Make a PowerPoint summarizing Chapter 5? It should include the major key points. C It doesn’t have to be fancy. Down below is the chapter attached.
Epidemiology
Holly B. Cassells
OUTLINE
Use of Epidemiology in Disease Control and Prevention
Calculation of Rates
Morbidity: Incidence and Prevalence Rates
Other Rates
Concept of Risk
Use of Epidemiology in Disease Prevention
Primary Prevention
Secondary and Tertiary Prevention
Establishing Causality
Screening
Surveillance
Use of Epidemiology in Health Services
Epidemiological Methods
Descriptive Epidemiology
Analytic Epidemiology
OBJECTIVES
Upon completion of this chapter, the reader will be able to do the following:
1. Identify epidemiological models used to explain disease and health patterns in populations.
2. Use epidemiological methods to describe the state of health in a community or aggregate.
3. Calculate epidemiological rates in order to characterize population health.
4. Understand the use of epidemiological methods in primary, secondary, and tertiary prevention.
5. Evaluate epidemiological study designs for researching health problems.
KEY TERMS
age adjustment of rates
age-specific rates
analytic epidemiology
attack rates
cause-and-effect relationship
crude rates
177
descriptive epidemiology
ecosocial epidemiology
epidemiological triangle
epidemiology
incidence rates
infant mortality rate
morbidity rates
mortality rates
natural history of disease
person–place–time model
prevalence rate
proportionate mortality ratio
rates
risk
risk factors
screening
screening programs
standardization of rates
surveillance
web of causation
Epidemiology is the study of the distribution and determinants of health and disease in human
populations (World Health Organization, 2017) and is the principal science of public health. It
entails a body of knowledge derived from epidemiological research and specialized
epidemiological methods and approaches to scientific research. Community health nurses use
epidemiological concepts to improve the health of population groups by identifying risk factors and
optimal approaches that reduce disease risk and promote health. Epidemiological methods are
important for accurate community assessment and diagnosis and in planning and evaluating
effective community interventions. This chapter discusses the uses of epidemiology and its
specialized methodologies.
178
Use of Epidemiology in Disease Control and Prevention
Although epidemiological principles and ideas originated in ancient times, formal epidemiological
techniques developed in the nineteenth century. Early applications focused on identifying factors
associated with infectious diseases and the spread of disease in the community. Public health
practitioners hoped to improve preventive strategies by identifying critical factors in disease
development.
Specifically, investigators attempted to identify characteristics of people who had a disease such
as cholera or plague and compared them with characteristics of those who remained healthy. These
differences might include a broad range of personal factors, such as age, gender, socioeconomic
status, and health status. Investigators also questioned whether there were differences in the
location or living environment of ill people compared with healthy individuals and whether these
factors influenced disease development. Finally, researchers examined whether common time
factors existed (i.e., when people acquired disease). Use of this person-place-time model organized
epidemiologists’ investigations of the disease pattern in the community (Box 5.1). This study of the
amount and distribution of disease constitutes descriptive epidemiology. Identified patterns
frequently indicate possible causes of disease that public health professionals can examine with
more advanced epidemiological methods.
In addition to investigating the person, place, and time factors related to disease, epidemiologists
examine complex relationships among the many determinants of disease. This investigation of the
causes of disease, or etiology, is called analytic epidemiology.
Even before the identification of bacterial agents, public health practitioners recognized that
single factors were insufficient to cause disease. For example, while exploring the cholera epidemics
in London in 1855, Dr. John Snow collected data about social and physical environmental
conditions that might favor disease development. He specifically examined the contamination of
local water systems. Snow also gathered information about people who became ill—their living
patterns, water sources, socioeconomic characteristics, and health status. A comprehensive database
helped him develop a theory about the possible cause of the epidemic. Snow suspected that a single
biological agent was responsible for the cholera infection, although the organism, Vibrio cholerae,
had yet to be discovered. He compared the death rates among individuals using one water well
with those among people using a different water source. His findings suggested an association
between cholera and water quality (Box 5.2).
B O X 5 . 1 Pe r s o n – Pl a c e – T i m e M o d e l
Person: “Who” factors, such as demographic characteristics, health, and disease status
Place: “Where” factors, such as geographic location, climate and environmental conditions,
and political and social environment
Time: “When” factors, such as time of day, week, or month and secular trends over months
and years
The epidemiologist examines the interrelationships between host and environmental
characteristics and uses an organized method of inquiry to derive an explanation of disease. This
model of investigation is called the epidemiological triangle because the epidemiologist must
analyze the following three elements: agent, host, and environment (Fig. 5.1). The development of
disease depends on the extent of the host’s exposure to an agent, the strength or virulence of the
agent, and the host’s genetic or immunological susceptibility. Disease also depends on the
environmental conditions existing at the time of exposure, which include the biological, social,
political, and physical environments (Table 5.1). The model implies that the rate of disease will
change when the balance among these three factors is altered. By examining each of the three
elements, a community health nurse can methodically assess a health problem, determine
protective factors, and evaluate the factors that make the host vulnerable to disease.
Conditions linked to clearly identifiable agents, such as bacteria, chemicals, toxins, and other
exposure factors, are readily explained by the epidemiological triangle. However, other models that
179
stress the multiplicity of host and environmental interactions have developed, and understanding
of disease has progressed. The “wheel model” is an example of such a model (Fig. 5.2). The wheel
consists of a hub that represents the host and its human characteristics, such as genetic makeup,
personality, and immunity. The surrounding wheel represents the environment and comprises
biological, social, and physical dimensions. The relative size of each component in the wheel
depends on the health problem. A relatively large genetic core represents health conditions
associated with heredity. Origins of other health conditions may be more dependent on
environmental factors (Mausner and Kramer, 1985). This model subscribes to multiple-causation
rather than single-causation disease theory; therefore it is more useful for analyzing complex
chronic conditions and identifying factors that are amenable to intervention.
After the discovery of the causative agents of many infectious diseases, public health
interventions eventually resulted in a decline in widespread epidemic mortality, particularly in
developed countries. The focus of public health then shifted to chronic diseases such as cancer,
coronary heart disease, and diabetes during the past few decades. The development of these chronic
diseases tends to be associated with multiple interrelated factors rather than single causative agents.
In studying chronic diseases, epidemiologists use methods that are similar to those used in
infectious disease investigation, thereby developing theories about chronic disease control. Risk
factor identification is of particular importance to chronic disease reduction. Risk factors are
variables that increase the rate of disease in people who have them (e.g., a genetic predisposition) or
in people exposed to them (e.g., an infectious agent or a diet high in saturated fat). Therefore their
identification is critical to identifying specific prevention and intervention approaches that
effectively and efficiently reduce chronic disease morbidity and mortality. For example, the
identification of cardiovascular disease risk factors has suggested a number of lifestyle
modifications that could reduce the morbidity risk before disease onset. Primary prevention
strategies, such as dietary saturated fat reduction, smoking cessation, and hypertension control,
were developed in response to previous epidemiological studies that identified these risk factors
(Box 5.3). The web of causation model illustrates the complexity of relationships among causal
variables for heart disease (Fig. 5.3).
BOX 5.2 Example of the Epidemiological Approach
An early example of the epidemiological approach is John Snow’s investigation of a cholera
epidemic in the 1850s. He analyzed the distribution of person, place, and time factors by
comparing the death rates among people living in different geographic sectors of London. His
geographic map of cases, shown here, is an early example of the use of geographic information to
formulate a hypothesis about the causes of an epidemic. Snow noted that people using a particular
water pump had significantly higher mortality rates from cholera than people using other water
sources in the city. Although the cholera organism was yet unidentified, the clustering of disease
cases around one neighborhood pump suggested new prevention strategies to public health
officials (i.e., that cholera might be reduced in a community by controlling contaminated drinking
water sources). As an immediate response, in September 1854, Snow persuaded local leaders to
remove the handle of the pump, which to this day can be seen on Broadwick Street in London
(Snow, 1855.)
180
John Snow’s map of London neighborhood showing location of cholera case cluster surrounding the
Broad Street water pump.
Published by C. F. Cheffins, Lith, Southhampton Buildings, London, England, 1854. In Snow J, editor: On
the mode of communication of cholera, ed 2, London, 1855, John Churchill. Retrieved from:
http://www.ph.ucla.edu/epi/snow/snowmap1_1854_lge.html.
181
Replica of the famous Broad Street pump. Broadwick Street, London.
Photo courtesy of http://commons.wikimedia.org/wiki/File:John_Snow_memorial_and_pub.jpg, Creative
Commons Attribution-Share Alike 2.0 Generic [CC BY-SA 2.0].
182
FIG. 5.1
Epidemiological triangle.
A newer paradigm, ecosocial epidemiology, challenges the more individually focused risk factor
approach to understanding disease origins. This ecosocial approach emphasizes the role of evolving
macro-level socioenvironmental factors, including complex political and economic forces, along
with microbiological processes, in understanding health and illness (Smith and Lincoln, 2011).
