Session 2 Lecture 1: Measurement-Scales, Numbers, Rates, Ratios and Risk Flashcards
What are the 3 key features of the sample population taken to make an inference about the entire population?
- Representative
- Unbiased
- Precise
State the two type of error that can occur in a study that may influence the results
- Chance (Random error)
- Bias (Systematic error)
How does Chance (Random error) occur?
- Due to sampling variation
- will reduce as sampling size increases
How does Bias (Systematic error) occur?
- Quantified by the difference between the true value and the expected value
- Does not reduce as sample size increases, bias remains the same
Where does the bias come from that leads to Systematic error?
- Selection biases
- Information biases in the data
Examples of selection biases which are a source of bias that can lead to Systematic error
-Study Sample (External Validity)
not representative of pop of interest
-Group selection within a study (Internal Validity)
groups within a study may not be comparable
-Healthy worker effect
workers usually exhibit lower overall mortality than the general pop
Examples of Information biases in the data which are a source of bias that can lead to Systematic error
-Recall error
differences in recollection from study participants regarding events or experiences from the past
-Observer/interviewer error
study observer/interviewer may have preconceived expectations or knowledge that may influence the result
-Measurement error
Differences in the measurements of participants
-Misclassification
participants put in wrong group
Define the term Prevalence
The proportion of people who have a disease at a given point in time
Calculating prevalence
Number of people with disease / total population (number of people)
Define the term Incidence
The number of new cases of a disease within a given timeframe
Calculating Incidence rate
Number of new cases / sum of the patient time at risk
- patient time at risk = Sum of all the patients times at risk
often reported as a rate: events per person per year (50 per 100000 person-years)
What is the Incidence Rate Ratio used for?
- To compare the incidence rate in one group to another
- relative measure between 2 groups
IRR = incidence rate in group A / incidence rate in group B
Group A= group you are interested in
Group B=comparing group
Calculating Odds
If the probability of an event is p then the odds of that event is given by:
Odds=p/1-p
What is the ratio of odds used for?
To compare two exposure groups
What is the Odds Ratio?
A relative comparison of the odds of an event happening in Group A compared to Group B
“The ratio of ratios”
How is the Odds ratio calculated?
odds of group A = a/b
——————————– = ad/bc
odds of group B= c/d
How is the absolute risk calculated?
Group A = a / a+b
Group B = c / c+d
Calculating relative risk
Absolute risk for group A/Absolute risk for group B
What is the null value for the Odds ratio and relative risk?
1
What is the interpretation if the Odds ratio (OR) and relative risk (RR) are < 1 respectively?
OR < 1 = numerator has a lower odds of the event
RR < 1 = numerator has a lower risk of the event
What is the outcome of the event when the desirable interpretation of odds ratio and relative risk is < 1?
Desired if the event is a bad outcome
What is the interpretation if the Odds ratio (OR) and relative risk (RR) are > 1 respectively?
OR > 1 = numerator has a greater odds of the event
RR > 1 = numerator has a greater risk of the event
What is the outcome of the event when the desirable interpretation of odds ratio and relative risk is >1?
Desired if the event is a good outcome
Calculating the Risk difference
Absolute risk in Group A - Absolute risk in Group B
=Absolute difference
What is the null value for Risk difference?
0
What are the issues to be aware of when comparing groups?
Confounding issues
What do Confounding issues affect?
They affect both exposure and outcome.
When comparing groups, the association or effect between an exposure and outcome is distorted by the presence of another variable
What can be done about confounding factors?
In stats, we can adjust for differences in known confounding factors
-Standardisation
however, a lot of confounding factors are unknown
