Midterm Study Questions Flashcards
What is the main weakness of the RCT?
Not always generalizable to the lager public because they are so tightly controlled
3 different types of randomization
- Simple randomization
- Stratification randomization
- Block randomization
What is simple randomization and what are limits/strengths?
Random number tables, tossing a coin, drawing different colored balls, etc.
+ It is simple, and when done with integrity it’s difficult to go wrong
- can suffer from ‘chance bias’
- groups may end up with unequal numbers
What is stratification randomization and limits/strengths?
In simple randomization you can end up with unbalanced groups in co-variate.
Stratification identifies covariates in the population and ensures they are evenly distributed between treatment groups
+ identical numbers in each group (NOT most important)
What is block randomization and what are limits/strengths?
Can balance the numbers in the group at anytime during the trial
+ ensures identical numbers in each group
- can lead to prediction of group allocation if block size is guessed. Does not guarantee all covariates are equally distributed
What is it that RCTs can demonstrate that cohort, case control, and case series studies cannot do?
Establishes causation (not just correlation)
What does each letter of the ABCDFIX tool stand for?
allocation, blinding, comparable groups, dropouts, follow-up, intention to treat, X-factors
What do the numbers mean in 2-6?
- 2 acceptable methods to allocate
- 3 need to be blinded: patients, providers, outcome measureres
- 4 comparison groups
- 5 dropouts? 5% is ok, more than 20% is hard to make outcomes valid
- 6 follow-up length 6 months to a year
What are the most common acceptable ways to ensure that subject allocation is concealed?
1) computerized allocation (not just computerized randomization) from an off-site location or
2) on site allocation via sealed opaque envelops preferable by an independent agent.
Can allocation be concealed even if the patients, doctors and outcome measures are not blinded?
Yes. Blinding and allocation are different.
What are the 4 Cs?
C = group comparisons
- Comparable at baseline (start)?
- Comparable co-interventions?
- Comparable attention?
- Comparable compliance?
Who are the 3 most important participants in an RCT to blind?
Doctor, Patient, Outcome Assessor
What is response bias?
Patients report symptoms in a way to please the provider
Who else can be blinded?
statistician, adjuticator, research assistants
What is an adjudicator?
An adjudicator is someone who presides, judges, and arbitrates during a formal dispute or competition
What are 5 things that are good to check when looking at drop outs?
study should report:
- Total number of drop outs
- Number of drop outs per group
- When they dropped out
- The reason they dropped out
- Whether or not patients dropping out were similar to those who stayed
What is the 5-20 rule? How is it used?
In most cases <5% drop outs do not threaten the results too much (but an ITT analysis should still have been done), 10-20% is a bigger threat and an ITT analysis or similar approach is expected, and >20% is such a big loss, that even an ITT analysis may not be able to offset the challenge to the validity of the outcomes.
What is a reasonably good length of time for follow up in a musculoskeletal study?
6 months to a year
What are four different types of “control” groups used in RCTs?
- Receive experimental therapy
- Receive the “usual” therapy
- Receive a placebo therapy
- Receive nothing
What are methods to deal with or prevent uneven distribution of confounders after randomization?
- Per-protocol analysis
- IIT
- Worst-case analysis
- Or both IIT and per-protocol analysis
What are the advantages of ITT analysis?
Corrects drop outs:
- Preserves randomization
- Helpsmaintainprognosticbalance
- Preservessamplesizebalance
- Helps prevent an overestimation of how good a treatment really is
- Betterreflectsrealpractice
What are the challenges of creating a sham group for interventions like manipulation and acupuncture?
Hard to implement the placebo effect
What method of ITT is thought to be the weakest?
Per-protocol results because it does not account for potential effects of dropouts or crossovers
Where are two places (outside the abstract) where one can usually find out if an intention to treat
analysis was done?
Methodology or result section of the paper
Order the proper sequence of the steps of an RCT:
- sampling
- application of inclusion and exclusion criteria
- randomization
- allocation
- assessing baseline characteristics
- treating and monitoring during the intervention phase
- post intervention follow up
- statistical analysis
- adjusting results as necessary
What are some of the x factors that can challenge internal validity?
- Conflict of interest
- Was therapy given at an acceptable “therapeutic” dose
- We’re the outcome measures used to judge success validated and are they accurate?
- Was the sampling process flawed?