Investigating the context of health will necessitate alternative research approaches, such as
qualitative and ecological studies and studies of social institutions and processes. In turn, the
examination of social and contextual origins will enlighten the interventions of public health
practitioners.
TABLE 5.1
A Classification of Agent, Host, and Environmental Factors that Determine the Occurrence of Diseases in Human
Populations
183
Modified from Lilienfeld DE, Stoley PD: Foundations of epidemiology, New York, NY, 1994, Oxford
University Press.
For example, Buffardi et al. (2008) analyzed the ecosocial and psychosocial correlates of diagnosis
of sexually transmitted infections (STIs) among young adults. Specifically, they examined STI
diagnosis within “contextual conditions” such as low income, “housing insecurity,” childhood
physical or sexual abuse, intimate partner abuse, gang participation, personal history of having
been arrested, and drug/alcohol use. It was determined that STIs were statistically associated with
housing insecurity, exposure to crime, and having been arrested. The researchers concluded that
ecosocial or contextual conditions strongly enhance STI risk by increasing sexual risk behaviors and
likelihood of exposure to infection.
In another study, Phillips (2011) applied an ecosocial perspective when examining the effects of
social/contextual factors on adherence to antiretroviral therapy (ART) among black men who tested
positive for human immunodeficiency virus (HIV). He examined both individual factors (e.g.,
psychological state of mind, psychological distress, illicit drug use) and interpersonal/social
184
contextual factors (e.g., partner status, housing status, patient–provider relationship, social capital
[groups/networks]). He concluded that adherence to the medication regimen was strongly
associated with homelessness and how well the individual tolerated the ART. Other factors
included the individual’s state of mind and illicit drug use. Practice implications included the
observation that providers should assess social and behavioral factors and intervene accordingly.
This would include identification of psychological distress or presence of substance abuse. He also
suggested assessment of housing status and facilitation of effective patient–provider relationships
to mitigate tolerability issues with ART.
185
FIG. 5.2 Wheel model of human–environment interaction.
Redrawn from Mausner JS, Kramer S: Mausner and Bahn epidemiology: an introductory text, ed 2,
Philadelphia, 1985, Saunders
BOX 5.3 Coronary Heart Disease (Chd) Risk Factors Supported by
Epidemiological Data From the Framingham Study
• Age
• Gender (male)
• Current cigarette smoking
• Hypertension
• High level of low-density lipoprotein (LDL) cholesterol
• Low level of high-density lipoprotein (HDL) cholesterol
• (Diabetes)1
• Family history of premature coronary heart disease2
1
Diabetes is not included in the Framingham Global Risk Score but is now considered to be a
186
coronary heart disease risk equivalent, meaning that persons with diabetes will be treated as
intensively as those with coronary heart disease.
2
Included in NCEP list of major risk factors but not in the Framingham Global Risk Score.
Data from the Executive summary of the third report of the National Cholesterol Education
Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in
adults (adult treatment panel III), JAMA 285:2486–2497, 2001. The 2013 recommendations from the
American College of Cardiology/American Heart Association indicate these risk factors predict 10year cardiovascular disease incidence (rather than CHD as in NCEP ATP III), Circulation, 2013.
http://dx.doi.org/10.1161/01.cir.0000437741.48606.98, 2013.
FIG. 5.3 The web of causation for myocardial infarction: A current view.
From Friedman GD: Primer of epidemiology, ed 5, New York, 2004, McGraw-Hill.
187
Calculation of Rates
The community health nurse must analyze data about the health of the community to determine
disease patterns. The nurse may collect data by conducting surveys or compiling data from existing
records (e.g., data from clinic facilities or vital statistics records). Assessment data often are in the
form of counts or simple frequencies of events (e.g., the number of people with a specific health
condition). Community health practitioners interpret these raw counts by transforming them into
rates.
Rates are arithmetic expressions that help practitioners consider a count of an event relative to
the size of the population from which it is extracted (e.g., the population at risk). Rates are
population proportions or fractions in which the numerator is the number of events occurring in a
specified period. The denominator consists of those in the population at the specified time period
(e.g., per day, per week, or per year), frequently drawing on demographic data from the U.S.
census. This proportion is multiplied by a constant (k) that is a multiple of 10, such as 1000, 10,000,
or 100,000. The constant usually converts the resultant number to a whole number, which is larger
and easier to interpret. Thus a rate can be the number of cases of a disease occurring for every 1000,
10,000, or 100,000 people in the population, as follows:
When raw counts or numbers are converted to rates, the community health nurse can make
meaningful comparisons with rates from other cities, counties, districts, or states; from the nation;
and from previous periods. These analyses help the nurse determine the magnitude of a public
health problem in a given area and allow more meaningful and reliable tracking of trends in the
community over time (Box 5.4).
Sometimes a ratio is used to express a relationship between two variables. A ratio is obtained by
dividing one quantity by another, and the numerator is not necessarily part of the denominator. For
example, a ratio could contrast the number of male births to that of female births. Proportions can
describe characteristics of a population. A proportion is often a percentage, and it represents the
numerator as part of the denominator.
B O X 5 . 4 U s i n g R a t e s i n E ve r y d a y C o m m u n i t y H e a l t h N u r s i n g P r a c t i c e
The following school situation exemplifies the value of rates:
A community health nurse screened 500 students for tuberculosis (TB) in Southside School and
identified 15 students with newly positive tuberculin test results. The proportion of Southside
School students affected was 15/500, or 0.03 (3%), or a rate of 30/1000 students at risk for TB.
Concurrently, the nurse conducted screening in Northside School and again identified 15 students
with positive tuberculin test results. However, this school was much larger than the Southside
School and had 900 potentially at-risk students. To place the number of affected students in
perspective relative to the size of the Northside School, the nurse calculated a proportion of 15/900,
or 0.017 (1.7%), or a rate of 17/1000 students at risk in Northside School.
On the basis of this comparison, the nurse concluded that although both schools had the same
number of tuberculin conversions, Southside School had the greater rate of tuberculin test
conversions. To determine whether rates are excessively high, the nurse should compare rates with
the city, county, and state rates and then explore reasons for the difference in these rates.
Morbidity: Incidence and Prevalence Rates
The two principal types of morbidity rates, or rates of illness, in public health are incidence rates
188
and prevalence rates. Incidence rates describe the occurrence of new cases of a disease (e.g.,
tuberculosis, influenza) or condition (e.g., teen pregnancy) in a community over a given period
relative to the size of the population at risk for that disease or condition during that same period.
The denominator consists of only those at risk for the disease or condition; therefore known cases or
those not susceptible (e.g., those immunized against a disease) are subtracted from the total
population (Table 5.2):
The incidence rate may be the most sensitive indicator of the changing health of a community
because it captures the fluctuations of disease in a population. Although incidence rates are
valuable for monitoring trends in chronic disease, they are particularly useful for detecting shortterm changes in acute disease—such as those that occur with influenza or measles—in which the
duration of the disease is typically short.
If a population is exposed to an infectious disease at a given time and place, the nurse may
calculate the attack rate, a specialized form of the incidence rate. Attack rates document the number
of new cases of a disease in those exposed to the disease. A common example of the application of
the attack rate is food poisoning; the denominator is the number of people exposed to a suspect
food, and the numerator is the number of people who were exposed and became ill. The nurse can
calculate and compare the attack rates of illness among those exposed to specific foods to identify
the critical food sources or exposure variables.
TABLE 5.2
Examples of Rate Calculations
A prevalence rate is the number of all cases of a specific disease or condition (e.g., deafness) in a
population at a given point in time relative to the population at the same point in time:
When prevalence rates describe the number of people with the disease at a specific point in time,
they are sometimes called point prevalences. For this reason, cross-sectional studies frequently use
them. Period prevalences represent the number of existing cases during a specified period or interval
of time and include old cases and new cases that appear within the same period.
189
Prevalence rates are influenced by the number of people who experience a particular condition
(i.e., incidence) and the duration of the condition. A nurse can derive the prevalence rate (P) by
multiplying incidence (I) by duration (D): (P = I × D). An increase in the incidence rate or the
duration of a disease increases the prevalence rate of a disease. With the advent of life-prolonging
therapies (e.g., insulin for treatment of type 1 diabetes and antiretroviral drugs for treatment of
HIV), the prevalence of a disease may increase without a change in the incidence rate. Those who
survive a chronic disease without cure remain in the “prevalence pot” (Fig. 5.4). For conditions such
as cataracts, surgical removal of the cataracts permits many people to recover and thereby move out
of the prevalence pot. Although the incidence has not necessarily changed, the reduced duration of
the disease (because of surgery) lowers the prevalence rate of cataracts in the population.
190
FIG. 5.4 Prevalence pot: The relationship between incidence and prevalence.
Redrawn from Morton RF, Hebel JR, McCarter RJ: A study guide to epidemiology and biostatistics, ed 3,
Gaithersburg, MD, 1990, Aspen Publishers.
