Gordis Epidemiology 6th Edition PDF

Any screen. Any time. Anywhere. Activate the eBook version of this title at no additional charge. Student Consult eBooks give you the power to browse and find content, view enhanced images, share notes and highlights—both online and offline. Unlock your eBook today. 1 Visit studentconsult.inkli...

Any screen. Any time. Anywhere. Activate the eBook version of this title at no additional charge. Student Consult eBooks give you the power to browse and find content, view enhanced images, share notes and highlights—both online and offline. Unlock your eBook today. 1 Visit studentconsult.inkling.com/redeem Scan this QR code to redeem your eBook through your mobile device: 2 Scratch off your code 3 Type code into “Enter Code” box 4 Click “Redeem” 5 Log in or Sign up 6 Go to “My Library” Place Peel Off It’s that easy! Sticker Here For technical assistance: email [email protected] call 1-800-401-9962 (inside the US) call +1-314-447-8200 (outside the US) Use of the current edition of the electronic version of this book (eBook) is subject to the terms of the nontransferable, limited license granted on studentconsult.inkling.com. Access to the eBook is limited to the first individual who redeems the PIN, located on the inside cover of this book, at studentconsult.inkling.com and may not be transferred to another party by resale, lending, or other means. 2015v1.0 Gordis Epidemiology NOTES TO INSTRUCTORS: Contact your Elsevier Sales Representative for teaching resources, including an image bank, for Gordis Epidemiology, 6e, or request these supporting materials at: http://evolve.elsevier.com/Gordis/epidemiology/  6th Edition Gordis Epidemiology David D. Celentano, ScD, MHS Dr. Charles Armstrong Chair and Professor Department of Epidemiology Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland Moyses Szklo, MD, MPH, DrPH University Distinguished Professor Department of Epidemiology Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland 1600 John F. Kennedy Blvd. Ste 1600 Philadelphia, PA 19103-2899 GORDIS EPIDEMIOLOGY, SIXTH EDITION ISBN: 978-0-323-55229-5 Copyright © 2019 by Elsevier, Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. With respect to any drug or pharmaceutical products identified, readers are advised to check the most current information provided (i) on procedures featured or (ii) by the manufacturer of each product to be administered, to verify the recommended dose or formula, the method and duration of administration, and contraindications. It is the responsibility of practitioners, relying on their own experience and knowledge of their patients, to make diagnoses, to determine dosages and the best treatment for each individual patient, and to take all appropriate safety precautions. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Previous editions copyrighted 2014, 2009, 2004, 2000, 1996 by Saunders, an imprint of Elsevier Inc. Library of Congress Cataloging-in-Publication Control Number: 2018949544 Publisher: Elyse O’Grady Senior Content Development Specialist: Deidre Simpson Publishing Services Manager: Catherine Jackson Book Production Specialist: Kristine Feeherty Design Direction: Ryan Cook Cover credit: An original watercolor “Remembering Baltimore” by Haroutune K. Armenian, MD, DrPH Professor Emeritus Department of Epidemiology Johns Hopkins Bloomberg School of Public Health Printed in Canada Last digit is the print number: 9 8 7 6 5 4 3 2 1  In Memoriam LEON GORDIS, MD, MPH, DrPH 1934–2015 Preface Epidemiology is one of the foundational disciplines of the previous edition. In the fifth edition, learning underlying public health. Clinical research relies heavily objectives were inserted in most chapters, and we have on epidemiologic methods and contemporary medical revised these and updated the examples throughout. care research, particularly in comparative effectiveness Additional new review questions have been added to studies and statistical approaches to “big data” (as in most chapters. A significant change has been to the the use of the electronic medical record for health presentation and order of the methods in epidemiology studies). As Dr. Leon Gordis wrote in his preface to that were previously presented at the end of Section I the fifth edition, “Epidemiology is the basic science of and more extensively in Section II. Rather than leading disease prevention and plays major roles in developing with the randomized trial (or the “experimental” design) and evaluating public policy relating to health and to and then comparing observational study design to the social and legal issues.” There are many uses of epi- gold standard, we have organized the presentation of demiology today. The majority of epidemiologic research epidemiologic methods along a study continuum from focuses on establishing etiologic associations between clinical observation, to case-series, to the use of ecologic putative risks and health outcomes. However, epidemi- studies, and then to cross-sectional investigations as ology is also widely used in the evaluation of primary the foundational approach to epidemiologic hypothesis and secondary prevention programs, comparisons of development. We then follow with case-control and interventions, and the evaluation of policy at the cohort designs, leading up to the randomized trial. population level. Epidemiologic findings commonly This more organically follows the development, in find their way into public media, providing the public our opinion, of how epidemiologic observations and and policy makers with data to guide personal decisions hypotheses are developed in the daily practice of doing regarding their behavior. Increasingly, the scrutiny epidemiology. focused