Epidemiology is the study of the distribution of disease and its determinants in human populations. Epidemiology usually takes place in an applied public health context. It focuses on the occurrence of disease by time, place, and person and seeks to identify and control outbreaks of disease through identification of etiological factors. Its approach is to identify associated risk factors and then work back to causes.
Historically, the focus of epidemiology was on large outbreaks, usually of infectious disease. The substance and methods of epidemiology have been applied to most forms of acute and chronic disease and many other physical and mental health conditions. Along with the broadening of the subject matter of epidemiology has come more focus on its methods. Along with the specialization of epidemiological methods has come the professional identification of persons as epidemiologists. In this entry, we address the origins, methodology, current topics, and professional issues related to epidemiology.
PROFESSIONAL ROLE
Epidemiology is the professional identification of an increasing number of persons who have received specialized training in departments of epidemiology in schools of public health. Many complete approved courses of study at the master’s or doctoral level leading to an M.P.H., Dr. P.H., or Ph.D. with epidemiology as an area of concentration. Some schools offer the M.P.H. to physicians after an abbreviated course of study. Like most other developing professions, graduate epidemiologists are protective of their professional identification and would seek to differentiate their professional credentials from those who have found their ways to epidemiological roles from other educational backgrounds. The problem in this developing professionalism is how to define a unique intellectual content or theory of epidemiology, when the applied nature of the discipline demands that the latest theories and methods of investigating disease be incorporated into any ongoing investigation.
ORIGINS
Langmuir (Roeche 1967, p. xiii) traces the origins of epidemiology as far back as Hypocrites’ report of an outbreak of mumps among Greek athletes, but the written history of mankind is full of major outbreaks of disease, which the practitioners of the time attempted to control with methods ranging from folk remedies and the available medicine, to religion and witchcraft, and even to civil and sanitary engineering. It was the careful and scientific observation of these outbreaks that led to the development of modern epidemiology.
Perhaps the most famous early epidemiological inquiry was John Snow’s careful observation of the house-by-house locations of incident cases of cholera during epidemics in London in the 1840s and 1850s. Snow’s correlation of new cases with some water supply systems, but not others, changed the conventional medical wisdom about how cholera was spread and did so before the microbial nature of the disease was discovered. Roeche’s Annals of Epidemiology provides a number of examples of the application of epidemiological methods in relation to local outbreaks in the earlier part of the twentieth century. More recent well-known investigations have included those relating to Legionnaires’ disease and AIDS.
Epidemiology as a discipline has grown along with the many associated disciplines of public health and medicine. Microbiology has identified the organisms and modes of transmission for many infectious diseases. Biochemistry, physiology, virology, and related basic medical sciences have provided the scientific background to support the development of the epidemiology of infectious disease. Other sciences ranging from genetics to sociology have contributed to the various specializations within epidemiology, which are broadly classified as infectious disease, chronic disease, cancer, cardiovascular disease, genetic, perinatal, reproductive, oral, occupational, environmental, air pollution, respiratory, nutritional, injury, substance abuse, psychiatric, social, and health care epidemiology as well as pharmaco-epidemiology.
A CONCEPTUAL PARADIGM
A typical paradigm in epidemiology focuses on the interaction of: host, agent, and environment. The host is typically a person but is sometimes another species or organism which provides a reservoir of infection that is subsequently transmitted to humans. In the presence of an epidemic, one tries to understand the characteristics of the host that provide susceptibility to the epidemic condition. Characteristics that are frequently considered are: (1) genetic characteristics; (2) biological characteristics, such as immunology, physiology, and anatomy; (3) demographic characteristics, such as age, sex, race, ethnicity, place of birth, and place of residence; (4) social and economic factors, such as socioeconomic status, education, and occupation; and (5) personal behaviors, such as diet, exercise, substance use, and use of health services.
The agents in this paradigm have classically been rats, lice, and insects. They may be biological, such as viruses or bacteria, as is the case in most infectious disease epidemiology. Typhus, cholera, smallpox, the Ebola virus, Legionnaires’ disease, and AIDS follow this model. Chemical agents have also contributed to significant epidemics. Some have been physiological poisons such as the mercury once used in hatmaking, lead in old paint, water pipes, or moonshine liquor; others have been carcinogens such as tobacco or PCBs found as a contaminant in oil spread on roads. Recent focus has been on alcohol and other drugs of abuse. Other physical agents have included asbestos, coal dust, and radioactive fallout. Guns, automobiles, and industrial equipment are also agents of injury and mortality. Sometimes the ”agent” has been a nutritive excess or deficiency, demonstrated in Goldberger’s 1915 discovery that a deficiency of niacin, part of the vitamin B complex, causes pellagra. Similarly, stress and other elements of lifestyle have been investigated as contributors to cardiovascular disease. An element in the consideration of agents is the degree of exposure and whether the agent alone is a necessary or sufficient cause of the disorder.
