Confounding and effective modification Flashcards
How do you test for effect modifcation
- Crude unajusted measure ofe assoc.
- strata specfic measures of assoc.
- comparison b/w the two and checking of the differance has a relative difference of 15% or more
What its is stratafication and what are its strengths and weaknesses?
- Descriptive and statsistical anayslsis of measure of assocation b/w E and O
- ex: young vs old
- Strength:
- straight fowards, initutive, enhances understanding of data
- weakness
- impractical for those simultaneously control of multiple confounders
Why do we control confounders?
more precise or accurate measure of assocation b/w exposure and outcome
WHat are two main impacts of confounders?
- Magnitude or strength
- More extreme or less extreme
- Direction
- can go from toawrds or away from a null or equal assoc.
What is effect modification?
modifies your measure of assocation by varying different strata within a 3rd variable.
Ex: Race may be a non confounding but if you broke it down african american or asian can be skewwed from the average
If interaction IS present the rearch must report the measures of association of each strata indivdually
What is confounding and how does it affect the measure of assocation?
Confounding is defining able a variable “cofounder” that disporationatly EFFECTING (NOT CAUSUAL) both your exposure and outcomes.
It help adjust your measures of assocatation to show more realistic values.
or
LACK OF APPARTENT EXCHANGABILITY
What is a crude measure of assocation?
A measure of assocation that still includes confounding variables.
funfact:
age is a confounding variable
What is the relationship b/w the confounding variable and exposure and outcome of interest?
- independtly assoc. with exposure
- independently asssocated with outcome
- Not directly in caussal path way linking exposure to outcome (independent)
What is a confounding variable?
A 3rd variable (that distorts an assocaiation b/w exposure and the outcome
or
alternative explanation of assocation
What is matching and what are its strengths and weaknesses?
- matching confounders to each group
- Strengths
- inutitive, some feels like it gives better analytics
- weakness
- difficult, time consumingm and potentially expensive
- doesnt control for an confounders other than thse matched
WHat are the steps of testing for confoudning?
- calculate crude measure of assoc. (OR/RR)
- “unadjusted assocation”
- Calculate outcome between exposire and outcome for each indivdual strat or categories(levels to the study like age)
- “adjusted measure assocation”
- Compare ( (Crude - Adjusted) / Crude) if 15% different then it is compounding variable
What do we want to do with the confounding variable?
Elinate, control or adjust
What is restriction and what are its strengths and weaknesses
- restricits studies to those who dont fall under the confounding variable
- strengths
- straightfoward, convient, and inexpensive
- no impact on internal valdity
- Weakness
- too narrow = negativly impacts ability to enrol subjects
- Not narrow enough introduction of other variables
- not ability to evaluate other strata
- can negatively impact external validity (generalizability)
what is Multivariate anaylsis and what are its strengths and weaknesses?
- stastically anaylsis of data by mathematically factoring out the effects of the confounding variables
- strength:
- simulataneously control for multipe confounding variables
- Stat. reggression, ORs can be obtain and intepreted
- Weakness
- process requres indivduals
- time consuming