on epidemiology may cause researchers and As with the previous edition, the sixth edition practitioners some discomfort, as the interpretation of consists of three sections. Section I addresses how basic epidemiologic principles can be subject to con- epidemiology is used to understand health and the siderable error. Our task is to make the thinking development of diseases in populations as well as the underlying epidemiology transparent. basis for interventions to influence the natural history This book is intended to be a basic introduction to of disease. The first six chapters provide the conceptual the definitions, logic, and use of the epidemiologic framework underlying the discipline of epidemiology method to elucidate factors influencing health and and present many of the basic principles of the disci- disease. We have tried to illustrate the principles with pline. Chapter 1 provides an overview of epidemiology, examples of how epidemiology is applied in the real using many historical examples to illustrate how the world. The examples selected include both “classic field developed. Chapter 2 is concerned with how examples” from the early days of the development of disease is transmitted in populations, both directly (in the discipline of epidemiology to contemporary exam- the case of infectious pathogens) and indirectly (for ples. Where appropriate, we draw on examples pertain- example, through a vector such as a mosquito or ing to clinical practice. contaminated air). The basic terms used in epidemics Upon the passing of Dr. Gordis in 2015, the sixth are presented and illustrated to guide the student in edition of this book has been revised by two new seeing how these principles and terms are used. Chapter authors, both of whom worked with and under Professor 3 addresses disease surveillance and how we measure Gordis and have been actively engaged in teaching morbidity in populations, while Chapter 4 is concerned epidemiology at Johns Hopkins for over four decades. with aspects of mortality and measures of disease impact We have generally retained the structure and organization in populations. Chapter 5 focuses on ways to detect vi  Preface vii disease in populations, comparing different approaches concept of risk is expanded to include the calculation to differentiate people who have a disease from those and interpretation of the attributable risk, the population who are disease free, articulating how screening tests attributable risks, and their use in evaluating the success can be adjusted to better diagnose those with or those of prevention programs. Causal inference is introduced without the disease in question. The issues of the in Chapter 14 and focuses on how to derive inferences reliability and validity of screening tests are of critical in epidemiologic investigations. Chapter 15 presents interest to both clinicians and to those planning for issues of bias, confounding, and interaction in epide- health services. Finally, Chapter 6 presents how the miologic studies and discusses how they influence causal natural history of disease can be used to best express inference. Finally, Chapter 16 addresses the role of disease prognosis, using examples of case-fatality and genetic and environmental contributions to the etiology survivorship. of disease, and presents new methods of genetic research Section II details the methods used by epidemiolo- commonly used in epidemiologic studies today. gists primarily to ascribe associations between a Section III addresses the uses of epidemiology in hypothesized exposure (risk) and a health outcome. everyday public health. The final four chapters address Chapter 7 discusses the initial observations made in some of the critical issues facing the field today. Chapter clinical practice (the case report) leading to a recognition 17 illustrates how epidemiologic principles and designs of an accumulation of cases that appear to have some described in Sections I and II are used in the evaluation commonalities (the case series). This is followed by an of health services. Chapter 18 addresses the use of introduction to the ecologic design and its analysis, epidemiology to evaluate screening programs, while with cautions as to its interpretation. Finally, cross- Chapter 19 details how epidemiology can be used to sectional (snapshot) studies are presented as the address major areas of public health policy. The final groundwork for hypothesis development. Chapter 8 chapter summarizes ethical issues confronted in the then provides an introduction to observational studies practice of epidemiology and reviews some of the as commonly used in epidemiology, addressing case- important professional issues confronted by the field control and cohort studies, which are then compared today. in Chapter 9. To this point, we are addressing exposures We have continued in Professor Gordis’ use of as they occur in populations, where we are observers illustrations and examples to demonstrate how epide- of exposures and their putative impacts on health miologic issues and principles are put into practice. outcomes. In Chapter 10 we then move to the “experi- We have updated examples extensively and added new mental” approach (randomized trial) in which the examples throughout the text. Many of the prior investigator “assigns” exposure or health interven- chapters have been extensively edited and updated, tion—generally randomly—to study participants to with some chapters being greatly expanded. The genetic address how this influences the health outcome. In epidemiology presentation has been heavily edited and this case the exposure is under the control of the updated, reflecting the amazing growth in genomics