Epidemiology:study Design

Front 1

Relative Risk

Back 1

-Risk of developing a disease relative to exposure.
-Comparing prevalence/incidence of dis. in 2 groups (1 exposed,1 not)
-The ratio of prevalence/incidence in 2 groups is the RR assoc. with that risk factor

Front 2

Attributable Risk

Back 2

-Diff in rate of condition bet exposed/not exposed
-Additional risk of dis. following exposure above that experienced by those not exposed
-(prev/inci of non exposed) -(prev/inci of exposed)
-Assumes risk in non exposed group = background risk of dis.
-bigger the AR the bigger the RF under observation

Front 3

Sources of Data

Back 3

-Repr. of pop
-Systemic Random Sampling=practicle,ie every 3rd animal through gate
-Stratified Random Sampling=divide pop into sub pop called Strata,random sample from each strata,same elements withing strata,incre.precision & deals with confounding factors
-stratify according to colour,hair etc
Cluster sampling=Div. into clusters of units,,select random sample of clusters and observe every unit,indiv. in clusters as diff as poss.,diff bet clusters a small as poss ie farm=cluster,herd=unit
Multistage Sampling=random sampling at diff hierarchal levels ie randomly select clusters=primary unit-herd,race yard,in each cluster select animals to test at random as the sampling unit=secondary unit,ie indiv cows

Front 4

Sample Size

Back 4

Power required
-sig level
-S.D.
-Signi. difference

Front 5

Target population

Back 5

=pop at risk
-study pop=pop that is sampled
Target pop->study pop--random sampling->sampling frame
Aims=
Precise (low variance)
Accurate (low bias)

Front 6

Estimation of disease occurence

Back 6

n=z2p(1-p)/d2
p=probable prevalence
z=desired confidence level (90% +/- 1.65se or 95% +/- 1.96se)
d=level of precision