Epidemiology Module 3 Flashcards
(31 cards)
What are the measures of association?
Risks, rates, prevalences, odds ratio
What is the cumulative incidence (CI)?
Risk. R=# people with new outcome/total population at risk
What is the incidence density (incidence rate) (IR)
new cases/total person-time of observation
How to provide information on absolute differences in outcomes among exposed and unexposed?
Additive measures. RD or IRD
What is the risk difference (RD)?
RD=R1-R0. R1=incidence proportion exposed and R0 = incidence proportion unexposed
How do you interpret RD?
RD= # per 100,000 = exposed experience # additional cases per 100,000 compared to the unexposed.
What is the incidence rate difference (IRD)?
IRD=IR1-IR0. IR1=incidence density exposed. with person-time.
How do you interpret IRD?
per 100,000 person-years. the exposed expierence # additional cases per 100,000 person-years
How do you provide information about the probability of the outcome in the exposed relative to the unexposed?
Multiplicative measures. RR, incidence rate ratio (IRR), odds ratio (OR)
What is the risk ratio (RR)?
RR=R1/R0=P(D|E)/P(D|noE)
How do you interpret RR?
times as likely to get outcome or % more/less likely. # times the risk of the outcome among unexposed.
What is absolute risk reduction (ARR)?
ARR=R0-R1 or placebo and treatment.
What is relative risk reduction (RRR)?
% reduction in risk in exposed group compared to unexposed group. RRR=(R0-R1)/R0=1-RR
What is number needed to treat (NNT)?
NNT=1/(R0-R1)=1/ARR. number of persons treated to prevent one outcome, i.e. the number of persons that have to be exposed to the risk factor to develop one case. e.g. for ever 11 women that smoke during pregnancy, one additional low-birth-weight baby will be born compared to women who did not smoke.
What is the incidence rate ratio (IRR)?
IRR=IR1/IR0. has person-time
How do you interpret IRR?
the rate of the outcome among the exposed is 3x the rate of the outcome among the unexposed, or x% more/less than.
What is the prevalence ratio?
PR=PR1/PR0.
How do you interpret PR?
x as likely or #% more/less likely or #x the prevalence of the outcome
What is the odds ratio (OR)?
ratio of the odds of exposure among those with the outcome to the odds of exposure among those without hte outcome. OR=ad/bc=[P(E|D)/P(noE|D)]/[P(E|noD)/P(noE|noD)]
What is a strict interpretation of OR?
odds of the outcome among the exposed is #x the odds of the outcome among hte unexposed.
What is a regular interpretation of the OR?
The exposed are #x as likely to have the outcome as the unexposed
When can OR be used?
for incidence or prevalence data
What is a drawback of the OR?
may over-estimate effect
When should you report confidence intervals and why?
measures of frequency and association are best reported with confidence intervals. used for estimate of precision of measure. equivalent to a two-sided hypothesis test (i.e. if 100(1-a)% CI contains the null value, then the result is not significant at the a level)
