A Pocket Guide to Epidemiology

279 Pages·2007·3.56 MB·English

by  David G. Kleinbaum, Kevin Sullivan, Nancy Barker

A Pocket Guide to Epidemiology
David G. Kleinbaum
Kevin M. Sullivan
Nancy D. Barker
A Pocket Guide to
Epidemiology
DavidG.Kleinbaum KevinM.Sullivan
DepartmentofEpidemiology DepartmentofEpidemiology
RollinsSchoolofPublicHealth RollinsSchoolofPublicHealth
EmoryUniversity EmoryUniversity
1518CliftonRoad,NE 1518CliftonRoad,NE
Atlanta,GA30322 Atlanta,GA30322
USA USA
[email protected] [email protected]
NancyD.Barker
2465TraywickChase
Alpharetta,GA30004
USA
[email protected]
LibraryofCongressControlNumber:2006933294
ISBN-10:0-387-45964-2 e-ISBN-10:0-387-45966-9
ISBN-13:978-0-387-45964-6 e-ISBN-13:978-0-387-45966-0
Printedonacid-freepaper.
(cid:1)C 2007SpringerScience+BusinessMedia,LLC
Allrightsreserved.Thisworkmaynotbetranslatedorcopiedinwholeorinpartwithoutthewritten
permissionofthepublisher(SpringerScience+BusinessMedia,LLC,233SpringStreet,NewYork,
NY10013,USA),exceptforbriefexcerptsinconnectionwithreviewsorscholarlyanalysis.Use
in connection with any form of information storage and retrieval, electronic adaptation, computer
software,orbysimilarordissimilarmethodologynowknownorhereafterdevelopedisforbidden.
Theuseinthispublicationoftradenames,trademarks,servicemarks,andsimilarterms,evenifthey
arenotidentifiedassuch,isnottobetakenasanexpressionofopinionastowhetherornottheyare
subjecttoproprietaryrights.
9 8 7 6 5 4 3 2 1
springer.com
Preface
Four years ago (2002), I (DGK) authored a unique educational program,ActivEpi
(Springer Publishers), developed in CD-ROM format to provide a multimedia
interactive “electronic textbook”onbasicprinciples and methodsof epidemiology.
In 2003,theActivEpi Companion Text,authoredby myself (DGK), KM Sullivan
and ND Barker and alsopublishedby Springer, was developed to provide a hard-
copyof the material contained in theActivEpi CD-ROM. The CD-ROM contains
15 chapters,with each consistingof a collectionof “activities” including narrated
expositions, interactive study questions, quizzes, homework questions, and web
links to relevant referenceson the Internet.
In the nearly three years since the publication of the ActivEpi CD-ROM, we
have received several suggestions from instructors of introductory epidemiology
courses as well as health and medical professionals to produce an abbreviated
version that narrows the discussion to the most “essential” principles andmethods.
Instructors expressed to us their concern that the material covered by the CD-
ROM (and likewise, the Companion Text)was too comprehensive to conveniently
fit the amount of time available in an introductory course.Professionals expressed
their desire for a more economically time-consuming version that would
conveniently fit their “after hours” availability.
To address these suggestions, we haveherewithproducedA Pocket Guide to
Epidemiology which provides a much shorter, more “essential” version of the
material covered by theActivEpiCD-ROM and Companion Text.We realize that
determining what is “essential” is not a simple task, especially since, from our
point of view, the original CD-ROM was already restricted to “essential” topics.
Nevertheless, to produce this text, we decided to remove fromthe originalmaterial
a great many fine points of explanation and complicated topics/issues about
epidemiologic principlesand methods, with our primarygoala“quicker read”.
A Pocket Guide to Epidemiology contains less than half asmany pages as the
ActivEpi Companion Text. We have continued to include in A Pocket Guide to
Epidemiology many of the study questions and quizzes that are provided in each
Lesson of the CD ROM, but we have eliminated homework exercises, computer
exercises, and Internet linkages from the original CD-ROM. Nevertheless, we
indicate throughout A Pocket Guide to Epidemiology how and where the
interested reader can turn to the ActivEpi CD ROM (or the Companion Text) to
pursue moredetailed information.
