Measurement: Scales, Numbers, Rates, Ratios and Risk Flashcards
What are the 2 types of error that can occur in a study
How does sample size affect the errors
Chance (Random error):
- Due to sampling variation
- Reduces as sample size increases
(More precision/ less uncertainty)
Bias (Systematic error):
- Quantified by difference between TRUE and EXPECTED value
- Doesn’t reduce as sample size increases
What are the 2 types of Bias
Selection Bias
Information Bias
What are 3 sources of Selection Bias?
Explain them
- External Validity (Study sample)
Sample not representative of entire population) - Internal Validity (Group selection within a study)
Groups within a study may not be comparable - Healthy worker effect
Workers usually have lower mortality than general population
What are 4 sources of Information Bias?
Explain them
Which error usually arises from another
- Recall error
Difference in recollection (amongst study participants ) about past events/ experiences - Observer/ interviewer error
Study observer may have preconceived ideas that can affect result - Measurement error
Differences in measurements of participants (Using same tool) - Misclassification
Participants put in wrong group. Usually arises from Observational OR Measurement
Compare Large Bias and High Precision
What is the ideal combo of bias and precision?
Large Bias: Continuous results that are very “Off-target”
High Precision: Results are very similar
No Bias+ High Precision
Define Prevalence
What is it useful for?
The proportion of people who have a disease at a given point in time
Useful to determine resource/ service allocation
Define Incidence
What is it useful for?
What is the equation for Incidence rate?
The number of new cases within a given timeframe
Useful when monitoring epidemics
IR: Number of new cases/ Sum of {patient time at risk*}
*No. of patients * Time they were at risk
What is the Incidence Rate Ratio?
The IRR compares the incidence rate in one group to another
What is Relative Risk also known as?
What is Relative Rate also known as?
What is the difference between the 2
Relative Risk= Risk Ratio
Relative Rate= Rate Ratio
Relative Risk: Underlying quantity we want to approximate
Relative Rate: Approximates relative risk if disease is rare/ if time-period is short
How is Odds different to Probability, p
How can an odds of 1 also be written
Rather than odds, what is more used to compare 2 groups
How do you calculate this for 2 groups, A and B
Odds= p/ (1-p)
1:1
Odds Ratio
Odds of Group A/ Odds of Group B
How do you calculate Absolute Risk of disease in a group
How do you calculate Relative Risk between 2 groups, A and B
What is Risk Difference
No. of diseased people/ Total people= Absolute Risk
Relative Risk: Absolute Risk for A/ Absolute Risk for B
Absolute risk in A minus Absolute risk in B
What is a confounder?
A factor that affects both Exposure AND Outcome
How can we adjust for differences in known confounding factors?
Why may this not work?
Standardisation: Using weighted averages to allow us to compare “like for like”
A lot of confounding factors are unknown