Morbidity rates are not available for many conditions because surveillance of many chronic
diseases is not widely conducted. Furthermore, morbidity rates may be subject to underreporting
when they are available. Routinely collected birth and death rates, or mortality rates, are more
widely available. Table 5.2 provides examples of calculating selected rates.
Other Rates
Numerous other rates are useful in characterizing the health of a population. For example, crude
rates summarize the occurrence of births (i.e., crude birth rate), mortality (i.e., crude death rates), or
diseases (i.e., crude disease rates) in the general population. The numerator is the number of events,
and the denominator is the average population size or the population size at midyear (i.e., usually
July 1) multiplied by a constant.
The denominators of crude rates represent the total population and not the population at risk for
a given event; therefore these rates are subject to certain biases in interpretation. Crude death rates
are sensitive to the number of people at the highest risk for dying. A relatively older population will
probably produce a higher crude death rate than a population with a more evenly distributed age
range. Conversely, a young population will have a somewhat lower crude death rate. Similar biases
can occur for crude birth rates (e.g., higher birth rates in young populations).
This distortion occurs because the denominator reflects the entire population and not exclusively
the population at risk for giving birth. Age is one of the most common confounding factors that can
191
mask the true distribution of variables. However, many variables, such as race and socioeconomic
status, can also bias the interpretation of biostatistical data. Therefore the nurse may use several
approaches to remove the confounding effect of these variables on rates.
Age-specific rates characterize a particular age group in the population and usually consider
deaths and births. Determining the rate for specific subgroups of a population and using a
denominator that reflects only that subgroup remove age bias:
To characterize a total population using age-specific rates, one must compute the rate for each
category separately. The reason is that a single summary rate, such as a mean, cannot adequately
characterize a total population. Specific rates for other variables can be determined in a similar
fashion (e.g., race-specific or gender-specific rates) (Table 5.3).
Age adjustment or standardization of rates is another method of reducing bias when there is a
difference between the age distributions of two populations. The nurse uses either the direct
method or the indirect standardization method. The direct method selects a standard population,
which is often the population distribution of the United States. This method essentially converts
age-specific rates for age categories of the two populations to those of the standard population, and
it calculates a summary age-adjusted rate for each of the two populations of interest. This
conversion enables the nurse to compare the two rates as if both had the standard population’s age
structure (i.e., without the prior problem of age distortion).
The proportionate mortality ratio (PMR) method also describes mortality. It represents the
percentage of deaths resulting from a specific cause relative to deaths from all causes. It is often
helpful in identifying areas in which public health programs might make significant contributions
to reducing deaths. In some situations, a high PMR may reflect a low overall mortality or reduced
number of deaths resulting from other causes. Therefore the PMR requires consideration in the
context of the mortality experience of the population.
TABLE 5.3
Comparison of U.S. Mortality Rates—2014 (Preliminary)
Death Rate
Crude death rate
Age-adjusted death rate
Age-specific death rates (years):
Modified from Mausner JS, Kramer S: Mausner and Bahn epidemiology: an introductory text, ed 2,
Philadelphia, 1985, Saunders. Rates from U.S. Department of Health and Human Services/National
Center for Health Statistics: Health, United States, 2015. Retrieved from:
http://www.cdc.gov/nchs/data/hus/hus15.pdf.
Active Learning Exercise
1. Compile a database of relevant demographic and epidemiological data for your community
by examining census reports, vital statistics reports, city records, and other sources in
libraries and agencies.
2. Using numerators from vital statistics and denominators from census data, compute crude
death and birth rates for your community.
3. Compare morbidity and mortality rates for your community with those of the state and the
nation. Determine whether your community rates are higher or lower, and hypothesize
about reasons for any disparities.
194
Concept of Risk
The concepts of risk and risk factor are familiar to community health nurses whose practices focus
on disease prevention. Risk refers to the probability of an adverse event (i.e., the likelihood that
healthy people exposed to a specific factor will acquire a specific disease). Risk factor refers to the
specific exposure factor, such as cigarette smoke, hypertension, high cholesterol, excessive stress,
high noise levels, or an environmental chemical. Frequently, the exposure factor is external to the
individual. Risk factors may include fixed characteristics of people, such as age, sex, and genetic
makeup. Although these intrinsic risk factors are not alterable, certain lifestyle changes may reduce
their effect. For example, weight-bearing exercise and taking calcium and hormonal supplements
may reduce the risk of osteoporosis for susceptible women.
Epidemiologists describe disease patterns in aggregates and quantify the effects of exposure to
particular factors on the disease rates. To identify specific risk factors, epidemiologists compare
rates of disease for those exposed with those not exposed. One method for comparing two rates is
subtracting the rate of nonexposed individuals from the exposed. This measure of risk is called the
attributable risk; it is the estimate of the disease burden in a population. For example, if the rate of
non–insulin-dependent diabetes were 5000 per 100,000 people in the obese population (i.e., those
weighing more than 120% of ideal body weight) and 1000 per 100,000 people in the nonobese
population, the attributable risk of non–insulin-dependent diabetes resulting from obesity would
be:
This means that 4000 cases per 100,000 people may be attributed to obesity. Thus a prevention
program designed to reduce obesity could theoretically eliminate 4000 cases per 100,000 people in
the population. Attributable risks are particularly important in describing the potential impact of a
public health intervention in a community.
A second measure of the excess risk caused by a factor is the relative risk ratio. The relative risk is
calculated by dividing the incidence rate of disease in the exposed population by the incidence rate
of disease in the nonexposed population. In the previous example, a relative risk of 5 was obtained
by dividing 5000/100,000 by 1000/100,000. This risk ratio suggests that an obese individual has a
fivefold greater risk of diabetes than a nonobese individual. In general, a relative risk of 1 indicates
no excessive risk from exposure to a factor; a relative risk of 1.5 indicates a 50% increase in risk; a
relative risk of 2 indicates twice the risk; and a relative risk of less than 1 suggests that a factor may
have a protective effect associated with a reduced disease rate.
The relative risk ratio forms the statistical basis for the risk factor concept. Relative risks are
valuable indicators of the excess risk incurred by exposure to certain factors. They have been used
extensively in identifying the major causal factors of many common diseases, and they direct public
health practitioners’ efforts to reduce health risks.
Community health nurses may apply the concept of relative risk to suspected exposure variables
to isolate risk factors associated with community health problems. For example, a community
health nurse might investigate an outbreak of probable foodborne illness. The nurse may compare
the incidence rate among those exposed to potato salad in a school cafeteria with the incidence rate
among those not exposed. The relative risk calculated from the ratio of these two incidence rates
indicates the amount of excess risk for disease incurred by eating the potato salad. A community
health nurse might also determine the relative risks for other suspected foods and compare them
with the relative risk for potato salad. Attack rates are the calculated incidence rates for foods
involved in foodborne illnesses. A food with a markedly higher relative risk than other foods might
be the causal agent in a foodborne epidemic. The identification of the causal agent, or specific food,
is critical to the implementation of an effective prevention program such as teaching proper foodhandling techniques.
195
196
Use of Epidemiology in Disease Prevention
Primary Prevention
The central goals of epidemiology are describing the disease patterns, identifying the etiological
factors in disease development, and taking the most effective preventive measures. These
preventive measures are specific to the stage of disease progression, or the natural history of
disease, from prepathogenesis through resolution of the disease process. When interventions occur
before disease development, they are called primary prevention. Primary prevention relies on
epidemiological information to indicate those behaviors that are protective, or those that will not
contribute to an increase in disease, and those that are associated with increased risk.
Two types of activities constitute primary prevention. Those actions that are general in nature
and designed to foster healthful lifestyles and a safe environment are called health promotion.
Actions aimed at reducing the risk of specific diseases are called specific protection. Public health
practitioners use epidemiological research to understand practices that are likely to reduce or
increase disease rates. For example, numerous research studies have confirmed that regular exercise
is an important health promotion activity that has positive effects on general physical and mental
health. Immunizations exemplify specific protection measures that reduce the incidence of
particular diseases.
Secondary and Tertiary Prevention
Secondary prevention occurs after pathogenesis. Those measures designed to detect disease at its
earliest stage, namely screening and physical examinations that are aimed at early diagnosis, are
secondary prevention. Interventions that provide for early treatment and cure of disease are also in
this category. Again, epidemiological data and clinical trials determining effective treatments are
crucial in disease identification. Mammography, guaiac testing of feces, and the treatment of
infections and dental caries are all examples of secondary prevention.
Tertiary prevention focuses on limitation of disability and the rehabilitation of those with
irreversible diseases such as diabetes and spinal cord injury. Epidemiological studies examine risk
factors affecting function and suggest optimal strategies in the care of patients with chronic
advanced disease.
Establishing Causality
As discussed earlier, a principal goal of epidemiology is to identify etiological factors of diseases to
encourage the most effective prevention activities and develop treatment modalities. During the last
few decades, researchers recognized that many diseases have not one but multiple causes.
