Observational studies Flashcards
What is a cohort study
- a group of individuals (cohort) who may share common characteristics are followed over time during which exposure/s and incidence of outcome/s are recorded
- aim to compare outcomes in the exposed and unexposed groups at the end of the study period
- can be prospective and retrospective
What is a case control study?
People with a condition (cases) are compared with a random sample of a similar population who do not have the condition (cases) with respect to their past exposure to the factor of interest
Exposure more common in cases, could be a risk factor
Exposure less common in controls, could be protective factor
What are the advantages and disadvantages of cohort studies
ADS
- stable for rare exposure or multiple outcomes
- temporal relationships can be assessed
- can estimate the risk of developing a disease in the population of interest
DISADS
- time consuming and expensive
- not suitable for rare outcomes or outcomes with long period between exposure and onset
- prone to selection, information and attrition (loss of follow up biases)
Advantages and disadvantages of case-control studies
ADS
- cheap and quick to perform
- most efficient design for rare diseases or diseases with long periods between exposure and onset
DISADS
- not suitable for rare exposure or multiple outcomes
- prone to recall, observer and selection biases
- cannot estimate the risk of developing a disease in the population of interest
Measures of association/ effect
- a basic aim of observational studies to quantify an association between the risk of disease (or other health outcomes) and the exposure risk factor
- we do this by comparing the occurrence of disease in the exposed group with the occurrence of the disease in the unexposed group
This is assigned by estimating the ratio (a relative measure) of the group (or odds) of developing the disease between two groups
How to measure association
Estimate the strength of association between disease and exposure
Risk ratio (RR)=
Risk in exposed group/ risk in unexposed group (cohort studies)
Odds ratio (OR) =
Odds in exposed group/ odds in unexposed group (cases control studies *)
- it is not possible to estimate the risk/ risk ratio of developing a disease in the population of interest
How to calculate risk
The number of new cases of a disease during periods
————————————————————
Total number of persons disease free at start of period
Example:
25 people in a 100 get a cold in a particular year = 25/100= 0.25, 25% or 1:4
How to calculate the odds
Number of persons who developed disease in a period
————————————————————
Number of persons who did not develop disease in a period
Example:
- 25 people in 100 get a cold in a particular year = 25/(100-25)= 25/75= 0.33, 33%, 1:3
Risk ratio and odds ratio interpretation
RR (or OR)
= 1
(no difference in the risk/ odds of developing disease between groups, exposure is potentially not a risk factor
> 1
(the risk/ odds are higher in the exposed group, exposure in a potential risk factor for the disease)
<1
( the risks/ odds are lower in the exposed group
————————————————————
Because risk and odds ratio are estimated from a sample, they should always be presented with 95% confidence intervals
How would you calculate the risk ratio and odds ratio
The 2x2 table and risk and odds formulae
- refer to presentation slides
What is meta-analysis
It is a statistical technique for summarising the results of several studies into one single effect estimate
The overall effect is calculated by averaging the individual studies weighted effect -> fixed/ random effects
Data usually used in meta-analysis are:
- mean different, risk difference
- risk ratio, odds ratio
It is used to produce more objective summaries combining multiple studies rather than intellectual interpretation of a single study
Considered the highest level of scientific evidence
