Logistic Regression: A Self-learning Text, Third Edition (Statistics in the Health Sciences)

Presentation

I. Overview

FOCUS

Basics of matching

Model for matched

data

Control for con-

founding and inter-

action

Examples from

case-control studies

II. Basic Features of
Matching

Study design procedure:

 Select referent group

 Comparable to index group on
one or more “matching factors”

Case-control study:

"

Our focus

Referent¼controls

Index¼cases

Follow-up study:

Referent¼unexposed

Index¼exposed

This presentation describes how logistic

regression may be used to analyze matched

data. We describe the basic features of match-

ing and then focus on a general form of the

logistic model for matched data that controls

for confounding and interaction. We also pro-

vide examples of this model involving matched

case-control data.

Matching is a procedure carried out at the

design stage of a study which compares two

or more groups. To match, we select a referent

group for our study that is to be compared with

the group of primary interest, called the index

group. Matching is accomplished by constrain-

ing the referent group to be comparable to the

index group on one or more risk factors, called

“matching factors.”

For example, if the matching factor is age, then

matching on age would constrain the referent

group to have essentially the same age struc-

ture as the index group.

In a case-control study, the referent group con-

sists of the controls, which is compared with an

index group of cases.

In a follow-up study, the referent group con-

sists of unexposed subjects, which is compared

with the index group of exposed subjects.

Henceforth in this presentation, we focus on

case-control studies, but the model and meth-

ods described apply to follow-up studies also.

EXAMPLE

Matching factor¼AGE

Referent group constrained to have

same age structureas index group

392 11. Analysis of Matched Data Using Logistic Regression