Epidemiologists who examine disease rates and conduct population-focused research often find
multiple factors associated with health problems. For example, cardiovascular disease rates may
vary by location, ethnicity, and smoking status. Even infectious diseases often require not only an
organism but also certain behaviors or conditions to cause exposure. Determining the extent that
these correlates represent associative or causal relationships is important for public health
practitioners who seek to prevent, diagnose, and treat disease.
Definitively establishing causality—particularly in chronic disease—is a challenge. The following
six criteria establish the existence of a cause-and-effect relationship:
1. Strength of association: Rates of morbidity or mortality must be higher in the exposed
group than in the nonexposed group. Relative risk ratios, or odds ratios, and correlation
coefficients indicate whether the relationship between the exposure variable and the
outcome is causal. For example, epidemiological studies demonstrated a higher relative risk
for heart disease among smokers than among nonsmokers.
2. Dose–response relationship: An increased exposure to the risk factor causes a concomitant
increase in disease rate. Indeed, the risk of heart disease mortality is higher for heavy
smokers than for light smokers.
3. Temporally correct relationship: Exposure to the causal factor must occur before the effect,
or disease. For heart disease, smoking history must precede disease development.
197
4. Biological plausibility: The data must make biological sense and represent a coherent
explanation for the relationship. Nicotine and other tobacco-derived chemicals are toxic to
the vascular endothelium. In addition to raising low-density lipoprotein (LDL) and
decreasing high-density lipoprotein (HDL) cholesterol levels, cigarette smoking causes
arterial vasoconstriction and platelet reactivity, which contribute to platelet thrombus
formation.
5. Consistency with other studies: Varying types of studies in other populations must observe
similar associations. Numerous studies using different designs have repeatedly supported
the relationship between smoking and heart disease.
6. Specificity: The exposure variable must be necessary and sufficient to cause disease; there is
only one causal factor. Although specificity may be strong causal evidence in the case of
infectious disease, this criterion is less important today. Diseases do not have single causes;
they have multifactorial origins.
The exposure variable of smoking is one of several risk factors for heart disease. Few factors are
linked to a single condition. Furthermore, smoking is not specific to heart disease alone. It is a
causal factor for other diseases such as lung and oral cancers. Additionally, smoking is not
“necessary and sufficient” to the development of heart disease, because there are nonsmokers who
also have coronary heart disease. Therefore the causal criterion of specificity more frequently
pertains to infectious diseases.
Although these criteria are useful in evaluating epidemiological evidence, it is important to note
that causality is largely a matter of judgment. In reality, absolute causality is only rarely established.
Rather, epidemiologists more commonly refer to suggested causal and associated factors. The effect
of confounding variables makes it difficult to ascertain true relationships between the exposure and
outcome variables. Confounding variables must be independently related to both the dependent
variable and the independent variable. Therefore these third variables can distort the true
relationship between the dependent and independent variables. For example, researchers
frequently control for the confounding effect of exercise and age when they examine the
relationship between diet and coronary heart disease. This is because persons with heart disease
may exercise less and be older than those without heart disease; that is, there is both an
independent relationship between exercise and heart disease and also between age and heart
disease. Without controlling for these factors, it is not possible to know if the association between
dietary fat and heart disease is real or if it can be attributed to differences in physical activity and
age.
By measuring the confounding variable, the researcher can statistically account for its effect in the
analysis (e.g., by using multiple logistical regression analysis or stratification). A biostatistics text
contains a discussion of these methods. Alternatively, matching subjects in treatment and control
groups with respect to the confounding variable minimizes the effects of the confounder. Again,
standardization for variables such as age is another method for managing spurious associations,
which makes true relationships more apparent. An understanding of such relationships facilitates
the practitioner’s interpretation and application of findings.
Screening
As explained previously, a central aim of epidemiology is to describe the course of disease
according to person, place, and time. Observations of the disease process may suggest factors that
aggravate or ameliorate its progress. This information also assists in determining effective treatment
and rehabilitation options (i.e., secondary or tertiary prevention approaches).
The purpose of screening is to identify risk factors and diseases in their earliest stages. Screening
is usually a secondary prevention activity because indications of disease appear after a pathological
change has occurred. In all forms of secondary and tertiary prevention, the identification of illness
prompts the nurse to consider which forms of upstream prevention could have interrupted disease
development.
Community health nurses commonly conduct screening programs. Community health nurses
may devote a large portion of their work activities to performing physical examinations; promoting
client self-examination; or conducting screening programs in schools, clinics, or community
settings. Although secondary prevention activities are important and provide vital information on
198
community health status, they focus on detecting existing disease. In contrast, primary prevention
and anticipatory guidance, which are hallmarks of community health nursing practice, attempt to
prevent the development of disease.
Community health nurses should consider several guidelines for screening programs. First,
nurses must plan and execute adequate and appropriate follow-up treatment for patients who test
positive for the disease. It is critical that nurses identify how to contact patients with positive
findings and where to refer them and then follow up with patients to determine whether they
accessed care. Health fairs have been criticized for the lack of consistent follow-up of screening
activities. Second, in the planning phase, the nurse should determine whether early disease
diagnosis constitutes a real benefit to clients in terms of improved life expectancy or quality of life.
Third, a critical prerequisite to screening is the existence of acceptable and medically sound
treatment and follow-up. In the past, public health providers debated the ethical and practical
arguments for implementing widespread HIV screening. Concern exists regarding the potential for
stigmatic consequences for and discrimination against those who screen positively for a test;
therefore people implementing screening programs should establish procedures for ensuring
confidentiality. These procedures, in conjunction with the development of effective antiviral
treatments, have encouraged earlier and more widespread identification of HIV-positive
individuals.
A screening program’s procedures must also be cost effective and acceptable to clients. Although
colonoscopy is a routine and effective screening procedure for colon cancer, it is neither simple nor
inexpensive. Although it is recommended periodically for all Americans with no known risk factors
beginning at age 50 years, less than 60% of that group have undergone the procedure (CDC, 2013b).
Annual screening using home fecal immunochemical tests (FIT) are not only cheaper but do not
require bowel preparation, anesthesia, or transportation. Therefore non-DNA FITs are suggested by
the U.S. Preventive Services Task Force as a more acceptable alternative to colonoscopy (Lin, et.al.,
2016). In short, a nurse should consider whether or not to screen a population on the basis of the
significant costs of screening programs and procedures, follow-up for clients who test positive, and
subsequent medical care (Box 5.5).
When developing a screening program, the community health nurse also must evaluate issues
related to the validity of the screening test. Detecting clients with disease is the purpose of
screening, and sensitivity is the test’s ability to do so correctly. Conversely, specificity is the extent to
which a test can correctly identify those who do not have disease. To obtain estimates of these two
dimensions, the nurse must compare screening results with the best available diagnostic procedure.
For a given test, the sensitivity and specificity tend to be inversely related to each other. When a test
is highly sensitive, individuals without disease may be incorrectly labeled as testing positive. These
false-positive results may cause stress and worry for clients and require further expense in the form
of testing to confirm a diagnosis. With a highly sensitive test, specificity may be lower and the test
may identify people as having the disease who are in fact disease free (i.e., more false-positive
results). If the sensitivity is low (and the specificity high), more patients who have the disease will
have negative test results. These patients will not be diagnosed and thus presumably will receive
care later in the disease process.
BOX 5.5 Guidelines For Screening Programs
• Screen for conditions in which early detection and treatment can improve disease outcome
and quality of life.
• Screen populations that have risk factors or are more susceptible to the disease.
• Select a screening method that is simple, safe, inexpensive to administer, acceptable to clients,
and has acceptable sensitivity and specificity.
• Plan for the timely referral and follow-up of clients with positive results.
• Identify referral sources that are appropriate, cost effective, and convenient for clients.
• Refer to evidenced-based screening recommendations published by the U.S. Preventive
Services Taskforce
(http://www.uspreventiveservicestaskforce.org/Page/Name/recommendations) and other
199
organizations.
Optimally, a screening test should be maximally sensitive and specific. To a large extent, this
depends on the precision of the test and the stringency of the cutoff point established for
determining a positive result. In the past, for example, the tuberculosis skin testing criterion was
based exclusively on a skin reaction of 10 mm of induration. As a result, some persons with the
disease were missed because their skin reactions were less than 10 mm (false-negative results), and
some with 10 mm of reaction were identified as having the disease but further testing showed they
did not have it (false-positive results). More recently, risk factors have been considered in addition
to induration, creating a more sensitive and specific TB screening algorithm. Currently, high-risk
individuals, such as those with HIV disease, are considered to have a positive skin test result with
5 mm of induration, which is more sensitive than the criteria of 15 mm set for those of low risk. Use
of the more stringent 5-mm criterion among those with HIV will lead to fewer false-negative
results, and use of the 15-mm cut point among low-risk individuals will lead to fewer false-positive
results (CDC, 2016). Box 5.6 shows the formula for calculating sensitivity and specificity.
Sensitivity and specificity reflect the yield of a screening test, which is the amount of detected
disease. One measure of yield is the positive predictive value of a test, which is the proportion of truepositive results relative to all positive test results. On the basis of Box 5.6, the formula is
. The
positive predictive value depends on the prevalence of undetected disease in a population.
