Measures of Association Flashcards
What types of studies are pertinent to measures of association?
- RCT
- case-control
- cohort studies
Why are outcomes measured in a variety of ways?
there are different data types
What are the 3 types of outcome measures?
- continuous objective e.g. blood glucose level
- continuous subjective measures e.g. pain levels
- categorical e.g. positive or negative disease test
In which 3 ways can the effects of new treatments with dichotomous outcomes be expressed?
- relative risk
- risk difference
- odds ratio
What do the measures (relative risk, risk difference, odds ratio) show?
they show the direction and magnitude of association between 2 variables
e.g. treatment and outcome
What is the definition of probability?
a measure of the chance of getting an outcome of interest from an event
Between which 2 numbers do probabilities always lie between?
probabilities always lie between 0 and 1
What does a probability of 0 and 1 mean?
a probability of 1 means that the outcome is certain
a probability of 0 means that the outcome cannot happen
In what other way can probabilities be measured?
as percentage chances
this is the probability x 100 so it ranges from 0-100%
What is risk a way of expressing?
risk is a way of expressing treatment effects
What is the definition of risk?
the probability of an outcome or event occurring
How is absolute risk calculated?
by dividing the number of events by the number of people at risk
What would the absolute risk be if 1 in every 100 patients suffers a side effect?
absolute risk = 1/100 = 0.01
What is meant by relative risk?
the ratio of the risk of an outcome in one group (e.g. exposed) compared to the risk of the outcome in another group (e.g. unexposed)
What is the calculation for relative risk?
risk in exposed/treatment group
/
risk in unexposed/control group
What does it mean when relative risk = 1?
the risk in the exposed group is the same as the risk in the unexposed group
What does it mean when the relative risk is < 1?
the risk in the exposed group is less than the risk in the unexposed group
What does it mean when relative risk > 1?
the risk in the exposed group is higher than the risk in the unexposed group
For relative risks <1, how is the percentage reduction calculated?
subtracting the relative risk from 1
(1-RR) x 100
Why is the size of the observed relative risk important?
Large values indicate a strong association
What should relative risk always be interpreted in association with?
confidence interval
What is meant by ‘odds’?
the probability of an event occurring divided by the probability of the event not occurring
How are odds calculated?
by dividing the number of times an event happens by the number of times it does not happen
If 1 in 100 patients suffer a side effect, what are the odds?
1/99 = 0.0101
What must be calculated before odds can be used to interpret results?
an odds ratio
What is an odds ratio?
the odds of the event in one group (e.g. exposed) divided by the odds in another group (e.g. non exposed)
What is the calculation for odds ratio?
odds in exposed group
/
odds in exposed group
What does it mean when the odds ratio = 1?
the odds in the exposed group is the same as the odds in the unexposed group
What does it mean when the odds ratio < 1?
the odds in exposed group are less than the odds in the unexposed group
What does it mean when the odds ratio > 1?
the odds in the exposed group is higher than the odds in the unexposed group
What values can an odds ratio take?
any value from 0 to infinity
What is the 95% confidence interval used for when reported with RR or OR?
It is used to estimate the precision of the OR or RR
What does a large and small confidence interval show when presented with OR or RR?
Low CI shows low precision of the OR, RR
Small CI shows high precision of the OR, RR
What does it mean if the confidence interval crosses 1?
this implies no statistical significant difference between exposed and unexposed groups
When is RR usually used in a study?
In studies where the subjects are initially chosen by their exposure and followed up over time to see their outcome
e.g. cohort studies, randomised controlled trials
When should relative risks not be used?
case-control or other retrospective studies
When should odd ratios be used in a study?
in case-control or other retrospective studies
they are commonly reported in logistic regression analysis
If a disease/outcome is rare, what is the difference between RR and OR?
The RR will be similar to OR
What is meant by risk difference?
the difference in risk of outcome between the exposed/treatment group and the unexposed/control group
Why is risk difference used in clinical practice?
it is meaningful for understanding if findings of a study are worthy of incorporating into practice
Does risk difference always warrant changing clinical practice?
the magnitude of the risk difference may not warrant changing practice
What is another way of evaluating clinical importance of treatment?
number needed to treat (NNT)
What is the calculation for NNT?
1/risk difference
When is NNT usually used?
in studies comparing an intervention and standard treatment
What is Akobeng’s definition of NNT?
the number of people who need to receive the treatment in order to achieve the required outcome in one of them