EPI Flashcards
screening: pop
has to matter to population
must have high risk group
early intervention with positive impact
screening test
sensitive and specific
test must be acceptable
minimal risk
diagnostic work up for positive test must be tolerable
false positive pt - no harm for screening
screening outcomes
decreased death, morbidity, interventions
lead time bias
increase in survival as measured form dx to death w/o actually lengthening life (just dx’ed earlier)
length biased sampling
the impression that detecting cancers through screening causes cancers to be less dangerous, when the reality is that less dangerous cancers are simply more likely to be detected by screening
match to decrease ______ bias
confounding
PICO
population/pt/problem
intervention
comparison
outcome
OR=1
no association btwn exposure and outcome
OR >1
harmful, exposure associated with increased odds of outcome
OR <1
protective, exposure associated with decreased odds of outcome
P Value signif
P < .05
Incidence
new cases
new drops of h20 in sink
“well”
people at risk
Incidence equ
new cases/ # of people at risk (/time)
attack rate equ
new cases in pop @ risk / # persons @ risk in population
mortality rate equ
of deaths / # ppl @ risk of death (denominator includes the “well ppl”
case fatality rate equ
of deaths per cases/ # of cases (denominator would be the prevalence)
mortality rate
drops down the sink drain
prevalence
“snapshot” picture of population (sink) at a certain time
prevalence equ
of ppl with dz/ population @ specific time
prevalence helps with…
Burden - helps direct resources.
4 types of foreground questions
intervention
dx
harm
prognosis
statistical power
ability to detect the true difference btwn groups
type 1 error
rejected when it is right
type 2 error
kept (failed to reject) when it’s wrong
Risk equ
of outcome/ (exposed) or (unexposed)
RR
R(exposed)/R(unexposed)
Relative risk
is the ratio of the probability of an event occurring
RR = 1
exposure not related to outcome
null
exposed = unexposed
RR > 1
exposure increases risk of death
RR < 1
exposure decreases risk of death
sensitivity
tp/tp+fn
specificity
tn/tn+fp
NPV
tn/tn+fn
PPV
tp/tp+fp
case series bias
investigator bias
ecological drawback
ecological fallacy - cant make conclusions for population based on individuals
cross sectional bias
recall
case control bias
selection, no randomization
cohort bias
selection bias, particpant drop out
