Study Types in Epidemiology

Transcript

Study Types in Epidemiology

Welcome to “Study Types in Epidemiology.” My name
is John Kobayashi. I’m on the Clinical Faculty at the
Northwest Center for Public Health Practice, at the School
of Public Health and Community Medicine, University
of Washington in Seattle. From 1982 to 2001, I was the
state epidemiologist for Communicable Diseases at the
Washington State Department of Health. Since 2001, I’ve
also been the foreign adviser for the Field Epidemiology
Training Program of Japan.

About this Module

The modules in the epidemiology series from the
Northwest Center for Public Health Practice are intended
for people working in the field of public health who are
not epidemiologists, but who would like to increase their
understanding of the epidemiologic approach to health and
disease.
This module focuses on descriptive and analytic epide-
miology and their respective study designs. Before you go
on with this module, we recommend that you become
familiar with the following modules, which you can
find on the Center’s Web site: What is Epidemiology in
Public Health? and Data Interpretation for Public Health
Professionals.
We introduce a number of new terms in this module. If you want to review their
definitions, the glossary in the attachments link at the top of the screen may be
useful.

Objectives
By now, you should be familiar with the overall approach
of epidemiology, the use of various kinds of rates to
measure disease frequency, and the various ways in which
epidemiologic data can be presented.
This module offers an overview of descriptive and
analytic epidemiology and the types of studies used to
review and investigate disease occurrence and causes.
By the end of this module, you should be able to: list the
differences between descriptive and analytic epidemiology,
describe the main types of epidemiologic studies and their

Northwest Center for Public Health Practice 1

Study Types in Epidemiology
Transcript

uses, identify and provide examples of person, place, and time, and describe the
main differences among case-control, cohort studies, and experimental studies.

Goals of Epidemiologic Studies

Epidemiology is a very common and important activity in
public health departments. It clarifies clinical and demo-
graphic characteristics of diseases and conditions. It iden-
tifies who is at risk, and provides clues to the causes of
disease. And finally, it guides preventive measures and
interventions.
Before we go on, I want to emphasize an important
point about epidemiologic studies. Unlike medical staff,
who are concerned about the health of individuals, epide-
miologists focus on the distribution and determinants of
events and diseases in groups of people.

Epidemiologic Study Types

Epidemiologic activity is often divided into two types,
descriptive and analytic, and each of these types of epide-
miology uses specific kinds of studies.
Descriptive studies examine patterns of disease occur-
rence, with a focus on person, place, and time. These
studies use relatively accessible data for program plan-
ning, to estimate caseloads, to determine the amount
of public health resources needed, or to identify high
risk groups. Many public health practitioners carry out
descriptive epidemiologic studies within their jurisdictions.
Epidemiologists also use descriptive studies to generate
hypotheses that need to be confirmed or ruled out by
analytic studies.
For example, early descriptive studies found that the majority of newly diagnosed
AIDS cases in the United States were among young urban men, which led to the
hypothesis that certain types of sexual behavior might cause AIDS.
Analytic studies are used to test hypotheses such as this. They are not usually
planned or conducted at the local level, because these types of studies typi-
cally require more resources and knowledge than are available in a local agency.
However, it’s important that public health practitioners understand analytic study
types in order to interpret these studies and make suggestions based on their own
observations and experiences.

Northwest Center for Public Health Practice 2

Study Types in Epidemiology
Transcript

The Five Ws of Epidemiology Studies

Epidemiologic studies ask five questions: what, who,
where, when, and why. These five Ws remind us to orga-
nize questions about possible exposures (or risk factors).
What includes the diagnosis or clinical information.
Who, where, and when include person, place, and time
information.
Why involves causes of disease, risk factors, and modes
of transmission.
We study what, who, where, and when, or in other
words, clinical data plus person, place, and time informa-
tion, through descriptive epidemiology, which investigates
the distribution of diseases or conditions.
We study why and how, or in other words, causes, risk factors, and modes of
transmission, through analytic epidemiology, which investigates the determinants of
diseases or conditions.
Let’s look at the questions who, where, and when in a little more detail.

