1.1: Meaning and Scope of Epidemiology

Epidemiology is a very old science, yet it did not flourish until after the "germ theory" of disease causation became established in the 1800s. Since that time, and until approximately 1960, epidemiology has been closely allied with microbiology in the battle against disease. Subsequent to 1960, epidemiology has become a more holistic discipline, and many factors in addition to the specific agent are investigated to determine their role as potential causes of disease (Schwabe 1982). Concurrently, the use of quantitative methods has become more widespread in epidemiologic research. In veterinary medicine the latter trend has been most pronounced in the last decade. As the emphasis both in veterinary education and practice shifts from the individual animal toward the population, the need for the veterinarian to have skills in quantitative methods will be accentuated. This text has been written in an attempt to assist veterinary students and veterinarians in developing quantitative epidemiologic skills that can be applied to population medicine. It contains a number of introductory epidemiological methods and examples of their application.

Epidemiology may be defined as the study of the patterns of disease that exist under field conditions. More specifically, epidemiology is the study of the frequency, distribution, and determinants of health and disease in populations. Thus, the epidemiology of a disease is the population analogue to the pathogenesis of disease in individuals, and in this context epidemiology is a fundamental science for medicine in populations.

To some, epidemiology is merely a set of methods; however, the use of these methods frequently leads the practitioner to a holistic, population· oriented way of thinking about health and disease that is quite different from the individual patient-oriented approach of clinical medicine. In many instances, the unit of concern in epidemiologic studies is not the individual but rather groups or categories of individuals such as the pen, herd, or flock. Despite this difference in unit of concern, epidemiology requires the same attention to detail and observer skills as clinical medicine and the other biologic sciences.

One method of exploring and understanding epidemiology is by elaborating the previous definition. First, it is noted that epidemiology is the study of the frequency and distribution of disease. Initial clues about the etiology of a disease .are often provided by its distribution. That is, information about what animals are affected and where and when a disease occurs often is suggestive of the causes of disease. Subsequently, it will be necessary to formally identify some of the determinants (causes) of the disease, (i.e., to explain why the disease occurs with the objective being to reduce its severity or frequency of occurrence). These details may be obtained by formally contrasting the characteristics of healthy versus diseased individuals, or by contrasting the characteristics of groups having a relatively high frequency of the disease versus groups having either none or a low frequency of the disease of interest. (Studies of the latter type are called case-control studies, and along with other types of analytic observational studies, they are introduced in Chapter 2 and elaborated in Chapter 6.)

Determinants, those factors that influence health and disease, are commonly called causes of disease. In epidemiology the word determinant is used to describe any factor that when altered produces a change in the frequency or characteristics of disease. Therefore, as will be stressed throughout this text, few diseases have a single cause. Host factors (such as age, breed, and sex) frequently are determinants of disease. Many determinants are external to the individual animal, as opposed to the internal factors that relate to the pathogenesis of disease. Putative causes of disease may be referred to as exposure or risk factors (or as independent, predictor, or explanatory variables) because they are suspected of producing the outcome of interest. The presumed effect, usually either health (as measured by productivity) or disease occurrence, is called the outcome, response, or dependent variable. (Variable refers to a property, factor, or characteristic of an individual or group being measured, rather than meaning "changeable.") For example, in a study of the association between immune status (e.g., level of serum antibodies) and the occurrence of disease, immune status is the independent and health status the dependent variable. If the impact of disease on the level of production were being studied, production would be the dependent variable and the presence or absence of disease the independent variable.

Disease and health are not redundant in the above definition, since in all epidemiologic studies both "diseased" and "healthy" animals should be present. As one example of their dual value, contrasting the characteristics of diseased versus healthy animals can provide valuable clues about the causes of disease. Nonetheless, health and disease are relative terms and their definitions usually depend on the circumstances in which they are applied. Hence some working definitions are in order.