The third element of this paradigm is the environment, which may promote the presence of an agent or increase host susceptibility to the agent. Many elements of the environment may be relevant to the disease process, including (1) the physical environment, such as climate, housing, and degree of crowding; (2) the biological environment, including plant and animal populations, especially humans; and (3) the demographic and socioeconomic environments of the host.
At the core of this simple paradigm is the potential for multiple paths of causation for an epidemic. Rarely is a single cause sufficient to ensure the onset of a disease or condition. Thus the approach considers alternative modes of transmission, such as the influence of a common agent versus transmission from host to host. For many agents there is an incubation period, with delayed onset after ”infection” or contact with the agent. Further, exposure may lead to a spectrum of disease with variation in the type and severity of the response to the agent. An important element to observe in any outbreak is who is not affected and whether that is due to immunity or some other protective factor.
METHODOLOGY
Field Methods. The historical method of epidemiology began with the observation that an epidemic was present, with the initial response beginning with ”shoe leather” investigation of who was affected and how. The initial approach focused on identified cases and their distribution over time and place, leading to a methodology that was often able to discern the risk factors for a disease even before a particular pathogen could be identified. The description of cases in terms of geography and environment as well as various demographic characteristics and exposures remains at the core of epidemiology.
Essential complements to epidemiologic field methods are careful clinical observation, measurement, and classification of the disorder, as well as laboratory identification of any pathogens and identification of potential risk factors for the observed disorder. This methodology is drawn from the methods of associated fields such as pathology, bacteriology, virology, immunology, and molecular genetics. Efforts at classification of pathogens have been augmented by the collection of libraries of reference specimens from past outbreaks.
A problem with the investigation of known outbreaks is that many types of epidemic may develop unobserved until a significant proportion of the population is affected. This has led to the development of surveillance strategies. Among these have been the reporting of multiple causes on death certificates and the mandatory reporting to the health department of many communicable diseases, such as tuberculosis and sexually transmitted diseases. The early detection and reporting of outbreaks make various public health interventions possible. Statistical analysis of mortality and morbidity has become a major component of the public health systems of most countries.
Case Registers. An extension to this approach involves case registers, in which identified mortality or morbidity is investigated and compiled in a statistical database, with subsequent investigation of the detailed context for each case. Case registers for particular types of diseases typically identify a case and then collect additional information through review of medical charts and direct field investigations and interviews. Although not as detailed as most case registers, large databases of treated disorders are gathered by various funders or providers of health services. These include the Health Care Finance Administration (Medicare and Medicaid) and various private insurance companies.
The systematic analysis of large databases, whether case registers or those of administrative agencies, makes it possible to monitor the prevalence of various conditions and to detect increases in prevalence or outbreaks long before they would be detected by individual clinicians. Analyses of mortality and morbidity are presented by most public health agencies. Typical methods include the presentation of rates of disorder, such as the number of cases per 100,000 population for geographic and demographic subpopulations. Analyses might ask whether rates are higher in one place compared to another, for a particular birth cohort, or for persons of a particular socioeconomic status. Such comparisons may be crude or adjusted for factors such as age and sex, which may bias such comparisons. Sometimes the analyses provide detailed age- and sex-specific comparisons or breakouts by other risk factors. Yet this approach is dependent on people with diseases being identified through the health system.
Prevalence Surveys. In order to discover the true prevalence of various conditions and risk factors, it is possible to conduct sample surveys of the population of a country or other geographic unit. This approach avoids potential reporting bias from the health care system and can identify conditions that might not otherwise be recognized or reported. Typically a representative sample of persons would be interviewed about their health, and sometimes various examinations or tests would be administered. Large comprehensive surveys of health and nutrition are conducted regularly by the National Center for Health Statistics. More specialized surveys of particular conditions such as blood pressure, substance use, and mental health have been conducted by federal, state, and private agencies. Because participation in most such surveys is voluntary, investigative procedures are usually limited to interviews and test procedures with little discomfort or risk. The limitation of prevalence surveys is that they are expensive and may have respondent selection biases. On the other hand, they have great potential for collecting detailed information about disorders and potential risk factors.
