Session 7 Flashcards
What do cohort studies select participants based on?
Their exposure status.
They compare outcomes between people who have been exposed and people who have not.
What are the two types of cohort studies, and how do they work?
Prospective - where a sample who have had an exposure are followed to see whether an outcome occurs.
Retrospective - where a sample who’s outcomes have already occurred are analysed to see the effects of an exposure, find the exposures that have occurred first.
What are cohort studies good for?
Looking at rare exposures, as they can follow these people to observe the effect.
Looking at multiple exposures and outcomes, but with one primary exposure.
Can be used to look for a cause of a disease.
Good for showing associations between exposures and outcomes, particularly when RCTs are not feasible - such as when unethical.
What are the drawbacks of cohort studies?
They can take a long time for the outcome to occur.
They can be expensive.
There can be a loss to follow up.
It can be difficult to measure confounding, to really assess the effect of one exposure.
Retrospective studies cannot guarantee the quality of the research and data.
Poor for rare outcomes as the sample size needs to be large.
What kind of biases are cohort studies prone to?
Selection bias - people with certain exposures are selected, which the population as a whole may not have.
Attrition bias - where people drop out of the study or cannot be contacted.
Healthy worker effect - when a cohort study looks at occupational exposures.
Recall bias - incorrectly identifying whether they have had an exposure or not, which can lead to misclassification bias.
When can information bias not be a bad thing?
When it affects both groups equally, it only skews the results when one group is affected more than the other.
Why are participants being recruited at different points in time a bad thing?
When a study stops at one point in time, different people may have had different exposure lengths of time and so they may be more prone to developing the disease.
What are person years?
The total number of years that people were observed from, calculated by adding the length of time from each person in the study together.
How is incidence rate of a cohort study calculated?
Frequency of outcome/ person years.
How is incidence rate ratio calculated?
The incidence rate of one group/ the incidence rate of the other.
It can be for exposed or unexposed groups.
How can the outcome of a cohort study be measured?
Through an odds ratio or a risk ratio.
If the effect is an incidence of an outcome then the incidence rate ratio can be measured.
What are confidence intervals affected by?
The sample size - a smaller sample size will widen the confidence interval.
When can internal and external comparisons be made?
When the sample has been selected from a reference population, external comparisons can be made.
The observed number of cases is divided by the expected number of cases in the general population.
Internal comparisons can always be made.
What type of epidemiological study is a cohort study?
It is an observational, analytical study.
When can confounding not be mitigated?
When it is unknown - this means that the data will be skewed.