investigator, not the study participant, a crucial differ- research in the past 5 years. In Chapter 16 we have ence in the randomized trial as compared to the cohort added a glossary of genetic epidemiology terms to or other observational study design. Chapter 11 dis- provide the student with some guidance for this cusses a series of issues involved in the conduct of somewhat complex field. Finally, new review questions randomized trials, including sample size, power, and have been added at the end of most chapters. generalizability; determining efficacy (vs. effectiveness); Our aim for this book is to allow the reader to ethical considerations; and the US Federal Drug appreciate how epidemiology can be used to respond Administration phases for evaluating new drugs. In to population health problems confronting society today. Chapter 12 we present issues on estimating risk, includ- Our expectation is not that the reader will be able to ing absolute and relative risk and their interpretation, conduct an epidemiologic investigation. Rather, we calculating and interpreting an odds ratio in a case- hope that there will be an appreciation of what epi- control study and in a cohort study, and doing so in demiology is, what the basic research and evaluation a matched-pairs case-control study. In Chapter 13 the designs are, and how to interpret the basic findings in viii Preface an epidemiologic study. We hope that the excitement are quintessential Baltimore, much as the “Painted we feel about the uses of epidemiology will come across Ladies” are identified with San Francisco. Much of Dr. to the reader of this text. Gordis’ research centered on pediatric and childhood The cover illustration selected for this edition of disease in Baltimore, as illustrated in many of the Gordis Epidemiology has special meaning. This original examples in this text. We are particularly proud to watercolor by Haroutune Armenian, MD, DrPH, created include this tribute by Dr. Armenian to Dr. Gordis and in August 2017, is titled “Remembering Baltimore.” to our first revision of his world-renowned text. This Professor Armenian was a professor of epidemiology, sixth edition has kept our mind on our friend and deputy chair to Professor Gordis for many years, and mentor. interim chair from 1993–94, until Jon Samet became chair. “Remembering Baltimore” truly captures the urban David D. Celentano landscape of Baltimore, Dr. Gordis’ adopted home for Moyses Szklo some 60 years. The distinctive rowhomes on the harbor August 2018 Acknowledgments This book reflects the contributions of several genera- Gange, Shruti Mehta, and Alvaro Munoz. To past co- tions of teachers of epidemiology at Johns Hopkins, instructors of the introductory course, we acknowledge first as the School of Hygiene and Public Health, and Bill Moss, Elizabeth Platz, and Jennifer Deal for their more recently as the Bloomberg School of Public dedication to educating scores of public health students Health. The course was developed by the Department of in the “art” of epidemiology. In particular, Dr. Deal Epidemiology faculty and was first taught as Principles has made outstanding contributions to our introduc- of Epidemiology by Dr. Abraham Lilienfeld, the chair tory course, and many of the examples introduced of the department from 1970–75. Dr. Leon Gordis in this edition come from her suggestions, for which became the course instructor following an acute illness we are particularly appreciative. The support of many experienced by Dr. Lilienfeld in the midst of teaching deans of the school is also appreciated, including D.A. the subject in 1974. Dr. Gordis then was the primary Henderson, Al Sommer, Mike Klag, and most recently lecturer for the following 30 years. In addition, Dr. Ellen MacKenzie. The course on which this book is Gordis taught epidemiology to many cohorts of School based would not exist without the long-term dedication of Medicine students for a similar period of time. This and knowledge of our colleague Allyn Arnold who has book was developed from these experiences, and Dr. served as the bridge from the Gordis years to the present. Gordis was the author of the first five editions of this Preparing the sixth edition of this book was a sig- very popular text. nificant undertaking for us. Our goal was to preserve The current authors were trained in public health Dr. Gordis’ voice—and humor—and to retain the style at Johns Hopkins and were actively engaged as members of the text as much as possible. We also sought to of the epidemiology teaching team for many years when update examples and to intersperse new illustrations they were junior faculty. Dr. Szklo taught the second of the epidemiologic principles we are presenting along course in the epidemiology sequence, Intermediate with time-honored classics that were included in earlier Epidemiology. Upon Dr. Gordis’ retirement, Dr. Celen- editions. tano became the director of Principles of Epidemiology, Youssef Farag, MBBCh, MPH, PhD, was invaluable which has recently been revised in content and renamed in preparing the sixth edition. He is a bright, talented, Epidemiologic Inference in Public Health 1. Its content and hardworking young physician-epidemiologist whom reflects this sixth edition of Gordis Epidemiology. we recruited to help us in this extraordinary endeavor. Many colleagues have made invaluable contributions While completing his PhD in epidemiology at Johns to this revision of Gordis Epidemiology. Chief among them Hopkins Bloomberg School of Public Health, Youssef was the late Dr. George W. Comstock, mentor, adviser, took on the minutia of preparing this text—from and sage scientist to both of us. We also acknowledge updating CDC figures