Interventional Study Designs Flashcards
What is the key difference b/w interventional & observational studies?
Investigator selects interventions and allocates study subjects to forced-intervention groups -can demonstrate Causation
What are the two buzz words for interventional study designs?
Randomization and phase
Which of the interventional studies has the highest evidence? And lowest evidence?
-Phase 4 (highest) -Phase 0 (lowest)
What is a phase 0 study?
(Exploratory, Investigational New Drug)
- First in human use/study ***
- Purpose: asses drug target actions, pharmokinetics a (is drug targeting what its supposed to and at what %)
- It’s to ask a question (not about drug efficacy)
ex) New Ca therapy, not trying to see if its effective, just if it can get to target - Healthy volunteers (or dz patients)
-Very small N (<20)
-Very short duration
– usually a single dose to just a few days (max)
Phase 1
(Investigational new drug)
-Asses safety/tolerance and pharmokinetics
-Healthy or Dz pats
-Small N (20-80..bigger than phase 0)
-Short duration (several days-few weeks.. longer than phase 0)
- can ppl tolerate drug?
- is it safe?
- Side effects?
- how long does it stay in system?
Phase 2
- Asses effectiveness (expands on phase 1)
- Dz volunteers (ppl with condition of interest)
-Larger N (few to several)
-Short to medium duration (several weeks - to several/few months) **
- restrictive inclusion and exclusion criteria
- ex) 128 volunteers with parkinson’s were randomly assigned into 3 groups
A one year long study would NOT be which phase studies?
0 or 1
Phase 3
***assess effectiveness (continue to asses safety and tolerability)
- diseased volunteers
- larger N (several hundred to few thousand)
- longer duration (few months to a year +) ***
- longer and more ppl than phase 2***
- can be a repeated study, or progression
Which phases are healthy patients never used in??
Phase 2, 3, (and 4)
When are 2 placebos used?
When the route of administration is different (iv vs inhalation)
What is the final phase before they must seek approval from FDA?
Phase 3
Phase 4
(post FDA approval)
- assess long-term safety, effectiveness, and optimal use (risks/benefits)
- Dz volunteers (expand use criteria - use more blended popln)
- Population N = few hundred to tens of thousands)
- Wide range of durations (many months to several years, ongoing)
- interventional or observational designs
- track data base of observations
What are the advantages and disadvantages of Interventional trials?
Advantages = demonstrate causation and Key designs used by FDA for approval, environment highly controlled.
Diasdav = expensive, complex, timely, ethical considerations, generalizability–external validity
Traditional - Exploratory Interventional Studies
- Trying to asses drug effectiveness, optimal use, does it get to the target
- No flexibility for clinicians
- so Pts either live with the terms or drop out -worry about generalizability
Pragmatic - Explanatory Interventional Studies
- Still trying to explore most effective tx BUT their goal is how to manage a condition
- gives clinicians more flexibility
- maximized applicability of the trials and results to usual care settings
- more practical management
What is the primary difference b/w simple and factorial interventional studies
Simple - only one randomization event
– tests single hypothesis
Factorial - multiple randomization events (2 or more)
– 2x2 or 2x2x2
–multiple hypotheses
What is the primary difference b/w parallel and crossover studies?
Parallel = no switching b/w intervention groups
Crossover (self-control) = switching/crossover b/w groups
- advantages of crossover:
– groups can act as their own contorl
– allows for smaller N bc each pt counts as two ppl
– allows b/w and w/in group comparison possible
When does wash out occur and what does that mean?
- during crossover studies, can “wash out” previous medicaiton givcen before crossing over
- occurs after the study has already started, so in the middle of the study
- when doing crossover study, it prevents the “hangover effect” and gives times for drug/tx to get out of system
***back to baseline***
What does Lead-in mean?And when does it occur?
Lead-in occurs before study starts -asses study protocol compliance (dry run/rhersal) -can act as a washing out phase -can determine amount of placebo effect - can decrease exclusion criteria
What are the most CLINICALLY relevant endpoints for interventional studies?