We authors view A Pocket Guide to Epidemiology as a stand-alone
introductory text on the basic principles and concepts of epidemiology. Our
primary audience for this textis the public health student orprofessional, clinician,
healthjournalist, and anyoneelse at any age or life experience that is interested in
learning whatepidemiology is all about in a convenient, easy to understand format
with timely, real-world health examples. We believe that the reader of this text
will also benefit from using the multi-media learner-interactive features of the
ActivEpi CD ROM electronic textbook to further clarify and enhance what is
covered in this more abbreviated (non-electronic) text. Nevertheless, we suggest
that, on its own, A Pocket Guide to Epidemiology will provide the interested
reader with a comfortable, time-efficient and enjoyable introduction to
epidemiology.
About the Authors
David G. Kleinbaum, KevinM.
Sullivan, and NancyBarker
David G. Kleinbaum is a Professor of
Epidemiology at Emory University's
Rollins School of Public Health in
Atlanta, GA, and an internationally
recognized expert in teaching
biostatistical and epidemiological
concepts andmethods at all levels.He is
the author of several widely acclaimed
textbooks including, Applied Regression
Analysis and Other Multivariable
Methods, Epidemiologic Research:
Principles and Quantitative Methods,
Logistic Regression-A Self-Learning
Text, and SurvivalAnalysis-A Self-Learning Text.
Dr. Kleinbaum has more than 25 years of experience teaching over 100 short
courses on statistical and epidemiologic methods to a variety of international
audiences, and has publishedwidely in both themethodological and applied public
health literature. He is also an experienced and sought-after consultant, and is
presently an ad-hoc consultant to all research staff at the Centers for Disease
Control and Prevention.
On a personal note, Dr. Kleinbaum is an accomplished jazz flutist, and plays
weekly in Atlanta with his jazzcombo, The Moonlighters JazzBand.
Dr. Kevin M. Sullivan is an Associate Professor of Epidemiology at Emory
University’s Rollins School of Public Health. He has worked in the area of
epidemiology and public health for over30 years and has over 80 publications in
peer-reviewedjournals andhas published chapters in several books. Dr. Sullivan
has used theActivEpi Companion Textbook and CD-ROM in a number of courses
he teaches, both in traditional classroom-based courses and distance learning
courses. He is one ofthe developers ofEpi Info,afreelydownloadable web-based
software package for the analysis of epidemiologic data published by the Centers
for Disease Control and Prevention. He is also the co-authorofOpenEpi, a freely
downloadableweb-based calculator for epidemiologicdata (www.OpenEpi.com).
Ms. Nancy Barker is a statistical consultant who formerly worked at the
Centers for Disease Control and Prevention. She is an Instructor in the Career
MPH at the Rollins School of Public Health at Emory University where she
teaches a distance learning course on Basic Epidemiology that uses ActivEpi CD
and ActivEpi Companion Text as the course textbooks. She also has been co-
instructor in several short courses on beginning and intermediate epidemiologic
methods in the Epi in Action Program sponsored by the Rollins School of Public
Health.
Contents
Preface . . . . . . v
Chapter 1 A Pocket-Size Introduction. . . 1
Chapter 2 The Big Picture – With Examples . 3
Chapter 3 How to Set Things Up? Study Designs . 21
Chapter 4 How Often does it Happen? Disease Frequency 43
Chapter 5 What’s the Answer? Measures of Effect. 71
Chapter 6 What is the Public Health Impact? . 91
Chapter 7 Is There Something Wrong? Validity & Bias 109
Chapter 8 Were Subjects Chosen Badly? Selection Bias 127
Chapter 9 Are the Data Correct? Information Bias . 139
Chapter 10 Other Factors Accounted For? Confounding
Interaction . . . . 161
Chapter 11 Confoundingcan be Counfounding– Several
RiskFactors . . . . 179
Chapter 12 SimpleAnalyses – 2x2 Tables are Not That
Simple . . . . . 191
Chapter 13 Control – What it’s All About . . 227
Chapter 14 How to Deal with Lots of Tables - Stratified
Analysis . . . . 245
Chapter 15 Matching – Seems Easy, but not That Easy. 257
Index . . . . . . 277
1
CHAPTER
A POCKET-SIZE INTRODUCTION
Epidemiology is the study of health and illness in human populations. For
example, a randomized clinical trial conducted by Epidemiologists at the Harvard
Schoolof Public Health showed that taking aspirin reducesheartattack risk by 20
to 30 percent. Public health studies in the 1950's demonstrated that smoking
cigarettes causes lung cancer. Environmental epidemiologists have been
evaluating the evidence that living near power lines may have a high risk for
childhood leukemia. Cancer researchers wonderwhyolder women are less likely
to be screened for breast cancer than younger women. All of these are examples
of epidemiologic research, because they all attempt to describe the relationship
between a health outcome and one or more explanations or causes of that outcome.
