Midterm Study Questions

Question 1

What is the main weakness of the RCT?

Answer 1

Not always generalizable to the lager public because they are so tightly controlled

Question 2

3 different types of randomization

Answer 2

1. Simple randomization
2. Stratification randomization
3. Block randomization

Question 3

What is simple randomization and what are limits/strengths?

Answer 3

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

Question 4

What is stratification randomization and limits/strengths?

Answer 4

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)

Question 5

What is block randomization and what are limits/strengths?

Answer 5

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

Question 6

What is it that RCTs can demonstrate that cohort, case control, and case series studies cannot do?

Answer 6

Establishes causation (not just correlation)

Question 7

What does each letter of the ABCDFIX tool stand for?

Answer 7

allocation, blinding, comparable groups, dropouts, follow-up, intention to treat, X-factors

Question 8

What do the numbers mean in 2-6?

Answer 8

• 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

Question 9

What are the most common acceptable ways to ensure that subject allocation is concealed?

Answer 9

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.

Question 10

Can allocation be concealed even if the patients, doctors and outcome measures are not blinded?

Answer 10

Yes. Blinding and allocation are different.

Question 11

What are the 4 Cs?

Answer 11

C = group comparisons

1. Comparable at baseline (start)?
2. Comparable co-interventions?
3. Comparable attention?
4. Comparable compliance?

Question 12

Who are the 3 most important participants in an RCT to blind?

Answer 12

Doctor, Patient, Outcome Assessor

Question 13

What is response bias?

Answer 13

Patients report symptoms in a way to please the provider

Question 14

Who else can be blinded?

Answer 14

statistician, adjuticator, research assistants

Question 15

What is an adjudicator?

Answer 15

An adjudicator is someone who presides, judges, and arbitrates during a formal dispute or competition

Question 16

What are 5 things that are good to check when looking at drop outs?

Answer 16

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

Question 17

What is the 5-20 rule? How is it used?

Answer 17

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.

Question 18

What is a reasonably good length of time for follow up in a musculoskeletal study?

Answer 18

6 months to a year

Question 19

What are four different types of “control” groups used in RCTs?

Answer 19

1. Receive experimental therapy
2. Receive the “usual” therapy
3. Receive a placebo therapy
4. Receive nothing

Question 20

What are methods to deal with or prevent uneven distribution of confounders after randomization?

Answer 20

• Per-protocol analysis
• IIT
• Worst-case analysis
• Or both IIT and per-protocol analysis

Question 21

What are the advantages of ITT analysis?

Answer 21

Corrects drop outs:

• Preserves randomization
• Helpsmaintainprognosticbalance
• Preservessamplesizebalance
• Helps prevent an overestimation of how good a treatment really is
• Betterreflectsrealpractice

Question 22

What are the challenges of creating a sham group for interventions like manipulation and acupuncture?

Answer 22

Hard to implement the placebo effect

Question 23

What method of ITT is thought to be the weakest?

Answer 23

Per-protocol results because it does not account for potential effects of dropouts or crossovers

Question 24

Where are two places (outside the abstract) where one can usually find out if an intention to treat
analysis was done?

Answer 24

Methodology or result section of the paper

Question 25

Order the proper sequence of the steps of an RCT:

Answer 25

• 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

Question 26

What are some of the x factors that can challenge internal validity?

Answer 26

• 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?