Screening for a rare disease such as phenylketonuria will yield a lower predictive value and more
false-positive results. In phenylketonuria, a low predictive value is acceptable, because the falsenegative result has very serious consequences. The predictive value is also affected by the nature of
the screened population. Screening only the individuals at high risk for a disease will produce a
higher predictive value and can be a more efficient way to identify those with health problems. For
example, diabetes screening in an American Indian, Mexican American, or African American adult
population should produce a higher predictive value than screening the general adult population.
B O X 5 . 6 S e n s i t i v i t y a n d S p e c i f i c i t y o f a S c r e e n i n g Te s t
Screening Test Result
Positive
Negative
Those With Disease
True positives (a)
False negatives (c)
Surveillance
200
Those Without Disease
False positives (b)
True negatives (d)
In addition to screening, surveillance is a mechanism for the ongoing collection of community
health information. Monitoring for changes in disease frequency is essential to effective and
responsive public health programs. Identifying trends in disease incidence or identifying risk factor
status by location and population subgroup over time allows the community health nurse to
evaluate the effectiveness of existing programs and to implement interventions targeted to high-risk
groups. Again, identifying new cases for calculating incidence rates is particularly useful in
evaluating morbidity trends. However, this form of surveillance data is more difficult to collect, and
public health practitioners can access the data only for selected diseases. Prevalence rates, mortality
data, risk factor data, and hospital and health service data can help indicate a program’s successes
or deficiencies.
The Centers for Disease Control and Prevention (CDC) coordinates a system of data collection
among federal, state, and local agencies. These groups compile numerous sets of data and base
some of these data sets on the entire population (e.g., vital statistics data) and other collections on
subsamples of the population (e.g., the National Health Interview Survey). The completeness of
data reporting is variable because not all diseases are reportable. For example, practitioners are
required to report only four sexually transmitted infections (i.e., HIV/AIDS, syphilis, gonorrhea,
and chlamydia) to local and state health departments. Furthermore, not all practitioners report
cases on a regular basis, and not all people with sexually transmitted infections actually seek care.
Studies have indicated that practitioners also underreport childhood communicable diseases, such
as chickenpox and mumps. The CDC conducts studies that estimate the magnitude of this
underreporting problem.
Practitioners have a continuing need for comprehensive and systematically collected surveillance
data that describe the health status of national and local subgroups. They use this information to
evaluate the impact of programs on specific groups in a community.
Healthy People 2020
Objectives for Data Collection and Reporting
PHI HP2020–7: Increase the proportion of population-based Healthy People 2020 objectives for
which national data are available for all major population groups.
PHI HP2020–8: Increase the proportion of Healthy People 2020 objectives that are tracked
regularly at the national level.
PHI HP2020–9: Increase the proportion of Healthy People 2020 objectives for which national
data are released within one year of the end of data collection.
PHI HP2020–10: Increase the percentage of vital events (births, deaths, fetal deaths) reported
using the latest U.S. standard certificates of birth and death and the report of fetal death.
From U.S. Department of Health and Human Services: Healthy People 2020: public health infrastructure,
n.d. Retrieved from: https://www.healthypeople.gov/2020/topics-objectives/topic/public-healthinfrastructure/objectives.
For example, the effectiveness of Healthy People 2020 depends on the availability of reliable
baseline and continuing data to characterize health problems and evaluate goal achievement as
listed in the Healthy People 2020 box. Healthy People 2020 addresses the ongoing need to extend the
inclusiveness of such data collection systems (U.S. Department of Health and Human Services
[USDHHS], n.d.). For example, simply documenting children’s mortality rates resulting from injury
is insufficient for the development of specific methods of injury prevention. Data on the number of
injured children and the nature of injury (e.g., motor vehicle accidents, drowning, abuse) across the
nation would increase the usefulness of surveillance information. The Health Indicators Warehouse
has compiled indicator data for initiatives like Healthy People and the Center for Medicare Services
so that health status and service quality can be monitored through April 2017 (National Center for
Health Statistics, n.d.). Databases such as Health, United States, Healthy People, and Health Status
Indicators available at the CDC webpage will be the primary source of evaluative information.
Nurses need to describe trends in health and illness according to a community’s locale,
demographics, and risk factor status to intervene effectively on behalf of the people. They must
201
compare the data for their locale with those of a relevant neighboring area (e.g., a census tract, city,
county, state, or nation) to gain perspective on the magnitude of a local problem (see Clinical
Example 5.1). Ideally, the nurse should have access to surveillance data at several different levels
over a given period. In some instances, community health nurses find it necessary to construct their
own surveillance systems that are tailored to specific health conditions or available programs in
their community. These smaller data collection systems help nurses evaluate programs when the
data are readily accessible and are compatible with data from large city or statewide surveillance
systems.
Clinical Example 5.1
In 1991, a cluster of neural tube defects (NTDs) occurred in babies born in Brownsville, Texas,
within a span of 6 weeks. Investigation indicated a rate of 27.1 cases per 10,000 live births, in
contrast to the U.S. rate of approximately 8 per 10,000 (Texas Department of Health, unpublished
report, 1992). The Brownsville rate was more than three times the national rate and represented an
increased risk in Hispanic women. This increased risk was partially attributable to cultural and
environmental factors, including lower socioeconomic status and migrant farm work. The
investigators implemented a surveillance program that obtained more accurate population-based
data. Additionally, the program implemented folic acid supplementation in Texas counties along
the Mexican border. NTD rates dropped to 13 per 10,000 after the supplementation effort.
Research suggests that 50% to 70% of NTDs may be preventable with folic acid supplementation.
This finding supports the fortification of bread and cereal products; in January 1998, the U.S. Food
and Drug Administration (FDA) mandated the addition of 140 µg of vitamin B per 100 g of most
grain products. It is estimated that there has been a 24% reduction in the number of NTDs since
grain fortification with folic acid began.
Dietary intake alone may be insufficient; also, the greatest risk to the fetus occurs within the first
3 to 8 weeks of pregnancy, a time when many women do not yet recognize their pregnancies.
Therefore the CDC and U.S. Preventive Services Task Force recommend that all women of
reproductive age consume 400 µg (0.4 mg) to 800 µg (0.8 mg) daily of synthetic folic acid in
addition to dietary sources such as cereal or grain products, leafy green vegetables, and vitamin
supplements.
From U.S. Preventive Services Task Force: Folic acid to prevent neural tube defects: preventive
medication, 2009. Retrieved from:
http://www.uspreventiveservicestaskforce.org/uspstf09/folicacid/folicacidrs.htm.
Research Highlights
Reducing Infant Mortality Rates Using the Perinatal Periods of Risk Model
The infant mortality rate is an accepted indicator for measuring a nation’s health. The rate is
representative of the health status and social well-being of any nation. Despite decreases in the past
50 years, infant mortality rates in the United States remain higher than in other industrialized
countries. Using overall infant mortality rates to determine the effectiveness of interventions does
not help communities focus on particular underlying factors contributing to the rates. Targeting
interventions to the factors most responsible for the infant mortality rate should help reduce the
rate more rapidly and effectively.
The Perinatal Periods of Risk (PPOR) model was developed to provide direction, focus, and
suggestions for effective interventions. The model helps users identify and rank four factors as they
contribute to the overall infant mortality rate: (1) mother’s health before and between pregnancies,
(2) maternal health care systems, (3) neonatal health care systems, and (4) infant health during the
first year of life. The PPOR model is based on two major theoretical constructs: age of fetus-infant
at death and birth weight. The PPOR model maps each death in a geographic region on the basis of
birth weight and age at death, including fetal, neonatal, and postneonatal periods. The lowest birth
weight infant deaths are combined into one cell named the maternal health cell. The three
remaining groups are put into cells suggesting the primary preventive focus for that group:
maternal health, newborn health, and infant health. Multiple interventions are important in
202
reducing infant mortality, and the PPOR model guides prioritizing interventions based on the cell
contributing the most to infant mortality rates (Peck et al., 2010).
The PPOR model has been used in several programs and research studies to improve mother
and infant health. In one intervention study (Chao et al., 2010) the PPOR model was successfully
used to improve birth outcomes in very-high-risk populations in the Antelope Valley area of Los
Angeles County. On the basis of assessment findings per PPOR directives, efforts were made to
infuse resources into the community and expand case management initiatives for high-risk
mothers. Long-term findings indicated that the PPOR model was useful for identifying risk and
social factors and that it helped mobilize community partnerships that resulted in widely
improved birth outcomes.
Data from Chao SM, Donatoni G, Bemis C., et al.: Integrated approaches to improve birth outcomes:
perinatal periods of risk, infant mortality review and the Los Angeles mommy and baby project,
Maternal and Child Health Journal 14:827–837, 2010; Peck MG, Sappenfield WM, Skala J: Perinatal
periods of risk: a community approach for using data to improve women and infants’ health,
Maternal and Child Health Journal 14:864–874, 2010.
