Epidemiological Study design 1 Flashcards
Types of bias in epidemiological studies
- Selection bias
- Information bias (exposure or outcome)
Pros of cross-sectional studies
- Quick
- Cheap and simple
What’s a cohort study
Undertaken to obtain evidence to refuter to support existence of association between cause and disease
Cons of cross-sectional studies
- Not useful for rare exposures/diseases
- Can’t assess correlation as we don’t know whether outcomes of exposures come first
Strengths of cohort studies
- Exposure occurred before disease so measurement of exposure not going to be biases by disease so therefore you can look at multiple outcomes easily
- Little risk of bias
- Good for rare exposures
Weaknesses of cohort studies
- Poor for rare outcomes as would have to study too many people
- Slow and expensive as would have to wait for outcomes
- Attrition bias (people may die)
Bias in cohort studies
- Exposed and unexposed cohorts should be representative otherwise there will be SELECTION BIAS
- ASCERTATION BIAS; when results of study are skewed due to factors you didn’t account for
- ATTRITION BIAS; loss to follow up
What’s case control studies
Define groups based on the outcome
Advantages of case control studies
- Efficient for rare outcomes or diseases with a long latency between exposure and disease
- Cheaper and quick
- Good for populations where follow up is hard
Disadvantages of case control study
- Prone to bias
- Can study only one outcome
- Poor for studying rare exposures
- Cannot allow direct calculation of incidence rates
Bias in control studies
- selection bias
- recall bias
What’s an ecological study
Association of an outcome and an exposure at population level and then compare the populations
Advantages of ecological studies
- Quick and cheap
- Vey early study of an association
- Generate hypotheses to examine at individual level
Disadvantages of ecological studies
- Different areas may have different recording of disease
- Different areas may have differences in measurements of exposures
- Generally not possible to correct for confounding factors
- Relationships found at population level may not be true at an individual level