on morbidity and mortality, to the assistance of many past and current colleagues, working closely with the National Cancer Institute to including Haroutune Armenian, Dr. Gordis’ deputy chair run new data analyses to illustrate key epidemiologic and acting chair when Dr. Gordis stepped down as points, and to finding references vaguely suggested by department chair, who contributed the original cover us. He led the significant reorganization of the chapters art, “Remembering Baltimore,” for this book. We also in the sixth edition, including rewriting entire new acknowledge our former chair, Jonathan Samet, and sections from scratch in several chapters. He also took Michel Ibrahim, who joined us as professor following on the initiative to update outdated examples from his his 2002 retirement as dean at the University of North knowledge of current medical and public health issues, Carolina–Chapel Hill. Others who have had major and his firm grasp of the relevant literature. His creative impacts on the teaching program in the department contributions facilitated simplifying and clarifying include Javier Nieto, Rosa Crum, Paul Whelton, Stephen conventionally challenging concepts in epidemiology. ix x Acknowledgments During a period of over one year, from our in-depth better book than would have occurred without her discussions during weekly meetings, numerous emails careful review. in between, and multiple revisions for each chapter, Preparing the sixth edition of Gordis Epidemiology has this project would never have run so smoothly without brought us many memories of Leon and his legacy at his commitment and calm and determined nature, for Johns Hopkins. The department has certainly changed which we are very grateful. We firmly believe that he since he stepped down as chair in 1993. Today we will be a future leader in epidemiology. are a significantly larger faculty, covering many more The chapter on the role of genetics in contemporary areas of epidemiology in greater depth, and using epidemiology was heavily influenced by our genetic tools unimaginable even a decade ago. At the same epidemiology colleagues Priya Duggal and Terri Beaty. time, the discipline remains grounded in the ideas This field has been changing so rapidly—and is tech- first set forth by Wade Hampton Frost at the dawn of nologically complicated to the naïve—that they assisted our school in 1919. This book is a testament to the us in doing a major revision in this sixth edition. We thought-leaders and giants of epidemiology who have cannot thank them enough for their contributions to studied and taught epidemiology at Johns Hopkins over this chapter. the past 100 years and hopefully will guide us into Charlotte Gerczak was invaluable in copy-editing our second century of practice, education, research, this volume. Charlotte worked for many years with and service. Jonathan Samet and is very experienced in working with practicing epidemiologists. Her gifted eye for grammar, David D. Celentano sentence structure, and meaning has made this a far Moyses Szklo Contents Section I 12 Estimating Risk: Is There an Association? 240 THE EPIDEMIOLOGIC APPROACH TO DISEASE AND INTERVENTION, 1 13 More on Risk: Estimating the Potential for Prevention, 259 1 Introduction, 2 14 From Association to Causation: Deriving Inferences 2 The Dynamics of Disease Transmission, 20 From Epidemiologic Studies, 269 3 The Occurrence of Disease: I. Disease Surveillance 15 More on Causal Inference: Bias, Confounding, and and Measures of Morbidity, 41 Interaction, 289 4 The Occurrence of Disease: II. Mortality and Other 16 Identifying the Roles of Genetic and Environmental Measures of Disease Impact, 65 Factors in Disease Causation, 307 5 Assessing the Validity and Reliability of Diagnostic Section III and Screening Tests, 94 APPLYING EPIDEMIOLOGY TO EVALUATION AND POLICY, 332 6 The Natural History of Disease: Ways of Expressing Prognosis, 123 17 Using Epidemiology to Evaluate Health Services, 333 Section II USING EPIDEMIOLOGY TO IDENTIFY THE CAUSE 18 Epidemiologic Approach to Evaluating Screening OF DISEASE, 147 Programs, 353 7 Observational Studies, 149 19 Epidemiology and Public Policy, 377 8 Cohort Studies, 178 20 Ethical and Professional Issues in Epidemiology, 395 9 Comparing Cohort and Case-Control Studies, 193 Answers to Review Questions, 409 10 Assessing Preventive and Therapeutic Measures: Randomized Trials, 197 Index, 411 11 Randomized Trials: Some Further Issues, 216 xi This page intentionally left blank  Section I THE EPIDEMIOLOGIC APPROACH TO DISEASE AND INTERVENTION T his section begins with an overview of the objec- and approaches for using mortality data in investigations tives of epidemiology, some of the approaches used relating to public health and clinical practice. Other in epidemiology, and examples of the applications issues relating to the impact of disease, including quality of epidemiology to human health problems (Chapter of life and projecting the future burden of disease, are 1). It then discusses how diseases are transmitted also discussed in Chapter 4. (Chapter 2). Diseases do not arise in a vacuum; they Armed with knowledge of how to describe morbidity result from an interaction of human beings with and mortality in quantitative terms, we then turn to their environment, including other people. An the question of how to assess the quality of diagnostic understanding of the concepts and mechanisms and screening tests that are used to determine which underlying the transmission and acquisition of people in the population have a certain disease (Chapter disease is critical to exploring the epidemiology of 5). After we identify people with the disease, we need human disease and to