Categorical data II Flashcards
Give examples of prospective studies
cohort studies
longitudinal studies
follow-up studies
Describe prospective studies
identify individuals with relevant characteristics then follow over time and record eventual outcomes
What are caveats for observational studies?
Only able to demonstrate associations between exposure and outcome, rather than causal relationship. A confounding factor may be related to both exposure and outcome seperately
Describe cohort studies
most ideal form of observational study
analogous to a clinical trial (without randomisation)
can follow the natural history of disease in time, from exposure to outcomes
Loss to follow up - connected to outcome is problem
Change in exposure / risk over time
How is the relative risk calculated?
(a/a+c)/(b/b+d)
How are confidence intervals calculated for RR?
logeRR +/- 1.96/SE(logeRR)
Describe hypothesis testing for RR
Ho is that RR=1
z = logeRR/SE(logeRR)
Describe case-control studies
quicker and cheaper if outcome is rare
selection of controls and matching are issues
recall accuracy
Describe the odds ratio
cannot use RR for case-control studies because cases and controls are fixed by study design
OR=ad/bc
How is the OR calculated for matched controls?
OR=b/c