Principles of Epi and Clinical research design

Question 1

Phase I

Answer 1

• Focus on PK and PD
• Provides estimation of initial drug safety and tolerability
• Subpopulations (renal failure, hepatic failure, very young) are important aspects

Question 2

Phase II

Answer 2

• Earliest attempt to establish efficacy in intended patient population
• Very narrow eligibility
• Leads to a homogenous population that is monitored carefully
• Determines dose and regimen for later studies
• Determines potential study endpoints and subsets of disease population

Question 3

Phase III

Answer 3

• objective is confirm therapeutic benefit
• Provides firm evidence of agents efficacy and safety
• Blinded and randomized
• Estimate the size of the treatment effect
• Predefined hypothesis which is tested

Question 4

Phase IV

Answer 4

• Post marketing surveillance aimed to accumulate longer term safety from large number of patients followed for extended periods
• Begin after drug reaches the market
• Shows how drug performs in everyday setting

Question 5

Which study is good for rare diseases/outcomes?

Answer 5

• Case control

Question 6

Which study is good for rar exposures?

Answer 6

• Cohort

Question 7

Which types of phases are open label?

Answer 7

• Phase I

- Phase IV

Question 8

Longitudinal study

Answer 8

• Data gathered from the same subjects repeatedly over a long period of time
• Observational
• Cohort studies are a type of longitudinal study

Question 9

Parallel group design

Answer 9

• If the patient remains int he sane group to which they are initially assigned

Question 10

Post hoc power analysis

Answer 10

• If the study finds a non signficant different between the two groups, there is a possibility to a type II error (saying a difference doesn’t exist when it does)
• Can calculate the minimum detectable difference, if it is within the CI then type II error might have occured, if it is outside of the CI, then the study was adequately powered

Question 11

Bias

Answer 11

• Lack of internal validity or incorrect assessment between the exposure and outcome in a target population

Question 12

Selection bias

Answer 12

• Study population does not represent target pop
• Lack of accurate sampling
• Lack of randomization

Question 13

Effect modification

Answer 13

• Analyses that report on small sub groups can also lead to misleading results

Question 14

Information bias

Answer 14

• Occurs during data collection
• Withdrawal bias: when losses are uneven between groups
• Detection bias: differing diagnostic procedures in the target pop (latent tb screen in patients on biologics)
• Misclassification bias is when exposure, outcome or other important variables are not determined accurately (recall or interview bias)

Question 15

Regression to the mean

Answer 15

A variable that shows an extreme value on any assessment will tend to be close to the center of distribution on a later measurement

Question 16

Confounder

Answer 16

Estimate of the effect of an exposure results from factors associated with both exposure and outcome
- Can be dealt with my matching, restrict enrollment, randomizing, stratifying, multivariate analysis , propensity scores

Question 17

Incidence

Answer 17

• Rate at which newly diagnosed cases develop over time in a the population
• Expressed in units of cases/person-time

Question 18

Prevalence

Answer 18

Total number of existing cases in a defined population either at a point in time or during some time period
- Can be expressed as a proportion, percentage or by actual numbers

Question 19

Cost effective analysis

Answer 19

• Special type of decision analysis that addresses questions of the cost of health interventions compared with health outcome
• Cost of a new medication worth paying by socieity?

Question 20

Sensitivity analysis

Answer 20

• Repeat of the primary analysis or meta analysis substituing alternative decisions or range of valves for decisions that were arbitatry or unclear

Question 21

Face validity

Answer 21

• Extent to which a test measures a construct as viewed by lay persons
• Is it a reasonable test?

Question 22

Construct validity

Answer 22

• Does it measure what it is supposed to measure?

Question 23

Criterion validity

Answer 23

• Is there a relationship between the test score and an outcome (sat and college success)

Question 24

Predictive validity

Answer 24

• Refers to the power or usefulness of a test to predict future outcomes

Question 25

Content validity

Answer 25

• Address the match between the test question and the subject its trying to assess

Question 26

Kappa test

Answer 26

• used when outcome is dichotomous

- > 80 is excellent, 60-80 is good, 40-60 is fair, 20-40 is minimnal, <20 bad

Question 27

What coefficient is used when outcomes are continuous/numeric

Answer 27

• Intraclass correlation

Question 28

Cronbach’s alpha

Answer 28

• Measurement of internal consistency

- Correlation of two tests that measure the same construct

Question 29

Power

Answer 29

The ability of the statistical test to identify a true difference if one exists (1-B)

Question 30

Post test probability

Answer 30

= pretest odds x LR+ (positive test) or LR- (negative test)

Question 31

Absolute risk reduction

Answer 31

• Subtraction

Question 32

Criteria for authorship

Answer 32

• Must be involved in conception/design and/or analysis and intrepretation
• Must be involved in drafting the article or revising it
• Must provide final approval
• Must have made a significiant contribution

Question 33

Which studies can never be exempt?

Answer 33

• Fetuses, IVF, prisoners

Question 34

Clinical equipoise

Answer 34

• Defined as honest professional disagreement among experts about the preferred treatment
• NO patient should be randomized to a treatment known to be inferior to the established SOC

Question 35

Therapeutic misconception

Answer 35

• TO deny the possibility that there may be major disadvantages to participating in research that stem from the nature of the research process itself

Question 36

When are both parents REQUIREd to BOTH give informed consent for their child

Answer 36

Greater than minimal risk and no potential for direct benefit

Question 37

When is assent not needed

Answer 37

• Intervention or procedure involved holds the prospect of direct benefit and is only available through the research