Measures of Effect Flashcards
Which ratios are relative measures of effect?
Risk Ratio & Odds Ratio
What is the definition of relative measures of effect?
They describe how an exposure changes the risks or odds of an outcome compared to the risks and the odds of an unexposed group
Which are the examples of absolute measures of effect?
Absolute risk
Number needed to treat
What is the calculation of risk ratio/relative risk?
Risk of exposed/Risk of unexposed
What is the interpretation of a RR < 1?
The risk of a certain outcome is lower in the exposed group than it is in the unexposed
What is the interpretation of RR > 1?
The risk of a certain outcome is higher in the exposed group compared to the unexposed group; exposure is a risk factor for the outcome
What is the interpretation of a risk ratio = 1?
This means there is no difference at all between the risk of the exposed vs. the unexposed groups
What is a confidence interval?
It is the range with whichin the actually risk ratio lies in with a probability of 95%
When is the difference between two groups statistically significant?
If the confidence interval does not include 1
What is the definition of risk?
The probability of an outcome which can be both positive or negative
How would you interpret an RR of 1.5?
The risk of the exposed group getting affected is 1.5 times greater than the risk of unexposed people getting affected, or 50% higher
What is the calculation for relative risk difference?
Risk Ratio - 1
What is absolute risk difference?
Risk of exposed group - Risk of unexposed group
What is the calculation for Odds Ratio?
Odds of exposed group/ Odds of unexposed group
What is the Odds Ratio?
Relative measure of effect
What is the definition of “Odds”?
The chance of having a certain outcome
What is the interpretation of OR < 1?
The chance of a certain outcome is lower in exposed group
What is the interpretation of OR > 1?
The chance of a certain outcome is greater in the exposed group
What is the interpretation of OR = 1?
There is no difference for the chance of a certain between an exposed group vs an unexposed one
When does the CI of Odds Ratio show statistical significance?
When the Odds Ratio does not contain 1
What is the difference between risks and odds calculations?
To find the risk of an outcome –> people with an outcome/all people
To find the odds of an outcome –>
people with an outcome/people WITHOUT an outcome
What are the differences between RR and OR?
OR –> all study types, more often used for case-control
RR –> cannot be used for case-control studies
Interpretation of RR is easier than OR, so RR is preferred
In which cases are RR and OR very similar?
When the outcome is really rare
What is the number needed to treat calculation?
1/absolute risk difference
What is the definition of number needed to treat?
Number of people you need to treat in order to help one person
OR
Number of people that need to be exposed to a treatment or risk factor for one single outcome to occur that would not be otherwise occurring
What is the purpose of the number needed to treat?
To measure the benefit of a treatment
In terms of a treatment (medication) what does a low NNT indicate?
Treatment is very effective –> very few people need to take the medicine in order to see an improvement
What is the difference between the number needed to treat vs the number needed to harm?
The number needed to harm indicates that the exposure has a negative effect on the group
What is the survival function?
The probability of exposed or unexposed group surviving past a specific time period (for instance; number of years)
What is the calculation of the survival function?
Number of patients surviving longer than t/ total number of patients
where t = a specific time period
What is the hazard function?
The probability of dying in the next few seconds, given that they are alive now, past the survival function
What is the purpose of the Kaplan Meier Plot?
It allows us to make statements on the survival rate based on time
Is Kaplan Meier Plot parametric or non-parametric?
Non-parametric model
What is the difference between a parametric and a non-parametric model?
Non-parametric tests are not seriously affected by outliers
