Observational Studies Flashcards
What is the equation of point prevalence
The number of people with disease at same time / total population at risk of disease at same time
What are 3 types of observational studies
- Cross sectional study
- Case control study
- Cohort study
What is a cross sectional study
A study that looks at the prevalence of a disease at one time point only. There is no follow up so exposure and outcome is measured at the same time.
What is a case-control study
You take people with the disease and look back at risk factors back in time.
What is a cohort study
You take people that are exposed and unexposed then follow in time prospective to identify if they develop the disease.
What are the 3 factors to consider in observational study that can change data
Bias
Confounding
Chance
What are the 2 main categories of bias
Selection bias
Information bias
What are the types of biases within selection bias
Sampling bias
Response bias
Healthy worker effect
Healthy reproducer effect
What is sampling bias
Some people may be chosen more than other so this would not represent the general population
What is response bias
If the study is using questionnaires some people may not respond so there is a question whether the groups of people that responded compared to people that did not respond are similar
What is the healthy worker effect
People that work are likely to be healthy compared to those that do not work as they can have a disability
What is the healthy reproducer effect
If Women that have children are selected can be more healthy than women that do not have children as they can be infertile so are unable to have children.
What are the types of bias within information bias
Recall bias Lost to follow up bias Social acceptability bias Interviewer bias Recording bias
What is recall bias
When people with an illness are likely to remember certain things more than healthy people
What is a recording bias
When ill people are more likely to have more extensive medical records compared to healthy people
What is a interviewer bias
If the interviewer knows the individual has the disease or not
What is social acceptability
When the person being interviewed knows what is is socially acceptable so may not be honest about drinking, smoking etc.
What is lost to follow up bias
When people that are lost are likely to have the outcome of interest drop out than those that are still within the study.
What is confounding
This is when there is a 3rd factor that is giving an association.
What is an ecological study
A study where observations are made as a whole group rather than individuals e.g regions such as a county or country etc.
How do we adjust for a confounders
With a suitable regression model
What are the biases that can occur in a cohort study
Healthy worker effect
Lost to follow up
What are the advantages of an cohort study
You can look at rare exposures
You can look at multiple diseases as an outcome
It eliminates recall bias as it can be prospective and not retro prospective
What are the disadvantages of cohort studies
Not useful for exploring rare disease as fewer individuals will develop it
What are the advantages of case-control studies
You can explore rare diseases
You can explore multiple risk factors
What are the disadvantages of case-control study
You can only look at one disease
Recall bias as it is retro prospective
You cant estimate absolute risk because we start with a diseased status
What are the biases that can occur in case-control study
Recall bias Recording bias Interviewer bias Response bias Sampling bias
In a case control study what are we interested in exploring
Odd ratio
If we do not find any bias, confounders or chance how do we show that there is a causality
Bradford hill criteria
