Lecture 5 Flashcards
1
Q
What is the time dimension of cohort
A
look at people over time… exposure to outcome
2
Q
what are potential biases in cohort
A
- nonresponse bias and loss to follow up
- bias in assessment of outcome
- information bias
3
Q
What is a confounder
A
3rd variable alters association between exposure and disease
- associated with both exposure and disease, not on causal pathway
4
Q
how to deal with confounding
A
- matching
- adjusting (ex: regression analysis)
- randomization
5
Q
when to use a cohort
A
- when we have an idea on exposures
- when we can minimize loss to follow up
- relatively short interval between exposure and disease
6
Q
when not to use a cohort
A
- lack of evidence to justify large and expensive study
- rare diseases
7
Q
advantages of cohort
A
- incidence
- rare exposures
- temporal relationship
- time to event analysis
- multiple outcomes
8
Q
disadvantages of cohorts
A
- lengthy and expensive
- may need large samples
- not for rare diseases or long latency
- environmental changes can influence
- loss to follow up
- sub clinical disease may go undetected at baseline
9
Q
When to use case control
A
- rare diseases
- new diseases
- diseases of little known etiology
- diseases which medical care usually sought
10
Q
Why is clear case definition important
A
sensitivity = correctly identify those with disease
Specificity = correctly identify those without disease
11
Q
what are limitations to case control
A
- not for weak associations
- need fair sample size
- misclassification of exposure
- recall bias
- reporting accuracy bias
- other forms of info bias
- for prevalent cases, hard to establish temporality
- finding good case and control groups
12
Q
what are advantages of case control studies
A
- relatively inexpensive
- can study multiple exposures
- can conduct in relatively short time period
- generally require less cases and controls
