case control studies Flashcards
basics of cause case control studies
outcome is starting point look back to the cause or exposure retrospectively, compare exposure history in groups with and without outcome using a control
case control over cohort
rare outcomes
long incubation or latency period
when data on exposure difficult/expensive to collect
dynamic populations avoids loss of follow up in css
methodology
define question/hypothesis then select cases and controls and measure exposures then analyse and interpret results
case definition
need to meet set criteria, broader criteria gives bigger sample sizes but more chance of incorrect identification
and definitions can change with new information
incident and prevalent cases
incident only uses new cases, minimalists recall bias, less likely to have changed behaviour, may have to wait for new diagnosis with rare cases
prevalent new and pre existing cases good for rare outcome and if onset of disease is hard to identify, less likely to recall past exposures and may have changed behaviour
selecting cases
representative sample, source eg clinical setting, routine collected data, disease registries
selecting controls
must be same pop. as cases, ratio of controls to cases changes power, can match individuals or groups
matching can be difficult and time consuming
measuring exposures
type, period of time, availability of resources to assess exposures
analysis
strength association, odds ratio
chance bias and confounding
bias
systemic study design error resulting in incorrect associations
minimise by objective over subjective recall on data and standardise collection
odds ratio/ cross product ratio
AxD/BxC
advantages of css
quick and cost effective
efficient for rare and long latency outcomes
can examine multiple exposures
disadvantages of css
temporal relationship difficult to examine only measures one outcome inefficient for rare exposures need representative sample no incident rates
