31. CRUDE AND ADJUSTED ESTIMATES Flashcards
- What are Crude Estimates?
- these are the estimates for the Exposure-Outcome
Association - before we apply any adjustments
- these are the original Association estimates
EXAMPLES:
- Odds Ratio
- Mean Difference
- etc.
- What are Adjusted Estimates?
- these are the estimates for the Exposure-Outcome
Associations - after we have applied Statistical Adjustment
- for any Potential Confounder
- What do Adjusted Estimates aim to do?
THEY AIM TO ANSWER THE QUESTION:
- what would the exposure-outcome association be if
everyone was effected the same
- if everyone had the same value in terms of the
Potential Confounder value
- What 6 effects can the adjustment have on the estimate of any given Exposure-Outcome Association?
- The association becomes weaker
- The association disappears
- The association appears
- The association becomes stronger
- The association is reversed
- The association remains unaffected
- What happens to the Association in this example?
THE ASSOCIATION BECOMES WEAKER
- the risk is decreased
INTERPRETATION:
- the potential confounder explains a small part of the
association between Exposure and Outcome
NB:
- the adjusted estimate can still be biased
- since we have not adjusted for all the Confounders
possible
- What happens to the Association in this example?
THE ASSOCIATION BECOMES WEAKER
- the risk is decreased
- there is a large difference between the Crude and
Adjusted Estimates
INTERPRETATION:
- the potential confounder explains a large part of the
association between Exposure and Outcome
- What happens to the Association in this example?
THE ASSOCIATION DISAPPEARS
- the association is no longer significant
- there is a small decrease between the Crude and
Adjusted Estimates
INTERPRETATION:
- the potential confounder explains all of the association
between the Exposure and the Outcome
AFTER ADJUSTING THE CONFOUNDER:
- there is no association between the Exposure and the
Outcome
- What can be said about the highlighted values?
THE VALUE HIGHLIGHTED IN BLUE:
- is less than one
- this means that it is a Protective Factor
THE VALUE HIGHLIGHTED IN PINK:
- is greater than one
- this means that it is a Risk Factor
- What happens to the Association in this example?
THE ASSOCIATION DISAPPEARS
- the association is now significant
INTERPRETATION:
- the potential confounder was masking the association
- adjustment has made the association appear
AFTER ADJUSTING THE CONFOUNDER:
- the highlighted value is no longer less than 1
- it is now a Risk Factor
- What happens to the Association in this example?
THE ASSOCIATION BECOMES STRONGER
- the risk is increased
INTERPRETATION:
- the potential confounder was diluting the association
- adjustment has made the association stronger
AFTER ADJUSTING THE CONFOUNDER:
- the highlighted Confidence interval has become wider
- What happens to the Association in this example?
THE ASSOCIATION BECOMES REVERSED
- the association is repaired
- the highlighted value became larger than 1
- it is now significant and a risk factor
INTERPRETATION:
- the potential confounder was reversing the association
between the Exposure and Outcome
AFTER ADJUSTING THE CONFOUNDER:
- the Association went from negative to positive
- What happens to the Association in this example?
THE ASSOCIATION REMAINS UNAFFECTED
- the change in estimates is too small to be significant
INTERPRETATION:
- the potential confounder has no effect on the
Exposure-Outcome Association
AFTER ADJUSTING THE CONFOUNDER:
- we see that this was not a Confounder in this specific
association
- What happens when we adjust for Mediators?
- we treat them exactly the same as Confounders
- they may exactly the same effects on the Exposure-
Outcome Association as the Confounders - the only difference is the Interpretation
- What happens if a factor that lies in the Causal Pathway between exposure and outcome
(such as a Mediator) is treated like a Confounder?
IT IS ADJUSTED AS ONE:
- this may lead to an underestimation of the true
Exposure-Outcome Association
- this is called Over-Adjustment
- What happens if the researchers adjusted for some of the potential Confounders, but not all of them?
WE HAVE AN OVERESTIMATION
- of the true Exposure-Outcome Association
- this is known as Residual Confounding
THIS CAN BE AVOIDED:
- by accurately measuring all of the Confounders
- by adjusting as many known Confounders as possible
