Stratification and Adjustment Flashcards Preview

epi3 > Stratification and Adjustment > Flashcards

Flashcards in Stratification and Adjustment Deck (16)
Loading flashcards...

multivariate analysis

analysis involving multiple dependent variables


multiple/multivariable regression

analysis involving multiple independent variable


Two types of multivariable analysis

1. stratification
2. regression


stratification requirement

1.requires there be no more than two or three stratification variables (in theory, 6)
2. stratification variables must be categorical


example of stratification

direct adjustment
indirect adjustment
Mantel-Haenszel method


regression requirement

1. can handle multiple variables
2. adjustment variables can continuous or categorical


examples of regression

linear regression
logistic regression
poisson regression
cox proportional hazard model


Def of stratification

separation of a sample into subgroups according to the levels of one or more discrete or categorical variables

used best when potential confounders or effect modifiers are few in number and relatively few categories


stratification and regression

analytical tool that are used to control for confounding, effect modifications and to summarize the associations of several predictor variables with disease risk in an efficient fashion


two methods of adjustments



why do we use adjustments

to compare summary measures between different groups

use of weighted averages to remove confounding

to compare oranges to apples


adjustments are only good when interactions are not present

when any difference in estimates across strata is due only to random variation; homogeneous across strata; confounding


direct adjustment

often used to adjust morbidity/mortality rates for age differences among populations

standard population provides the WEIGHTs used to adjust average rates in different study populations

the same set of weights is applied to strata in different populations


indirect adjustment

standard population provides the RATES that are applied to weights derived from strata of the study populations

used to compare populations in which true age-specific rates are unknown or their estimates are unstable

when data such as age specific data is missing


Direct adjustment population

standard populations could be used

1. you can make up your population
2. one o the study groups (smallest one you have)
3. all groups combined
4. minimum variance standard population


Standard Mortality Ratio SMR

the ratio of the number of deaths observed in the study population to the number you would expect if the study population had the same stratum specific rates as the standard population