Case-Control Studies. The case-control method is used extensively in epidemiology. This approach typically starts with a sample of cases of a particular disorder and identifies one or more samples of persons who appear similar but who do not have the disorder. Careful comparison of the groups has potential for identifying risk factors associated with having the disorder or, alternatively, factors that are protective. The most important issue in designing case-control studies is the designation of an appropriate control group in such a way that the process selecting controls neither hides the real risk factors nor pinpoints apparent but false ones. Control groups are frequently designated by geographic or ecological variables and are often matched on individual characteristics such as age and sex. These methods are somewhat similar to those used in natural experiments, in which two groups differ on one or more risk factors and the rates of disorder are compared, but the case-control method starts with identified cases.
Cohort Studies. Unlike the prevalence survey and most case-control studies, which are cross-sectional or retrospective in nature, a cohort study endeavors to identify groups of persons who do not have the disorder or disorders in question and then follows them prospectively through the occurrence of the disease, with ascertainment of factors likely to contribute to the etiology of the disease. Such studies may start with a general sample of the population or may select groups based on the presence or absence of hypothesized risk factors prior to onset of the disorder. They then follow the samples for incident disorder. Unlike cross-sectional studies, cohort studies provide the possibility of determining causal order for risk factors and thus avoid the possibility that apparent risk factors are simply consequences of the disease. The main problem with true cohort studies is that they must last as long as it takes for the disease to develop, which may even take longer than the working life of the investigator. Sometimes, however, cohorts can be found that have been identified and assessed historically and thus can be compared in the present. Examples may be insurance groups, occupational groups, or even residents of certain areas. Then the task is to gather the information on exposure and relate it to health outcomes.
Experimental and Quasi-Experimental Methods. A true experiment is dependent on random assignment of persons or groups to conditions, but one would hardly assign a person to a condition known to produce a serious disease. Nonetheless, there are research designs that approximate true experiments. The most direct is an intervention study in which an intervention that cannot be provided to everyone is provided to one group but not another—or provided to the alternative group at a later time. A quasi-experimental design would identify a group that has been exposed to a risk factor and compare it to another that has not. If the factors that govern the initial exposure appear to be truly accidental, then the quasi-experiment may be nearly as random as a true experiment. In such circumstances or in true randomization, one can determine the effect of the risk factor on outcomes with little risk of the outcome determining the status of the risk factor.
Alternative Methods. There are a number of research designs found in epidemiology. Those presented above are typical but certainly not exhaustive of the alternatives. At the core of nearly all of these methods, however, is the identification of factors associated with the incidence or prevalence of particular types of disorder.
CURRENT ISSUES
In recent epidemiological literature there has been some concern about the future development of epidemiology. A part of this discussion focuses on the basic paradigm being used and whether it has shifted from an ecological approach consistent with the spread of infectious disease and related pathogens to individual risk factors that are more closely related to the current emphasis on chronic diseases. The growing influence of molecular and genetic methods increases this tendency. It has been suggested that this focus on individual and proximate causes blames the individual rather than fixing the prior cause. Some authors have suggested that epidemiology should regain its public health orientation by focusing more on ecological issues affecting the risk status of groups rather than on individual-level factors, and one suggested that the primary focus should be on fighting poverty. A more integrative approach has been suggested by Susser and Susser (1996a, 1996b), who criticize the field for continuing to rely on a multiple risk factor ”black box” approach, which is of declining utility, in favor of an approach that encompasses multiple levels of organization and causes from the molecular to the societal.
A second area of concern, identified by Bracken (1998), is the relationship of epidemiology to corporate litigation when various rare disorders are related to sources of risk such as cellular telephones, breast implants, and the like. The issue is the extent to which epidemiologists can provide risk information to journalists and the public without undue burden from unbridled discovery and from the inconsistencies of findings of rare exposures related to rare diseases.
A third issue appears to be the search for professional identity in a discipline that has applicability across the full range of health professions, human and otherwise. The professional positions identified in a 1999 Internet search indicated that there is a continuing market for persons with epidemiological training but that these positions are distributed widely in departments supporting specific disciplines or located in various public health settings. Thus the job titles are often designated in a hyphenated form, such as psychiatric-epidemiologist, cancer-epidemiologist, and genetic-epidemiologist. For a more complete introduction to the principles and methods of epidemiology see Kleinbaum, Kupper, and Morgenstern (1982), Littlefeld and Stolley (1994), or MacMahon and Trichopolous (1996).