As stated, epidemiologists describe the course of disease over time. These secular trends are
changes that occur over years or decades, such as the fairly recent, significant decline in lung cancer
deaths in men and the gradual increase in lung cancer deaths in women. Frequently,
epidemiologists document the associated patterns of treatment and intervention. In many instances,
studies conducted by clinical epidemiologists provide this information. Cancer registries are a form
of surveillance that document the prevalence and incidence of cancer in a community and
document its course, treatment, and associated survival rates. The Surveillance, Epidemiology and
End Results (SEER) program of the National Cancer Institute compiles national cancer data from
existing cancer registries covering approximately 28% of the U.S. population (National Cancer
Institute, 2017).
Public health practitioners need to conduct community surveys of population segments to plan
for the segments’ health. For example, a survey of the disabled population that assesses prevalence
may also evaluate the adequacy of current services and project future needs.
203
Use of Epidemiology in Health Services
Epidemiological approaches, such as the ones presented here, can be used to describe the
distribution of disease and its determinants in populations. However, epidemiological principles
are also useful in studying population health care delivery and in describing and evaluating the use
of community health services. For example, analyzing the ratio of health care providers to
population size helps determine the system’s ability to provide care. The clients’ reasons for seeking
care, the clients’ payment methods, and the clients’ satisfaction with care are also informative.
Regardless of whether community health nurses or other health services professionals collect these
data, the information is essential for those who strive to improve clients’ access to quality health
care.
Health services epidemiology focuses on the population’s health care patterns. In particular,
public health practitioners are concerned with the accessibility and affordability of services and the
barriers that may contribute to excess morbidity in at-risk groups. Traditionally, children are a
vulnerable group, and they are a particular focus of health services research. Studies examining
poverty rates and care access have underscored the need to expand insurance coverage to those
who do not have private medical insurance and do not qualify for Medicaid programs or the State
Children’s Health Insurance Program.
Ultimately, nurses must apply epidemiological findings in practice. It is essential that they
incorporate study results into prevention programs for communities and at-risk populations.
Furthermore, the philosophy of public health and epidemiology dictates that nurses extend their
application into major health policy decisions, because the aim of health policy planning is to
achieve positive health goals and outcomes for improved population health.
A goal of policy development is to bring about desirable social changes. Epidemiological factors,
history, politics, economics, culture, and technology influence policy development. The complex
interaction of these factors may explain the challenges with application of epidemiological
knowledge. Lung disease in the United States exemplifies the incomplete progress in implementing
effective health policy. In the early 1950s, studies identified and conclusively linked cigarette
smoking to lung cancer and heart disease (Doll and Hill, 1952). Beginning in the 1950s, public
policies to address this health threat have included cigarette taxes, cigarette package warning labels,
smoking restrictions in public areas, the institution of smoke-free workplaces, and restrictions on
selling tobacco to minors. Despite the successes of the past 60 years, approximately 20% of
Americans continue to smoke, with rates particularly high among young adults, suggesting a
continued need for focused and effective public health policy. Community health nurses should
exercise “social responsibility” in applying epidemiological findings, but doing so will require the
active involvement of the consumer. Community health nurses collaborating with community
members can combine epidemiological knowledge and aggregate-level strategies to effect change
on the broadest scale.
Active Learning Exercise
Consult Healthy People 2020 to find the national goals for selected causes of morbidity and
mortality. Identify groups at an increased risk for these selected diseases. What are the approaches
suggested by these documents for reducing the rates of disease? How can this information be
useful in planning for your community?
204
Epidemiological Methods
Two epidemiological methods—descriptive epidemiology and analytic epidemiology—are used by
community health nurses. This section describes both and gives examples of how they are used in
population health.
Descriptive Epidemiology
Descriptive epidemiology focuses on the amount and distribution of health and health problems
within a population. Its purpose is to describe the characteristics of both people who are protected
from disease and those who have a disease. Factors of particular interest are age, sex, ethnicity or
race, socioeconomic status, occupation, and family status. Epidemiologists use morbidity and
mortality rates to describe the extent of disease and to determine the risk factors that make certain
groups more prone to acquiring disease.
In addition to “person” characteristics, the place of occurrence describes disease frequency. For
example, certain parasitic diseases, such as malaria and schistosomiasis, occur in tropical areas.
Other diseases may occur in certain geopolitical entities. For example, gastroenteritis outbreaks
often occur in communities with lax water quality standards. Time is the third parameter that helps
define disease patterns. Epidemiologists may track incidence rates over a period of days or weeks
(e.g., epidemics of infectious disease) or over an extended period of years (e.g., secular trends in the
cancer death rate).
These person, place, and time factors can form a framework for disease analysis and may suggest
variables associated with high versus low disease rates. Descriptive epidemiology can then generate
hypotheses about the cause of disease, and analytic epidemiology approaches can test these
hypotheses (Box 5.7).
B O X 5 . 7 A n E x a m p l e o f D e s c r i p t i ve E p i d e m i o l o g y
The person–place–time model is illustrated by two measles outbreaks in Utah:
• Person: Initially, an unvaccinated 15-year-old student had contracted measles, likely from a
trip to Europe. Subsequently, six more students contracted the illness.
• Place: Salt Lake County; three cases were “school transmission,” and three cases were
“household transmission.”
• Time: The index case traveled to Europe during March 3–17, 2011. He attended school on
March 21 and subsequently became ill. The other cases occurred between April 5 and April 17,
2011.
From Centers for Disease Control and Prevention: Two measles outbreaks after importation—Utah,
March–June 2011, Morbidity and Mortality Weekly Report 62(12):222–225, 2013. Retrieved from:
http://www.cdc.gov/mmwr/pdf/wk/mm6212.pdf.
205
FIG. 5.5
Cross-sectional, or prevalence, study.
Analytic Epidemiology
Analytic epidemiology investigates the causes of disease by determining why a disease rate is
lower in one population group than in another. This method tests hypotheses generated from
descriptive data and either accepts or rejects them on the basis of analytic research. The
epidemiologist seeks to establish a cause-and-effect relationship between a preexisting condition or
event and the disease (see previous section on causality). To determine this relationship, the
epidemiologist may undertake two major types of research studies: observational and experimental.
Observational Studies
Epidemiologists frequently use observational studies for descriptive purposes, but they also use them
to discover the etiology of disease. The investigator can begin to understand the factors that
contribute to disease by observing disease rates in groups of people differentiated by experience or
exposure. For example, differences in disease rates may occur in the obese compared with the
nonobese, in smokers compared with nonsmokers, and in those with high stress levels compared
with those with low stress levels. These characteristics (i.e., obesity, smoking, and stress) are called
exposure variables.
Unlike experimental studies, observational studies do not allow the investigator to manipulate
the specific exposure or experience or to control or limit the effects of other extraneous factors that
may influence disease development. For example, life stress is related to depression. People with
low socioeconomic status also have high depression rates. People with low socioeconomic status
frequently experience greater life stresses; therefore the confounding factor of socioeconomic status
makes it more difficult to demonstrate the effect of stress on depression. The three major study
designs used in observational research are cross-sectional, retrospective, and prospective studies.
Cross-Sectional Studies
Cross-sectional studies, sometimes called prevalence or correlational studies, examine relationships
between potential causal factors and disease at a specific time (Fig. 5.5). Surveys that
simultaneously collect information about risk factors and disease exemplify this design. For
example, the National Health and Nutrition Examination Survey (NHANES) has collected crosssectional data regarding current dietary practices, physical status, and health in adults and children
in the United States since the early 1960s (CDC, 2017). Data from the NHANES studies have been
analyzed and compared over the years by a number of researchers and have provided important
health information.
For example, NHANES studies have tracked contemporary behavior issues such as unsweetened
206
beverage consumption in U.S. youth (Kit et al., 2013) and fast food in adults, showing recent
declines in the total daily calories consumed from these sources. Although overall fast-food
consumption has declined from approximately 13% of daily caloric intake to 11.3% in the period
ending 2010, this decline was not shared by all groups, with non-Hispanic African Americans
consuming more fast food than other groups, as did all persons who were overweight (Fryar and
Ervin, 2013). NHANES has conducted interviews and physical examinations on youth, nutrient
studies on children, and dietary surveys of older Americans, contributing important data that
suggest risk factors that can be examined through more rigorous study designs.
Although a cross-sectional study can identify associations among disease and specific factors, it is
impossible to make causal inferences because the study cannot establish the temporal sequence of
events (i.e., the cause preceded the effect). For example, the NHANES was unable to determine
whether high salt intake precedes hypertension—thus making it a causal factor—or whether they
are unrelated. Therefore cross-sectional studies have limitations in discovering etiological factors of
disease. These studies can help identify preliminary relationships that other analytic designs may
explore further; therefore they are hypothesis-generating studies.
207
FIG. 5.6
Retrospective, or case-control, study.
