Observational Studies Flashcards
Why are observational studies done?
Hypothesis generating, RCTs are expensive, study of rare events, ethics
What is the difference between cases and controls?
Cases have the disease of interest
Controls don’t have the disease and are usually age and sex matched with cases
What is the selection bias in case-control studies?
Cases may not represent exposure distribution in all cases in the source population, control may not represent exposure distribution in those without disease
How is selection bias minimised?
Minimised at design stage = ensure cases and controls are representative of and come from the same source population
What is observer bias in case-control studies?
Knowledge of case/control status may influence data collection
Impact = identify more (spurious) risk factors in cases
How is observer bias minimised?
Use of standardised objective instruments and blind researchers to case/control studies
What is recall bias and how can it be minimised?
Cases and controls recall prior exposures differently
Minimise period of recall and measure exposure data objectively (medical notes, third party verification)
What does survivor bias cause?
Will detect factors that increase survival among the diseased as risk factors for the disease
What are the strengths of case-control studies?
Rapid and cheap, ideal for rare diseases/outcomes, useful for diseases with long latent periods, can simultaneously examine a large number of potential exposures
What are the weaknesses of case-control studies?
Bias, temporal relationship can be difficult to establish, can’t compare incidence rates
What are the uses of cohort studies?
Impact of infrequent/unusual exposure, multiple outcomes related to infrequent exposure, disease incidence, temporal sequence, how risk changes over time
What questions must be considered when calculating risk?
What is the risk of outcomes for those exposed?
How do the risks compare?
How is the risk of disease in exposed individuals calculated?
Disease present/(disease present + disease absent)
How is the risk of disease in unexposed individuals calculated?
Disease present/(disease present + disease absent)
What is the relative risk ratio?
Compares risk across exposure groups
Risk of disease in exposed/risk of disease in unexposed
What does the relative risk ration mean?
RR >1 = exposure harmful
RR <1 = exposure protective
RR = 1 then exposure does nothing
What determines the type of cohort study done?
The type of follow up needed
What are some features of prospective cohort studies?
Time consuming, expensive, defined cohort (detailed exposure records with standardised measurement)
What are some features of retrospective cohort studies?
Faster answers, don’t have to employ people to conduct regular follow-up (cheap), quality of records must be checked
What are the advantages of cohort studies?
Temporality, no recall bias, can study multiple outcomes associated with rare exposures, measures incidence
What are the disadvantages of cohort studies?
Requires large investment of time and money if prospective, requires large sample size, loss of follow-up bias, inefficient for rare diseases, uncontrolled confounding
What biases are present in cohort studies?
Selection = initial cohort not representative of population, loss of follow up Information = misclassification of exposure/outcome Confounding = unmeasured day-to-day exposures
What is the odds ratio in case-control studies?
Risk of disease given exposure, when disease studies is rare
What is a cross-sectional study?
Carried out a single point in time, most frequently as a survey
What are the benefits of cross-sectional studies?
Particularly good for estimating point prevalence, quick and cheap
What is the disadvantage of cross-sectional studies?
Can’t establish causation
What should be considered in a critical appraisal?
Where research came from, how the research was done, what research showed, how reliably it can be applied to clinical care
What do CASP checklists contain?
Set of eight critical appraisal tools
What is the null hypothesis?
Hypothesis that there is no significant difference between specified populations
What does a p value <0.1 mean for the null hypothesis?
There is weak evidence against the null hypothesis
What does a p value <0.05 mean for the null hypothesis?
There is strong evidence against the null hypothesis
How is the null hypothesis affected if the p value is <0.01?
There is very strong evidence against the null hypothesis