1) Patient-oriented (Patient-Oriented Endpoints = POEM’s)
***happen to people***
- death, stroke/MI, hospitalization, preventing need for dialysis
- pt’s would rather hear they have a decreased risk for death or the above things, than hearing “you lowered my BP # or cholesterol”
2) DOEMs - BP, Cholesterol, lab values
- can be used to help predict POEMs but are measures to evaluate whats happening
What are some examples of Dz-oriented Endpoints (DOE’Ms)?
- BP
- cholesterol
- change in SCr
- surrogate markers
- used in place of evaluating POEMs
What is the goal/purpose of randomization?
To make sure groups are as equal as possible, based on confounders
- decreases biases
- allows group allocation to be taken out of researchers hands
What are 3 forms of randomization?
Simple
Blocked
Stratified
Describe simple randomization:
equal probability for allocation
Describe blocked randomization:
Ensures balance w/in each intervention group
- more hands on
- ex = every N^th person stop and look if groups are even, if unbalanced then put the next ppl where needed
Describe stratified randomization:
ensures balance with known confounding variables
- stratified individualistic on particular strata
- gender, age, dz severity etc
- can pre-select levels to be balanced within each confounder
What are the 3 types of Masking?
Single-blind
Double-blind
Open-Label (unmasked/unblinded)
What is the main difference b/w single and double blind studies?
Single: subject not informed which tx group they’re in, but researcher knows
Double: neither subject or researcher knows -prevents interviewer, response bias
What is open-label masking?
Subjects and researcher both know what intervention is being used
- can be a continuation of a double blind study in an open fashion to see if benefit continues
What kind of studies can be used to assess the adequacy of blinding?
Post-hoc survey’s
What is another name for “Dummy” therapy?
Placebo therapy
- inert tx’s made to look ID in all aspects to the active tx
– dosage forms, dosing frequency, monitoring, therapy requirements
What are Double-Dummy tx’s?
Using more than one placebo
– primarily if there are different modes of tx (ex = inhalation vs injection)
What is the placebo effect?
Improvement in condition by power of suggestion of being ‘treated’
- can be as large as 30-50%
- benefit just from being in study and getting tx “tender, love and care”
Hawthorne-effect
Subjects change their behavior solely due to awareness of being studied/observed
What are post-hoc sub-group analyses? What are the positives and negatives?
- another analysis after study has been completed and data has already been collected (data-dredging or fishing)
***They are not accepted as appropriate by most when NOT prospectively planned***
- Negatives: reduces power and increases risk of Type 2 error
***Are accepted when prospectively planned and delineated in the methods that it will occur***
-Positives: can allow for hypothesis generation and development of future studies
What are the two ways to handle data collected from subjects who dropped out or were lost to f/u?
- Include them anyway
- Ignore them
If USE data even if subject dropped out what type of principle is this?
What are the benefits?
Intention to Treat
- Most CONSERVATIVE
- maintains group allocation and preserves randomization
- preserves baseline characteristic and group balance at baseline (which contrls for confounders)
- maintains statistical power
What are the 2 most common forms of group allocation procedures?
1) non-random = subjects don’t have equal probability of being selected or assigned to each group
2) random = most used
- subjects DO have equal chance of being assigned to each group
When managing Drop outs/ LTFU, you just IGNORE them (including only compliant or completing subjects), what types of studies are these? (theres 2)
1) Per-protocol/Efficacy analysis = compliance is pre-determined
- introduces bias (bc commonly over estimates effects) and decreases generalizability
2) Treating them As treated = ignore original group assignment and count them wherever they fit best
- use data as they were tx not as their intended randomized group
- not common and can introduce contamination bias
What are some was to 1) assess adherence/compliance and 2) imoprove adherence?
1) drug levels, pill counts at each visit, bottle counter-tops
*** stay engaged with patients***
2) frequent f/u appointments, communication with subjects, tx alarms/notifications, medication dosage containers