Disease may affect individuals in either a subclinical or clinical form. Clinical disease represents the state of dysfunction of the body detectable by one or more of a person's senses. In contrast, subclinical disease represents a functional and/or anatomical abnormality of the body detectable only by selected laboratory tests or diagnostic aids. Although subclinical disease usually is less serious for the individual than clinical disease, it may be more important for the population because of its frequency. As a general rule, regardless of the primary cause(s) of the disease, the number of animals subclinically diseased will be much larger than the number clinically diseased. In this regard, it is particularly important to make a distinction between infection and disease. infection with most agents (including microorganisms and parasites) of so-called infectious diseases does not lead to clinical disease in the majority of infected animals. In many cases the infected animals appear to be healthy. For present purposes, an animal that is neither clinically nor subclinically diseased is by definit.ion healthy. Most populations comprise varying proportions of healthy, subclinically diseased, and clinically diseased individuals, with the proportions being subject to change over time.

Although health in humans has been defined as a state of complete physical, mental, and spiritual well being, in veterinary medicine, produc1.ivity is often used as a surrogate measure of health. In domestic animal populations, whether a disease is present or not is usually less important than the frequency with which the disease occurs and its subsequent impact on productivity. In this context, whereas disease may limit productivity, disease per se may not be the most important limiting factor of production. Other factors (such as management decisions, improper housing, or inadequate feeding practices) may have a greater impact on production in many situations (Williamson 1980). The association of these factors with health status may be investigated in a manner similar to studying the impact of disease on production using the techniques described in this text.

Due partly to the premise that the herd or flock is more important than the individual, the unit of concern for the epidemiologist frequently is an aggregate or population of animals, not an individual (e.g., it is more important that the feedlot is healthy than that a particular animal is healthy). Even when the individual is the unit of concern (e.g., in a study of the effect of vaccination on the health status of individuals), epidemiologic techniques are limited to groups (categories of individuals) rather than t.o an individual. Epidemiologists do observe individuals within the groups, but the conclusions are based on the experience of the group. Despite this limitation, inferences derived from groups may be extrapolated under certain circumstances to individuals (see 1.5). ("Population" is used through out this text in two senses- first, to describe the total number of animals in a group being studied who are biologically at risk of the event under study, and second, to ref er to the larger number of individuals of a particular type or species about which inferences are being made, based on information from a sample.)

One dimension for conceptualizing the structure of populations is that they are composed of a number of levels of organization. For example, the levels of organization from smallest to largest may be conceptualized in the following manner: cells of similar structure or function form organs, organs form body systems, and individuals are composed of body systems. Litters, pens, or herds are composed of a number of individuals; a collection of herds in one geographic area would form a local industry; and the local industries together would make up a larger animal industry, such as the swine or dairy industry. Each higher level of organization has characteristics beyond those of the lower levels. Individuals have more properties or characteristics than the sum of all the body systems; likewise, herds of animals have more properties than the individuals that compose them.

The level of organization selected for a specific study (the sampling unit in observational studies and the experimental unit in field trials) is the unit of analysis for that study. The unit of analysis often is not the individual animal; for example, if pens of pigs are the sampling units in an observational study, the unit of analysis would be the pen. Recognition of the correct unit of analysis is important for a number of reasons in addition to those already described. The unit of analysis may constrain the causal inferences about individuals that can be drawn from a sample (see 5.6.1) and, in addition, the unit of analysis is the basis for determining the degrees of freedom used in statistical testing.

It should be obvious from the definition and the preceding discussion that the setting for most epidemiologic work is the field (farm, animal clinic, city, nation, etc.) rather than the laboratory. Thus, epidemiologic observations relate to and are derived from field situations, although the analysis of data based on these observations may be conducted in a laboratory environment. Suitably stored and analyzed data will give the epidemiologic laboratory the same essential role in population medicine as the clinical pathology or microbiologic laboratory in individual animal medicine. In another sense, epidemiology is the diagnostic tool for populations, analogous to the role of clinical medicine as the diagnostic tool for individuals.

Finally, all animals, including humans, are possible subjects for epidemiologic study. Historically, epizootiology has been used to describe studies of disease in animal populations, and epidemiology for similar studies in human populations. Since a literal translation of "epidemiology" is the study (logos) of what is upon (epi) the population (demos) and because of the many similarities between human and animal medicine, there is little need to continue to use the term epizootiology. For those wishing to retain the distinction betwt'en studies of disease in animals and humans, the linguistic problems associated with this carried to the extreme would result in terms such as epiornithology, epiicthyology, and epiphytology to describe the study of diseases in populations of birds, fish, and plants respectively.