Retrospective Studies
Retrospective studies compare individuals with a particular condition or disease and those who do
not have the disease. These studies determine whether cases, or a diseased group, differ in their
exposure to a specific factor or characteristic relative to controls, or a nondiseased group. To make
unambiguous comparisons, investigators select the cases according to explicitly defined criteria
regarding the type of case and the stage of disease. Investigators also select a control group from the
general population that is characteristically similar to the cases (Fig. 5.6).
Frequently, people hospitalized for diseases that are not under study become controls if they do
not share the exposure or risk factor under study. For example, a researcher may select patients
with heart disease to be controls in a study of patients with lung cancer. However, this choice may
introduce serious bias because the two groups often share the risk factor of smoking. The methods
of data collection must be the same for both groups to prevent further introduction of bias into the
study. Therefore it is desirable for interviewers to remain unaware of which subjects are cases and
which are controls.
In retrospective studies, data collection extends back in time to determine previous exposure or
risk factors. Investigators analyze study data by comparing the proportion of subjects with disease,
or cases, who possess the exposure or risk factors with the corresponding proportion in the control
group. A greater proportion of exposed cases than of exposed controls suggests a relationship
between the disease and the risk factor.
Investigators often use retrospective study designs because these designs address the question of
causality better than cross-sectional studies. Retrospective studies also require fewer resources and
less data collection time than prospective studies. Many examples of retrospective, or case-control,
studies exist in the literature. One classic example is Doll and Hill’s (1952) investigation of risk
factors for lung cancer. They compared exposure rates for cases in which lung cancer was
diagnosed with those in the control group, in whom cancer was diagnosed outside the chest and
208
oral cavity. The researchers recorded detailed smoking histories in all subjects. Compared with the
controls, a significantly higher proportion of patients with lung cancer smoked. This study yielded
the hypothesis that smoking may be etiologically related to lung cancer.
Prospective Studies
Prospective studies monitor a group of disease-free individuals to determine whether and when
disease occurs (Fig. 5.7). These individuals, or the cohort, have a common experience within a
defined period. For example, a birth cohort consists of all people born within a given period. The
study assesses the cohort with respect to an exposure factor associated with the disease and thus
classifies it at the beginning of the study. The study then monitors the cohort for disease
development. The investigator compares the disease rates for those with a known exposure and the
disease rates for those who remain unexposed. The study observes subjects prospectively; therefore
it summarizes data collected over time by the incidence rates of new cases (Box 5.8). Again,
comparing two incidence rates produces a measure of relative risk:
209
FIG. 5.7
Prospective, or cohort, study.
B O X 5 . 8 C o m p a r i s o n o f T i m e F a c t o r s i n R e t r o s p e c t i ve a n d P r o s p e c t i ve
Study Designs
Cohort Study
Girls with bacteriuria → Women with renal disease
Girls with sterile urine → Women without renal disease
Case-Control Study
Girls with bacteriuria ← Women with renal disease
Girls with sterile urine ← Women without renal disease
PAST—————PRESENT (BEGINNING)—————FUTURE
Comparison of time factors in prospective design (i.e., cohort) and retrospective design (i.e.,
case-control) approaches to studying the possible effect of childhood bacteriuria on renal
disease in adult women.
The relative risk indicates the extent of excess risk incurred by exposure relative to nonexposure.
A relative risk of 1 suggests no excess risk resulting from exposure, whereas a relative risk of 2
suggests twice the risk of having disease from exposure.
210
Prospective studies, or longitudinal, cohort, or incidence studies, are advantageous because they
obtain more reliable information about the cause of disease than do other study methodologies.
These studies establish a stronger temporal relationship between the presumed causal factors and
the effect than do retrospective and cross-sectional studies. Calculations of incidence rates and
relative risks provide a valuable indicator of the level of risk that exposure creates.
Ethical Insights
The Tuskegee Syphilis Study
In 1932 the U.S. Public Health Service (PHS) began a longitudinal-experimental study of 600
African American sharecroppers, 399 of whom had syphilis and 201 who did not. The study was
conducted in one of the poorest counties of Alabama, and the subjects were unaware that they had
syphilis; they were told they were being treated for “bad blood.” Enticed by the promise of free
medical care and meals, the subjects joined the study without knowledge of their disease, its
treatment, or the study procedures. The experimental group was initially treated with ineffective
doses of the treatments of the time—bismuth or mercury—and later with aspirin. Even when
penicillin became available in the late 1940s, these subjects were actively denied treatment. For 40
years, these men were followed up by PHS investigators affiliated with the Tuskegee Institute and
hospital, who claimed to be observing the differences in the progression of the disease in blacks in
comparison with the control group. During the course of the study, many subjects died of syphilis
or other causes, numerous wives became infected, and children were born with congenital syphilis.
In 1972, a former venereal disease interviewer, Peter Buxtun, “blew the whistle” on the study,
and reports were published in newspapers. Only after the public became outraged about the
unethical nature of the study did the CDC and the PHS move to end it. In 1973, the National
Association for the Advancement of Colored People won a $10 million class action suit on behalf of
the subjects. In 1997, President Bill Clinton formally apologized to the few survivors and their
families for the harm inflicted on these men and their families in the name of public health
research.
The Tuskegee Study raises questions about how a study could proceed without informing and
seeking consent of participants, how available treatment could be withheld, and how government
researchers could pursue an unethical research plan without periodic review and questioning.
Furthermore, the racial and discriminatory issues suggest disturbing questions for researchers and
practicing nurses to contemplate, one being that the Tuskegee Study contributes to a legacy of
distrust that minorities may harbor toward both the health care delivery system and research
programs.
Data from Centers for Disease Control and Prevention: The Tuskegee timeline, 2013. Retrieved from:
http://www.cdc.gov/tuskegee/timeline.htm; Infoplease: The Tuskegee syphilis experiment, 2005,
Pearson Education. Retrieved from: www.infoplease.com/ipa/A0762136.html.
However, certain disadvantages are inherent in the prospective design. It is costly in terms of
resources and staff to monitor a cohort over time, and lengthy studies result in subject attrition.
Problems arising from the nature of chronic diseases may compound these logistical dilemmas.
Frequently, chronic diseases have long latency periods between exposure and symptom
manifestation. Furthermore, the onset of chronic conditions may be insidious, making it extremely
difficult to document the incidence of disease. In addition, many diseases do not have a unifactorial
cause (i.e., single variable) because many interacting factors influence disease. These problems do
not negate the benefits of prospectively designed epidemiological studies; rather, they suggest a
need to carefully plan and tailor a study specifically to the disease and the study’s purpose.
The literature contains numerous prospective studies. In many cases, these studies have been
instrumental in substantiating causal links between specific risk factors and disease. A classic
example is a Doll and Hill cohort study of subjects who eventually developed lung cancer during
the follow-up period (1956). Doll and Hill originally completed questionnaires on a cohort of
physicians in Great Britain. Next, they classified the subjects according to several variables,
emphasizing the number of cigarettes smoked. In 4½ years, they accessed death certificate data.
These data revealed a higher mortality rate resulting from lung cancer and coronary thrombosis
211
among smoking physicians compared with nonsmokers. The death rate for heavy smokers was 166
per 100,000 versus 7 per 100,000 for nonsmokers. Combining these two incidence rates in a measure
of excess risk indicated that heavy smokers were 23.7 times more likely to develop lung cancer than
nonsmokers:
These findings in a prospective study provided strong epidemiological support for smoking as a
risk factor for lung cancer.
Another well-known prospective study is the Framingham Heart Study, which has followed an
essentially healthy cohort of Framingham, Massachusetts residents for more than 50 years. Findings
from the study suggested that serum cholesterol level and other risk factors are associated with the
future development of cardiovascular disease (Kramarow et al., 2013). The Framingham Study and
subsequent “offspring studies” helped form the basis for later experimental studies aimed at
reducing serum cholesterol through diet modification or drug therapy to ultimately lower the
incidence rate of coronary heart disease.
The Nurses’ Health Study was initiated in 1976 with 121,700 registered nurses, with the intent of
examining the long-term consequences of oral contraceptives. The initial cohort still returns
questionnaires every 2 years, and data have been collected on diet and nutrition, smoking, hormone
use, and menopause as well as various chronic illnesses. In 1989 the Nurses’ Health Study II was
initiated to study lifestyle issues, contraception, and illness patterns in younger women, and in 2008
a third study was begun looking at similar issues in another cohort.
These studies continue to monitor nurses’ changing health status and risk factors and to examine
factors associated with the development of numerous health conditions in women, such as breast
cancer and heart disease (Nurses’ Health Study, 2017). For example, research using Nurses’ Health
Study data has determined that regular use of nonsteroidal antiinflammatory drugs (i.e.,
acetaminophen) does not reduce the incidence of skin cancer (Jeter et al., 2012) or breast cancer
(Eliassen et al., 2009). Similarly, research from the nurses’ health studies have shown that although
consumption of sugar-sweetened beverages is associated with higher risk of type 2 diabetes,
caffeine intake lowers the risk (Bhupathiraju et al., 2013). Last, the Nurses’ Health Study suggested
the deleterious effects of exposure to foods containing trans-fats. Groundbreaking research found a
50% increase in cardiovascular disease risk for women who consumed the highest trans-fat intakes
(Willet et al., 1993). This initial finding set the stage for sweeping policy changes in FDA labeling of
these fats on Nutrition Facts boxes on food packages, on restaurant menus in many cities, and an
ultimate 2015 FDA determination that partially hydrogenated oils are not “Generally Recognized as
Safe” (Curtis et al., 2016). These changes have led to a reduction in population levels of serum
cholesterol.
