Lecture 4

Question 1

What is the objective of Case-Control?

Answer 1

To find possible predictors of outcomes

Question 2

What are the overall steps for case control?

Answer 2

• cases from a specified population are sampled
• controls are sampled
• the 2 groups (case vs control) are retroactively compared
• calculate odds ratio

Question 3

Can cases and controls come from different populations?

Answer 3

absolutely not! They must be as similar as possible

Question 4

What’s a good study method to to study a uncommon condition? DOUBLE CHECK

Answer 4

Case-control case! It allows us to maximize the effort and finances gone to recruit and find people from this small population

Question 5

What are the limitations of case control studies?

Answer 5

More vulnerable to biases, Can’t consider more than one disease, not feasible for rare exposures

Question 6

What biases in particular affect case control studies?

Answer 6

1. Sampling bias (unrepresentative samples) 2. Recall bias (retroactive and based on recall)

Question 7

How can sampling bias be solved?

Answer 7

random sampling

Question 8

How can recall bias be solved?

Answer 8

recorded data`

Question 9

What is the objective of a RCT?

Answer 9

To observe treatment effect

Question 10

What is the definition of a RCT?

Answer 10

Quantitative, comparative, and controlled experiments in which a group of a investigators studies two or more interventions by administering them to groups of individuals who have been randomly assigned to recieve each intervention.

Question 11

What’s a good alternative to having a control group with no treatment at all?

Answer 11

control could have generic/current treatment vs new treatment. there should be a difference

Question 12

Why are we so worried about confounding variables?

Answer 12

We might not be able to correctly see the pattern that leads from variable of interest to affect. Other things could influence our data!

Question 13

Define blinding

Answer 13

Being unaware of the condition in which a participant is assigned

Question 14

What are the types of RCTs by binding?

Answer 14

1. not blinded: both researchers and participants are aware of the group they are a part of
2. Single-binding: only patient is unaware of the group
3. Double blinding: both the treating physician and participant is in the intervention or the control condition
4. Triple blinding: participants, treating physicians, and the trial investigator are all blinded (via coded values so even statistician is )

Question 15

List the Types of RCT by phase with brief description

Answer 15

Phase 1: is it safe? (tested on willing human volunteers)
Phase 2: does it work?
Phase 3: how well does it work?
Phase 4: is there any adverse affects that we could have missed?

Question 16

Limitations of RCT?

Answer 16

May not be ethical, or feasible. Recall bias again (when collecting history for ex), small number of participants can be inaccurate

Question 17

What are hill’s criteria of causality?

Answer 17

1. Strength (strong association) 2. Consistency (same result across different studies and situations) 3. Specificity (very specific association and effect) 4. Temporality (effect has to occur after the cause) 5. Biological gradient (greater exposure should lead to greater effect) 6. Plausibility (consistent with current body of knowledge) 7. Coherence (epidemiological findings and lab findings agree)

Question 18

What are the 3 types of observational studies?

Answer 18

Cohort, Prevalence, and Case control

Question 19

How is the defined population split up in a case-control study?

Answer 19

First disease (cases) and not diseased (controls) are split, then those two categories each are split into those exposed/not exposed.

Question 20

What are case control studies useful for?

Answer 20

• studying uncommon conditions (lots of useful information with limited participants)
• generating a hypothesis for further studies
• Require less resources and less time consuming

Question 21

What is the main strength of RCTs?

Answer 21

randomizing subjects means that on average all other possible causes are equal between the two groups