All of these examples share several challenges: they must choose an appropriate
study design, they must be careful to avoid bias, and they must use appropriate
statistical methods to analyze the data. Epidemiology deals with each of these
three challenges.
2
CHAPTER
THE BIG PICTURE - WITH EXAMPLES
The field of epidemiology was initially concerned with providinga methodological
basis for the study and control of population epidemics. Now, however,
epidemiology has a much broader scope, including the study of both acute and
chronic diseases, the quality of health care, and mental health problems. As the
focus of epidemiologic inquiry has broadened, so has the methodology. In this
overview chapter, we describe examples of epidemiologic research and introduce
several importantmethodological issuestypically considered in such research.
The Sydney Beach Users Study
Epidemiology is primarily concerned with identifying the important factors or
variables that influence a health outcome of interest. In the Sydney Beach Users
Study, the key question was “Is swimming at the beaches in Sydney associated
with an increased risk ofacute infectious illness?”
In Sydney, Australia, throughout the 1980s,complaints were expressed in the local
news media that the popular public beaches surrounding the city were becoming
more and more unsafe for swimming. Much of the concern focused on the
suspicion that thebeacheswere being increasingly pollutedby waste disposal.
In 1989, the New South Wales Department of Health decided to undertake a
study to investigate the extent to which swimming and possible pollution at 12
popular Sydney beaches affected the public's health, particularly during the s
Summer months when the
beaches weremost crowded. The
primary research question of
interest was: are persons who
swim at Sydney beaches at
increased risk for developing an
acute infectious illness?
The study was carried out by
selecting subjects on the beaches
throughout the summer months
of 1989-90. Those subjects
eligible to participate at this initial interview were then followed-up by phone a
week later to determine swimming exposure on the dayof thebeach interview and
subsequent illness statusduring the weekfollowing the interview.
Water quality measurements at the beaches were also taken on each day that
subjectsweresampled inorder to match swimming exposure topollutionlevels at
the beaches.
Analysis of the study data lead to the overall conclusion that swimming in
polluted water carried a statistically significant 33% increased risk for an
infectious illness when compared to swimming in non-polluted water. These
4 Chapter 2. The Big Picture – With Examples
results were considered by health department officials and the public alike to
confirm that swimming in Sydney beaches posed an important health problem.
Consequently, the state and local health departments together with other
environmentalagencies in the Sydney areaundertook a program to reduce sources
of pollution of beach water that lead to improved water quality at the beaches
during the1990’s.
Summary
(cid:153) The Sydney Beach Users Study is an example of the application of
epidemiologic principles and methods to investigate a localized public health
issue.
(cid:153) Thekeyquestion in the Sydney Beach Users Study was:
o Does swimming at the beaches in Sydney, Australia (in 1989-90) pose an
increased health risk foracuteinfectiousillnesses?
o The conclusion was yes, a 33% increased risk.
Important Methodological Issues
We provide a general perspective of epidemiologic research by highlighting
several broad issues that arise during the course of most epidemiologic
investigations.
There are many issues to worry about when planning an epidemiologic research
study (see Box below). In this chapter we will begin to describe a list of broad
methodological issues that need to be addressed. We will illustrate each issue
using thepreviouslydescribed Sydney BeachUsers Studyof 1989.
Issuesto consider when planningan epidemiologic research study
Question Definea question of interest and key variables
Variables What to measure: exposure (E), disease (D), and control (C) variables
Design What study design and sampling frame?
Frequency Measures of disease frequency
Effect Measures of effect
Bias Flaws in study design,collection, or analysis
Analysis Perform appropriate analyses
The first two issues require clearly defining the study question of interest,
followed by specifying the key variables to be measured. Typically, we first
should ask: What is the relationship of one or
more hypothesized determinants to a disease
or health outcome ofinterest?
Adeterminantis often called anexposure variable and is denoted by the letter
E. The diseaseorhealth outcome is often denoted as D. Generally, variables other
than exposure and disease that are known to predict the health outcome must be
taken into account. We often call these variables control variables and denote
them using theletterC.
Next, we must determine how to actually measure these variables. This step
requires determining the information-gathering instruments and survey
questionnairesto be obtainedor developed.