Quiz 2

Question 1

What is systematic review?

Answer 1

Analysis of lit to address a specific question using rigorous procedures – including defined inclusion and exclusion criteria and literature search procedure – to identify and summarize existing studies and synthesize an answer

Question 2

What is a problem with systematic review?

Answer 2

Heterogenity

Question 3

What is meta analysis?

Answer 3

1. A systematic review which uses statistical methods to combine data from multiple studies of the same topic

Question 4

How does meta analysis differ from systematic review?

Answer 4

M: A quantitative and more objective type of systematic review.

Goes beyond critique to conduct secondary statistical analysis on outcomes of studies

Results are combined to produce overall statistic

Question 5

What is a problem with meta anaylsis?

Answer 5

Heterogenity

Question 6

What is a clinical practice guideline?

Answer 6

Statements that include recommendations intended to optimize patient care

Assess the benefits and harms of alternative care options

Question 7

What is CPGs based on?

Answer 7

1. Systematic assessment of the best available research and practice experience.
2. Large committees weigh all available evidence and expert opinions to estimate the benefits and harms of diagnostic procedures and treatments

Question 8

What is CPGs used for?

Answer 8

Tell how evidence was evaluated

Question 9

What is a population?

Answer 9

the set of all subjects or measurements

Question 10

What is a sample?

Answer 10

The group from which you will collect data

Question 11

What is a parallel design?

Answer 11

Each subject is randomized to either the experimental or the control group and receive the intervention assigned for the duration of the study

Question 12

What is crossover design?

Answer 12

1. Each subject receives all interventions in a specified sequence.
2. There is a washout period between treatments.
3. Balanced
4. Requires fewer subjects because each serves as her/his own control.

Question 13

How are subjects selected in crossover design?

Answer 13

Randomization of sequence helps to account for temporal trends

Question 14

Identify the comparison between parallel and crossover designs?

Answer 14

P: Groups assigned different treatments, shorter duration, sample size-large, no carryover effect, acute cases
C: Each patient receives both treatments, longer, smaller, carryover effect, not in acute but chronic and stable cases

Question 15

What are the advantages of crossover?

Answer 15

1. Small sample size
2. Balances cofounders
3. Can examine participant by intervention interaction

Question 16

What are the disadvantages of crossover?

Answer 16

1. Potential carry over
2. Longer follow up
3. More loss follow up with longer trials and exposure to multiple interventions
4. Parallel takes less time but larger sample size

Question 17

What scenarios would cross overs be the most advantageous?

Answer 17

1. Chronic, stable diseases with no permanent cure
2. Interventions with quickly reversible effect
3. Drugs with short half-life
4. Endpoints with large intrasubject variation
5. Short interventions are necessary

Question 18

What is factorial design?

Answer 18

1. An experiment that manipulates >1 independent variable (factors) at one time
2. Uses the same study population test >1 intervention
3. Measures every combination of variables and levels

Question 19

How are each factors measured in factorial design?

Answer 19

A number of different settings called levels

Question 20

How are factorial designs subjects selected?

Answer 20

Randomly assigned to a treatment combination from a pool of all possible combinations of factor levels

Question 21

What is an example of factorial design?

Answer 21

RCT which contains 2 levels: “on” active treatment condition and “off” control treatment condition

Question 22

What are the advantages of factorial design?

Answer 22

1. Investigation of results of interactions between factors
2. Interactions aren’t detected by one factor but many
3. Allows effects of a factor to be estimated at a number of levels of other factors yielding conclusions that are valid over a range of experimental condition

Question 23

Distinguish the differences among factorial, parallel, and crossovers regarding rondomizizatinon and outcomes?

Answer 23

Question 24

What is non-equivalent control group?

Answer 24

1. Compares groups where subjects are not randomly assigned to groups
2. Researcher has no control over the intervention
3. Used when a natural and common phenomenon affects a population

Question 25

What are the advantages of non-equivalent control group?

Answer 25

1. Lower temporal bias than cross-sectional design 2. Can be done when randomization is impractical or unethical

Question 26

What are the disadvantages of non-equivalent control groups?

Answer 26

1. Potential maturation bias and history bias 2. Possible regression to the mean 3. Selection bias

Question 27

What is time series methods?

Answer 27

1. Add more observations to pre-post test so that trends over time can be observed 2. Decrease probability of maturation bias 3. Multiple observations are made, before and after intervention 4. Track trends

Question 28

What is time series method useful for

Answer 28

1. Examining trends in an outcome over time 2. The possibility of an immediate treatment effect 3. The sustainability of a treatment effect over time

Question 29

What is double dummy design?

Answer 29

1. Use for blinding in trials investigating active interventions that can not be made indistinguishable

Question 30

How is double dummy design laid out?

Answer 30

Uses 2 placebos, one for each intervention

Question 31

How are subjects selected in double dummy?

Answer 31

Randomized to receive one active treatment and one placebo not knowing which For comparison of 2 treatments, A and B each subject takes 2 sets of treatment, either active A and placebo B or placebo A and active B.

Question 32

What is adaptive design?

Answer 32

1. Uses accumulating data from trial to modify aspects of the trial 2. Trial arm may be dropped

Question 33

What can be changed in adaptive design?

Answer 33

1. Sample size can be recalculated 2. Dose 3. End point 4. Randomization ratio

Question 34

What is the reason adaptive design is stopped?

Answer 34

Efficacy, futility, ssafety

Question 35

What is stratified random sampling?

Answer 35

1. Population to be sampled is 1st divided into groups based on characteristics believed to be importnat 2. Probabilities may be different between the groups 3. Assures that the different characteristics on which strata are based are sufficiently represented in the sample 4. Resulting data must be analyzed using statistics that account for stratification

Question 36

What is cluster random sampling?

Answer 36

1. Researchers divide the population to sampled into smaller groups or clusters 2. Researchers then randomly select clusters to be in the sample. 3. Unit randomized is not individual subjects but clusters of subjects

Question 37

What is cluster random sampling useful for?

Answer 37

1. When the target of an intervention is a system rather than an individual patient 2. Better able to test if a new standard of care, recommendation, or systems wide policy is affecting outcomes

Question 38

What is block random sampling?

Answer 38

1. Individuals are arbitrarily divided into blocks (subsets) with even numbers of subjects and then randomly assigned from those blocks to the different treatment groups. 2. Blocks have no significance other than an assignment group. 3. Guarantees equal numbers in the two treatment groups.

Question 39

What is non-inferiority design?

Answer 39

1. Not the same as equivalence 2. Design produces a statistical measure that shows that the tested intervention is clinically not significantly inferior to a standard treatment or active control (a real drug not placebo). 3. Specifies a statistical tolerance (boundary) that is deemed not clinically significant

Question 40

Why do we use non-inferiority?

Answer 40

1. New product similar efficacy but expect to have other advantages such as safety or cost. 2. Establishing superiority difficult 3. Testing vs placebo unethical