preventing and controlling ways to describe the natural history of disease in many infectious diseases. quantitative terms; this is essential for assessing the To discuss the epidemiologic concepts presented severity of an illness and for evaluating the possible in this book, we need to develop a common language, effects on survival of new therapeutic and preventive particularly for describing and comparing morbidity interventions (Chapter 6). and mortality. Chapter 3 therefore discusses morbid- This first section, then, introduces the student to ity and the important role of epidemiology in disease the nomenclature of epidemiology, surveillance and its surveillance. The chapter then presents how measures ramifications for determining the health of populations, of morbidity are used in both clinical medicine and and then focuses on screening and the natural history public health. Chapter 4 presents the methodology of disease. 1  Chapter 1 Introduction I hate definitions. ultimate aim is to intervene to reduce morbidity —Benjamin Disraeli (1804–1881, British Prime Minister 1868 and and mortality from the disease. We want to develop 1874–1880) a rational basis for prevention programs. If we can identify the etiologic or causal factors for disease and reduce or eliminate exposure to those factors, What Is Epidemiology? we can develop a basis for prevention programs. In Epidemiology is the study of how disease is distributed addition, we can develop appropriate vaccines and in populations and the factors that influence or deter- treatments, which can prevent the transmission of mine this distribution. Why does a disease develop in the disease to others. some people and not in others? The premise underlying The second objective of epidemiology is to epidemiology is that disease, illness, ill health, and determine the extent of disease found in the com- excellent health status are not randomly distributed in munity. What is the burden of disease in the com- human populations. Rather, each of us has certain munity? This question is critical for planning health characteristics that predispose us to, or protect us services and facilities and for estimating how many against, a variety of different diseases. These charac- future health care providers should be trained. teristics may be primarily genetic in origin, the result A third objective is to study the natural history of exposure to certain environmental hazards, or the and prognosis of disease. Clearly, certain diseases behaviors (good and bad) that we engage in. Perhaps are more severe than others; some may be rapidly most often, we are dealing with an interaction of genetic, lethal, whereas others may have extended durations environmental, and behavioral and social factors in the of survival. Many diseases are not fatal but may development of disease. affect quality of life or be associated with disability. A broader definition of epidemiology than that given We want to define the baseline natural history of a previously has been widely accepted. It defines epi- disease in quantitative terms so that as we develop demiology as “the study of the distribution and new modes of intervention, either through treatments determinants of health-related states or events in or through new ways of preventing complications, we specified populations and the application of this study can compare the results of using these new modalities to control of health problems.”1 What is noteworthy with the baseline data to determine whether our about this definition is that it includes both a description new approaches have truly been effective. of the content of the discipline and why epidemiologic Fourth, we use epidemiology to evaluate both investigations are carried out. existing and newly developed preventive and therapeutic measures and modes of health care delivery. For example, does screening men for prostate cancer Objectives of Epidemiology using the prostate-specific antigen (PSA) test improve What are the specific objectives of epidemiology? First, survival in people found to have prostate cancer? to identify the etiology, or cause, of a disease and its Has the growth of managed care and other new relevant risk factors (i.e., factors that increase a person’s systems of health care delivery and health care risk for a disease). We want to know how the disease insurance had an impact on the health outcomes is transmitted from one person to another or from a of the patients involved and on their quality of life? nonhuman reservoir to a human population or why it If so, what has been the nature of this impact and arises due to risk behaviors the person engages in. Our how can it be measured? 2  1 Introduction 3 Finally, epidemiology can provide the foundation health problems in the early 19th century. Clearly, for developing public policy relating to environmental cholera is no longer a major problem in the United problems, genetic issues, and other social and behavioral States today, but in many low-income and war-torn considerations regarding disease prevention and health countries of the world it remains a serious threat, with promotion. For example, is the electromagnetic radiation many countries periodically reporting outbreaks of that is emitted by cell phones, electric blankets and cholera that are characterized by high death rates, often heating pads, and other household appliances a hazard as a result of inadequate or inaccessible medical care. to human health? Are high levels of atmospheric ozone Let us compare the major causes of death in the or particulate matter a cause of adverse acute or chronic United States in 1900 and 2014 (Fig. 1.2). The catego- health effects in human populations? Is radon