Describing What, or Clinical Information

Clinical information from a single or several ill persons has
often been extremely important in alerting epidemiolo-
gists to a new problem. With this information, public health
practitioners might look for other cases of the same type
or establish some type of surveillance system to determine
the extent of the problem. (You can find more information
about surveillance systems in the module “Introduction to
Surveillance.”)
Public health practitioners might also use clinical infor-
mation to propose a new idea or hypothesis for solving an
existing problem, which could lead to an analytic study.
The “what,” or clinical, information includes: symp-
toms (or a condition reported by the patient, such as pain),
signs (or observed evidence of disease, such as blood in
the stool), laboratory results, whether hospitalization was
required, and whether the patient lived or died.

Describing Who, or Personal Information

Personal information about populations is another part of
descriptive epidemiology.

Northwest Center for Public Health Practice 3

Study Types in Epidemiology
Transcript

Personal information can be demographic, for example, age, sex, marital status,
personal habits including exercise, smoking, alcohol use, and so on.
Personal information can also be socioeconomic, for example education, occupa-
tion, income, place of work, or residence.
Finally, personal information can be cultural, which would include such categories
as ethnicity, eating and dietary habits, and religious preferences.

Describing Where, or Place

Place information may include where people become sick,
such as in the home or a vacation spot.
Place also includes where an exposure occurred, such as
in a restaurant serving contaminated food or a cruise ship.
And finally, place also includes the source of contamina-
tion, such as a farm or a poultry-packing plant.

Describing When, or Time

Time information can be analyzed in three basic ways:
trends, patterns, or epidemics.
A trend is a movement or change in something over
time. Long-term trends of disease occurrence are called
secular trends. We usually report them by year. This graph,
for example, shows reported measles cases in the United
States by year. Measles vaccine was first used in 1963.
The graph shows the dramatic decrease in measles reports
following the introduction of the vaccine.
Patterns are cyclical trends seen consistently over several
years, often as seasonal trends. This graph, for example,
shows pneumonia and influenza deaths reported by 122
cities in the United States.
In this graph we also see peaks above the epidemic
threshold. These peaks, corresponding to increased influ-
enza activity, show influenza epidemics. Epidemics are
increased occurrences of disease or conditions above the expected number for a
particular time and place.
An epidemic may involve many cases, or it may involve a single case, depending
on the expected number. For example, a single naturally acquired polio case in the
United States would be an epidemic, since polio has been eliminated from the west-
ern hemisphere for many years.
This graph illustrates a Salmonella epidemic in Washington State in 1999, caused
by contaminated orange juice. The effects of a rapid investigation and intervention
are clearly demonstrated in this graph.

Northwest Center for Public Health Practice 4

Study Types in Epidemiology
Transcript

Now we will pause for the first of several interactive exercises about the material
we have just covered. Please note that the exercises sometimes take several seconds
to load.

Exercise 1

Descriptive Study Types

Descriptive studies, as the name implies, describe condi-
tions or diseases. In particular, descriptive studies identify
patterns in person, place, and time. Here are three of the
most basic types: a case report, a case series, and an inci-
dence study. These types of study involve no comparison
group. They are merely descriptive.
A case report is a detailed description of the person,
place, and time information of a specific case of disease
or condition. Case reports usually are about unexpected
symptoms in a disease, an unexpected event while treating
a patient, or unique therapeutic approaches.
A case series describes the person, place, and time information about a group of
cases. It can be retrospective, looking back in time, or prospective, looking forward
in time, and usually involves a small number of patients, such as those who were
given similar treatment.
Case reports and series permit discovery of new diseases, unexpected effects,
and provide data for generating hypotheses. Data from a case series may be used in
analytic studies to investigate possible causal factors.
An incidence study describes the incidence, or number of new cases of a disease
or condition, during a specific time in a specific population. One advantage of inci-
dence studies is that they allow calculation of true rates of disease occurrence for a
better estimation of risk.