Experimental Studies
Another type of analytic study is the experimental design, called the randomized clinical trial (Fig.
5.8). Epidemiological investigations apply experimental methods to test treatment and prevention
strategies. The investigator randomly assigns subjects at risk for a particular disease to an
experimental or a control group. The investigator observes both groups for the occurrence of
disease over time, but only the experimental group receives intervention, although often the control
group receives a placebo. The primary statistical analysis is based on “intention to treat,” that is, all
subjects remain assigned to the original treatment group, regardless of whether subjects may have
decided on their own to discontinue or change their therapy. For example, if a subject in a drug trial
who is assigned to the active medication experiences side effects possibly from this medication and
therefore discontinues the medication, this subject still is considered to be within the active drug
group for the purpose of statistical testing. The change in category from treatment to no treatment,
or vice versa, is called a crossover and may decrease the likelihood of finding a significant effect for
212
the active treatment.
Theoretically, it is possible to introduce a harmful exposure or risk factor as the experimental
factor; however, ethical considerations usually prohibit the use of human subjects for these
purposes. For example, it is unacceptable to require an experimental group to smoke cigarettes in
an experiment; therefore the investigator uses case-control or cohort epidemiological designs. This
limitation usually restricts experimental epidemiological studies to prophylactic and therapeutic
clinical trials. Experimental studies testing vaccines and medications for safety and efficacy are
examples.
The experimental design is also useful for investigating chronic disease prevention. Thus
experimental studies may help evaluate community health nursing interventions. For example, they
may help determine the effectiveness of a sex education program in preventing high rates of
teenage pregnancy or the feasibility of an AIDS prevention program among intravenous drug users.
Randomized trials were used to evaluate the Nurse-Family Partnership program, which established
the long-term positive effects nurse home visits had on high-risk pregnant women and their
children in comparison with those who did not receive home visits (Olds et al., 2014).
FIG. 5.8
case Study
Experimental study, or clinical trial.
Application of the Nursing Process
Using an Epidemiological and Public Health Approach to Managing a Foodborne
Outbreak
Nurses working in schools, daycare centers, camps, and other facilities where food is served must
be cognizant of safe food-handling principles. Furthermore, they must be aware of the potential for
transmitting disease if proper procedures are not followed. Outbreaks of foodborne illness must be
assessed and managed, and often it is the community health nurse who initiates and participates in
this process. The following is a scenario in which the nurse utilized the nursing process to analyze
and intervene in such an epidemic.
Assessment
On Wednesday, October 4, the school nurse at Greenly Elementary School saw eight students who
complained of abdominal cramping, diarrhea, and fever. Parents of the sick students were called,
213
and the students were sent home. On Thursday, the nurse was alerted to a large number of absent
students and teachers. Specifically, 62 students and 10 teachers were absent. Most reported
diarrhea symptoms. Because the absentee rate of 10% exceeded the average daily rate of 4% for the
620-student school and because the nurse determined that the large number of diarrhea cases
suggested an epidemic, the local public health department was notified.
Public health officials arrived at the school and began to assess students still at school and those
who were recovering at home. Stool culture specimens were collected and sent to the state
laboratory. Results indicated that the organism causing illness was in most cases Shigella sonnei, the
most commonly found form of the bacteria. Persons with severe symptoms were referred to their
physicians for possible antibiotic therapy. Food histories of meals eaten both at school and outside
of school were taken.
Friday saw a continuing increase in absenteeism of students and staff reporting gastrointestinal
illness. Public health specialists defined the criteria for identifying cases on the basis primarily of
positive laboratory results, symptoms of diarrhea or vomiting, fever with nausea or abdominal
pain, or all of these. Cafeteria staff were interviewed, and it was determined that one staff member
had had diarrhea over the previous weekend but had returned to work on Monday. Public health
staff continued to take dietary histories of affected and unaffected persons and constructed rates of
illness for all foods served in the cafeteria beginning on Friday of the previous week. These data are
displayed in the following table.
From the data, it can be seen that students who ate lunch at school on Tuesday and ate fajitas
and salad had higher rates of illness than those who did not. Therefore it was concluded that the
outbreak of Shigella could be attributed to a food source.
Diagnosis
Determining the likely cause of the outbreak was important in specifying a diagnosis and directing
the planning of an intervention. The following diagnosis was formulated:
Increased risk for infectious diarrhea among elementary school children related to inadequate
hygiene and food-handling practices as evidenced by a 19% increase in reported cases within a 4day period.
Planning
The school nurse, in conjunction with public health specialists, determined that several groups
should be targeted in order to eliminate the further spread of disease. They identified a need to
assist families in understanding the nature of the disease, how to care for their children who were
ill, and how to prevent the spread at home. Within the school, there was a need to review foodhandling practices and the training that cafeteria workers received. Staff, including teachers, also
required information about Shigella and how it should be prevented in the everyday lives of
students. Needs of special ages and developmental levels of children were also important. A
formal plan of what needed to be done, by whom, and when was drawn up. Research into the
nature and prevention of Shigella was gathered from the CDC and the local health department,
among other sources. Health department staff developed a plan to release information to the public
about the prevention of gastrointestinal illnesses, as many of these diseases are easily spread and
so many students were already ill.
Long-Term Goal
• An absence of cases of infectious diarrhea
Short-Term Goals
• Treatment and recovery of all identified cases of diarrhea
• Implementation of an effective program of hygienic practices among students and staff
• Implementation of a food-handling program for all cafeteria workers
• Adequate informing of the larger community in order to prevent spread of the epidemic
Intervention
The school nurse took a central leadership role, directing action within the school aimed at staff,
214
students, and student families. Teaching of appropriate hand washing was stressed. Hand-washing
facilities were inspected for soap, paper towels, and running water. Food preparation guidelines
were reviewed with staff, and policies regarding remaining at home when ill were reiterated. The
health department staff provided technical assistance and made recommendations. They informed
community physicians about surveillance and reporting requirements and provided information
regarding case identification and treatment regimens. Daycare centers and preschools were
advised to watch for diarrhea outbreaks and to adhere to strict hand-washing and diaper-handling
practices, as these facilities tend to be high-risk areas for the transmission of organisms such as
Shigella. The media were contacted to elicit their help in disseminating correct and useful
information to the community.
Evaluation
Immediate evaluation involved monitoring the decline in Shigella cases both within the school and
in the larger community. The school nurse noted that rates of absenteeism returned to normal on
the following Monday. She determined that all classes had received hygiene instruction within the
following 2 weeks and that all teachers had received a flyer with specific information about
Shigella, its care, and its prevention. She observed that bathrooms had filled soap dispensers, that
friendly signs reminding students to wash hands were posted near sinks, and that students were
given the opportunity to wash hands before lunch and snacks. The public health department
likewise continued surveillance activities after encouraging physicians to collect and submit stool
culture specimens for suspected cases and to report cases to the health department. Rates of
diarrhea declined rapidly in the week after the school outbreak. The infection did not spread to
other schools or community groups. This outcome can be attributed to successful epidemic
management, yet surveillance remains critical if the public’s health is to be protected.
Levels of Prevention
Primary
• Teach students and staff about hand washing and hygienic practices.
• Maintain a system that promotes safe food-handling practices.
• Exclude those with symptoms from school or food handling.
Secondary
• Collect stool culture specimens from all symptomatic individuals.
• Treat those with advanced diarrhea symptoms with antibiotics.
• Exclude those with positive culture results from food handling, and exclude those with
symptoms from school.
• Advise families and individuals in the care of those with diarrhea.
Tertiary
• Treat and counsel those determined to be carriers of Shigella.
Information on Shigella infections is available at https://www.cdc.gov/shigella/index.html
Number Exposed by Meal and Food Item (N = 143)
215
∗ Odds ratios were calculated with the formula: ad/bd.
Modified from Texas Department of Health: Shigella outbreak in an elementary school, Dis Prev
News 55(6):1–3, 1995.
216
Summary
Epidemiology offers the community health nurse methods to quantify the extent of health problems
in the community and provides a body of knowledge about risk factors and their association with
disease. At each step of the nursing process, epidemiological applications support the practice of
the community health nurse. Compiling descriptive data from surveys or studies contributes to the
nurse’s understanding of the community’s health level. In assessing community problems,
epidemiological rates describe the magnitude of disease and provide support for community
diagnoses. Epidemiological studies suggest interventions and their potential efficacy—information
that is useful in planning prevention and intervention approaches. Evaluation studies using
epidemiological methods, either reported in the literature or conducted by community health
nurses, are essential for providing optimal research-based care.
217