in homes ries of causes have been color coded as described in a significant risk to human beings? Which occupations the caption for this figure. In 1900 the leading causes are associated with increased risks of disease in workers, of death were pneumonia and influenza, followed by and what types of regulation are required to reduce tuberculosis and diarrhea and enteritis. In 2014 the these risks? leading causes of death were heart disease, cancer, chronic lower respiratory diseases, and unintentional CHANGING PATTERNS OF COMMUNITY injuries. What change has occurred? During the 20th HEALTH PROBLEMS century there was a dramatic shift in the causes of A major role of epidemiology is to provide clues to death in the United States. In 1900 the three leading changes that take place over time in the health problems causes of death were infectious diseases; however, now presenting in the community. Fig. 1.1 shows a sign in we are dealing with chronic diseases that in most situ- a cemetery in Dudley, England, in 1839. At that time, ations are not communicable or infectious in origin. cholera was the major cause of death in England; the Consequently, the kinds of research, intervention, and churchyard was so full that no burials of persons who services we need today differ from those that were died of cholera would henceforth be permitted. The needed in the United States in 1900. sign conveys an idea of the importance of cholera in The pattern of disease occurrence currently seen in the public’s consciousness and in the spectrum of public developing countries is often similar to that which was Fig. 1.1 Sign in cemetery in Dudley, England, in 1839. (From the Dudley Public Library, Dudley, England.) 4 SECTION I The Epidemiologic Approach to Disease and Intervention 1900 2014 Pneumonia and influenza Heart disease Tuberculosis Cancer Diarrhea and enteritis Chronic lower respiratory disease Heart disease Accidents (unintentional injuries) Fig. 1.2 Ten leading causes of death in the United Stroke States, 1900 and 2014. Although the definitions of the Stroke diseases in this figure are not exactly comparable in Kidney disease Alzheimer’s disease 1900 and 2014, the bars in the graphs are color coded Accidents Diabetes mellitus to show chronic diseases (pink), infectious diseases Cancer Influenza and pneumonia (purple), injuries (aqua), and diseases of aging (white). (Redrawn from Grove RD, Hetzel AM. Vital Statistics Rates of Senility Kidney disease the United States, 1940–1960. Washington, DC: US Govern- Diphtheria Suicide ment Printing Office; 1968; and Kochanek KD, Murphy SL, 0 50 100 150 200 250 0 50 100 150 200 250 Xu JQ, Tejada-Vera B. Deaths: Final data for 2014. Natl Vital Stat Rep. 2016;65(4):1–122. [Hyattsville, MD: National Center Death rates per 100,000 Death rates per 100,000 for Health Statistics.]) TABLE 1.1 Ten Leading Causes of Death and Their Percentages of All Deaths, United States, 2014 Rank Cause of Death No. of Deaths Percent (%) of Total Deaths Age-Adjusted Death Ratea All Causes 2,626,418 100.0 724.6 1 Diseases of the heart 614,348 23.4 167.0 2 Malignant neoplasms (cancer) 591,699 22.5 161.2 3 Chronic lower respiratory diseases 147,101 5.6 40.5 4 Accidents (unintentional injuries) 136,053 5.2 40.5 5 Cerebrovascular diseases 133,103 5.1 36.5 6 Alzheimer disease 93,541 3.6 25.4 7 Diabetes mellitus 76,488 2.9 20.9 8 Influenza and pneumonia 55,227 2.1 15.1 9 Nephritis, nephrotic syndrome, 48,146 1.8 13.2 and nephrosis 10 Intentional self-harm (suicide) 42,773 1.6 13.0 All other causes 687,939 26.2 a Rates are per 100,000 population and age-adjusted for the 2010 US standard population. Note: Percentages may not total 100 due to rounding. Data from Centers for Disease Control and Prevention, Xu JQ, Murphy SL, Kochanek KD, Arias E. Mortality in the United States, 2015. NCHS data brief, no 267. Hyattsville, MD: National Center for Health Statistics; 2016. https://www.cdc.gov/ nchs/data/databriefs/db267_table.pdf. Accessed April 17, 2017. seen in the United States in 1900: infectious diseases immunodeficiency virus (HIV) infection has emerged remain the leading causes of death. However, as and the incidence of tuberculosis has increased, infec- countries become industrialized they increasingly tious diseases are again becoming major public health manifest the mortality patterns currently seen in problems. Table 1.1 shows the 10 leading causes of developed countries, with mortality from chronic death in the United States in 2014. The three leading diseases becoming the major challenge (this is com- causes—heart disease, cancer, and chronic lower respira- monly referred to as the “epidemiologic transition”). tory diseases—account for almost 55% of all deaths, However, even in industrialized countries, as human an observation that suggests specific targets for  1 Introduction 5 prevention if a significant reduction in mortality is to disease. Why should we identify such high-risk groups? be achieved. First, if we can identify these high-risk groups, we can Another demonstration of changes that have taken direct preventive efforts, such as screening programs place over time is seen in Fig. 1.3, which shows the for early disease detection, to populations who may remaining years of expected life in the United States not have been screened before and are most likely to at birth and at age 65 years for the years 1900, 1950, benefit from any interventions that are developed for and 2014, by race and sex. the disease. In sub-Saharan Africa, targeted HIV The number of years of life remaining after birth counseling and testing to men who are not aware of has dramatically increased in all of these groups, with their status can effectively reduce epidemics