients to switch groups.
What are ways to test for compliance? And improve compliance?
Drug levels
Pill counters
Bottle counter tops
~
Frequent communication
Treatment alarms
Dosage containers and medication blister packs
What are case control studies?
OBSERVATIONAL studies that know if an individual is diseased or not but tries to determine exposure.
Are case control studies retrospective or prospective?
Retrospective
What are the strengths of case control studies?
- Good for assessing multiple exposures of one outcome.
- Good when diseases are rare
- calculating OR
- less expensive
- useful when ethical issues are a roadblock
- useful for dynamic populations
- useful when disease has a long incubation period.
When selecting cases what is so important?
That they are diagnosed correctly! You want to avoid misclassification.
Which is more difficult and why, selection of cases or controls?
Selection of controls because they must match up to the cases as accurately as possible.
It’s also the major determination on if the study is valid (internal validly)
Controls are randomly selected. How is this different from randomization?
Randomly selected means a number of people are randomly picked from a population.
Randomization is systematically ensuring that 2 groups are equal.
Where do the controls come from?
Institutions/organizations
Friends/spouses
Outbreak sources
Is is possible to be a case and a control in the same study?
Yes,
In an outbreak with multiple exposures
In change of risk (seasonal)
What is a Case-crossover design?
A case control study that addresses the issue of temporality.
Subjects are their own controls during times when they’re not at risk.
What is a Nested Case Control study?
A case control study that is conducted after or from a cohort study.
Where do the samples Coke from I’m a nestled case control study?
Survivor sample- non-diseased people who never got sick.
Base sampling- non-diseased at the start of the study period.
Risk set-
What are selection and recall bias?
Selection- bias when selecting subjects
Recall- subjects not recalling information accurately
What are the 2 types of matching?
Individual and group