if they are most of the improvement having occurred from 1900 linked to care, started on antiretroviral therapy, and to 1950 and much less having occurred since 1950. If continued in care. we look at the remaining years of life at age 65 years, Second, if we can identify such groups, we may be very little improvement is seen from 1900 to 2014. able to identify the specific factors or characteristics What primarily accounts for the increase in remaining that put them at high risk and then try to modify those years of life at birth are the decreases in infant mortality factors. It is important to keep in mind that such risk and in mortality from childhood diseases. In terms of factors may be of two types. Characteristics such as diseases that afflict adults, especially those 65 years and age, sex, and race, for example, are not modifiable, older, we have been much less successful in extending although they may permit us to identify high-risk the span of life, and this remains a major challenge. groups. On the other hand, characteristics such as obesity, smoking, diet, sexual practices, and other lifestyle factors may be potentially modifiable and may Epidemiology and Prevention thus provide an opportunity to develop and introduce A major use of epidemiologic evidence is to identify new prevention programs aimed at reducing or changing subgroups in the population who are at high risk for specific exposures or risk factors. 90 80 70 1900 60 1950 2014 50 40 Years 30 20 10 0 White White Black Black White White Black Black males females males females males females males females At birth At age 65 years Fig. 1.3 Life expectancy at birth and at 65 years of age, by race and sex, United States, 1900, 1950, and 2014. (Redrawn from National Center for Health Statistics. Health, United States, 1987 DHHS Publication No. 88–1232. Washington, DC: Public Health Service; March 1988; National Center for Health Statistics. Health, United States, 2015: with special feature on racial and ethnic health disparities; 2016. https://www.cdc.gov/nchs/hus/contents2015. htm#015. Accessed May 2, 2017.) 6 SECTION I The Epidemiologic Approach to Disease and Intervention medical advice. Our objective with secondary prevention PRIMARY, SECONDARY, is to detect the disease earlier than it would have been AND TERTIARY PREVENTION detected with usual care. By detecting the disease at In discussing prevention, it is helpful to distinguish an early stage in its natural history, often through among primary, secondary, and tertiary prevention screening, it is hoped that treatment will be easier and/ (Table 1.2). or more effective. For example, most cases of breast Primary prevention denotes an action taken to prevent cancer in older women can be detected through mam- the development of a disease in a person who is well mography. Several recent studies indicate that routine and does not (yet) have the disease in question. For testing of the stool for occult blood can detect treatable example, we can immunize a person against certain colon cancer early in its natural history but colonoscopy diseases so that the disease never develops or, if a is a better test, although far more expensive and invasive. disease is environmentally induced, we can prevent a The rationale for secondary prevention is that if we person’s exposure to the environmental factor involved can identify disease earlier in its natural history than and thereby prevent the development of the disease. would ordinarily occur, intervention measures may be Primary prevention is our ultimate goal. For example, more effective and life prolonged. Perhaps we can we know that most lung cancers are preventable. If we prevent mortality or complications of the disease and can help to stop people from ever smoking, we can use less invasive or less costly treatment to do so. eliminate 80% to 90% of lung cancer in human beings. Evaluating screening for disease and the place of such However, although our aim is to prevent diseases from intervention in the framework of disease prevention occurring in human populations, for many diseases, are discussed in Chapter 18. such as prostate cancer and Alzheimer disease, we do Tertiary prevention denotes preventing complications not yet have the biologic, clinical, or epidemiologic data in those who have already developed signs and symp- on which to base effective primary prevention programs. toms of an illness and have been diagnosed (i.e., people Secondary prevention involves identifying people in who are in the clinical phase of their illness). This is whom a disease process has already begun but who generally achieved through prompt and appropriate have not yet developed clinical signs and symptoms treatment of the illness combined with ancillary of the illness. This period in the natural history of a approaches such as physical therapy that are designed disease is called the preclinical phase of the illness and to prevent complications such as joint contractures. is discussed in Chapter 18. Once a person develops clinical signs or symptoms it is generally assumed that TWO APPROACHES TO PREVENTION: under ideal conditions the person will seek and obtain A DIFFERENT VIEW Two possible approaches to prevention are a population- based approach and a high-risk approach.2 In the population-based approach, a preventive measure is TABLE 1.2 Three Types of Prevention widely applied to an entire population. For example, Type of prudent dietary advice for preventing coronary disease Prevention Definition Examples or advice against smoking may be provided to an entire Primary Preventing the initial Immunization, population using mass media and other health education development of a reducing approaches. An alternate approach is to target a high-risk disease exposure to group with the preventive measure. Thus screening for a risk factor cholesterol in children might be restricted to children Secondary Early detection of Screening for existing disease cancer from high-risk families. Clearly, a measure applied to to reduce severity an entire population must be relatively inexpensive and complications and noninvasive. A measure that is to be applied to a Tertiary Reducing the Rehabilitation high-risk subgroup of the population may be more impact of the for stroke expensive and may be more invasive or inconvenient disease but also has to be able to correctly identify individuals  1 Introduction 7 with the disease. More on screening tests is discussed in Chapter 18. Population-based approaches can be considered public health approaches, whereas high-risk approaches more often require a clinical action to identify the high-risk group to be targeted. In most situations, a combination of both approaches is ideal. Often a high-risk approach, such as prevention coun- seling, is limited to brief encounters with physicians. These approaches are discussed further in Chapter 19. Epidemiology and Clinical Practice Epidemiology is critical not only to public health but also to clinical practice. The practice of medicine is Fig. 1.4 “You’ve got whatever it is that’s going around.” (Al Ross/The dependent on population data. For example, if a physi- New Yorker Collection/The Cartoon Bank.) cian hears an apical systolic murmur, a heart sound produced when blood flows across the heart valves, how does he or she know whether it represents mitral concepts. What is portrayed humorously here is a true regurgitation? Where did this knowledge originate? The commentary on one aspect of pediatric practice—a diagnosis is based on correlation of the clinical findings pediatrician often makes a diagnosis based on what (such as the auscultatory findings—sounds heard using the parent tells him or her over the telephone and on a stethoscope) with the findings of surgical pathology what he or she knows about which illnesses, such as or autopsy and with the results of echocardiography, viral and bacterial infections, are “going around” in the magnetic resonance, or catheterization studies in a large community. Thus the data available about illness in group of patients. Thus the process of diagnosis is the community can be very helpful in suggesting a population based (see Chapter 5). The same holds for diagnosis, even if they are not conclusive. Data regarding prognosis. For example, a patient asks his physician, the etiology of sore throats according to a child’s age “How long do I have to live, doctor?” and the doctor are particularly relevant (Fig. 1.5). If the infection occurs replies, “Six months to a year.” On what basis does early in life, it is likely to be viral in origin. If it occurs the physician prognosticate? He or she does so on at ages 4 to 7 years, it is likely to be streptococcal in the basis of experience with large groups of patients origin. In an older child, Mycoplasma becomes more who have had the same disease, were observed at the typical. Although these data do not make the diagnosis, same stage of disease, and received the same treatment. they do provide the physician or other health care Again, prognostication is based on population data (see provider with a good clue as to what agent or agents Chapter 6). Finally, selection of appropriate therapy to suspect. is also population based. Randomized clinical trials that study the effects of a treatment in large groups of patients are the ideal means (the so-called gold Epidemiologic Approach standard) for identifying appropriate therapy (see How does the epidemiologist proceed to identify the Chapters 10 and 11). Thus population-based concepts cause of a disease? Epidemiologic reasoning is a mul- and data underlie the critical processes of clinical tistep process. The first step is to determine whether practice, including diagnosis, prognostication, and an association exists between exposure to a factor (e.g., selection of therapy. In effect, the physician applies a an environmental agent) or a characteristic of a person population-based probability model to the patient on the (e.g., an increased serum cholesterol level) and the examining table. presence of the disease in question. We do this by Fig. 1.4 shows a physician demonstrating that the studying the characteristics of groups and the charac- practice of clinical medicine relies heavily on population teristics of individuals.  8 SECTION I The Epidemiologic Approach to Disease and Intervention If we find there is indeed an association between an exposure and a disease, is it necessarily a causal 40 Viruses relationship? No, not all associations are causal. The second step therefore is to try to derive appropriate inferences about a possible causal relationship from the 20 patterns of the associations that have been previously Percent of children positive found. These steps are discussed in detail Chapter 14. Epidemiology often begins with descriptive data. Group A streptococci For example, Fig. 1.6 shows rates of gonorrhea in the United States in 2015 by state. Clearly, there are marked 40 regional variations in reported cases of gonorrhea. The first question to ask when we see such differences 20 between two groups or two regions or at two different times is, “Are these differences real?” In other words, are the data from each area of comparable quality? Before we try to interpret the data, we should be satisfied Mycoplasme pneumoniae 10 that the data are valid. If the differences are real, then we ask, “Why have these differences occurred?” Are