Descriptive Studies Compared

Let’s take a look at how these three descriptive study types
relate to each other using West Nile virus as an example.
Remember, case reports describe individual data,
perhaps on a form such as this one, or a report in a medical
journal. West Nile virus in a transfusion or transplant recipi-
ent might be reported as a case report.
A case series for West Nile virus, on the other hand,
might be presented in a table that lists and describes all
West Nile cases in the Northwest.

Northwest Center for Public Health Practice 5

Study Types in Epidemiology
Transcript

An incidence study could use a map to display all the new West Nile cases in the
northwestern United States in 2006.
Now, let’s look at analytic studies.

Overview of Analytic Studies

As you may recall, descriptive studies are useful if very little
is known about a new disease or condition. They can be
used to generate hypotheses on risk factors and causes of
disease.
Analytic studies, on the other hand, are usually larger
and more complex than descriptive studies. They’re often
used to assess determinants of diseases, focus on risk
factors and causes, and analyze the distribution of expo-
sures and diseases. A key feature of analytic studies is that
they use comparison groups.
In contrast to descriptive studies, which generate
hypotheses, analytic studies are used to test hypotheses. They are used to look for
and measure associations.
(For more information on these kinds of measurements, see the module on
Measuring Risk in Epidemiology.)

Descriptive and Analytic Studies Compared

To see how descriptive and analytic studies differ, let’s look
at ways to study severe acute respiratory syndrome, or
SARS.
A descriptive study of SARS, for example, might be a
case series describing person, place, and time information
about the first 100 cases, or patients, with SARS.
In contrast, an analytic study might be used to measure
risk factors for the reasons why people get SARS, such as
contact with animals or infected people.

Types of Analytic Studies

Epidemiologists conduct two main types of analytic studies: experimental and
observational.
Experimental studies use a randomized selection process. A process based on
chance is used to assign study subjects to different exposure groups.
Experimental studies may be either clinical, such as studying a new drug to
prevent influenza, or community-based, for example, studying the overall effective-
ness of a new drug in preventing influenza in a community.

Northwest Center for Public Health Practice 6

Study Types in Epidemiology
Transcript

Unlike experimental studies, observational studies are

non-randomized. In these studies the investigator does not
assign exposures to the study subjects, but simply observes
the patterns of exposure as they occur.
Observational studies fall into four categories: cohort,
case control, cross-sectional, and ecologic.
Cohort studies observe groups of people who experi-
ence and who don’t experience an exposure, such as flu
vaccine, to compare how many in each group become ill,
or in this case, get influenza.
In contrast, in case-control studies the investigator
compares a group of people who already are ill, such as
with influenza, to a similar group who didn’t get ill. The investigator records who
had exposures of interest. In the case of influenza this includes who had been
vaccinated.
Cross-sectional studies are a snapshot of a specific period in time, with the aim of
finding the same kind of relationships as experimental studies, but often at less cost.
A cross-sectional study of influenza might look at the number of cases at a specific
time in different occupations involving indoor or outdoor work.
Ecologic studies look at group data, not individual data. In the case of influenza,
an ecologic study might compare the rates of influenza cases in different Pacific
Northwest counties with immunization levels in those counties.
We’ll look at these different types of studies in more detail in the next few slides.

Experimental Studies
Experimental studies involve assigning subjects to exposures
randomly and following them over time to determine if
they develop or recover from disease. Experimental studies
fall into two categories: clinical trials and community trials.
Clinical trials use data from individual people. The
investigator randomly determines the type of exposure, for
example, to a new drug to treat cancer. The study partici-
pants are then followed to determine the effect of treat-
ment on them. People who received the new drug are
compared with people who received an older drug or no
drugs at all.
In community trials, on the other hand, the study group
is the entire community, rather than individuals. Researchers might investigate
whether a media campaign to reduce smoking was effective. The researchers would
select a community to receive the media campaign. They would then compare the

Northwest Center for Public Health Practice 7

Study Types in Epidemiology
Transcript

smoking rates over time in this community with those in another community that did
not to receive the intervention.
Many consider experimental studies to be the “gold standard” of epidemiol-
ogy because they closely resemble experiments in scientific laboratories. However,
experimental studies are frequently very expensive and take a long time to perform.
Also, ethical reasons may prevent performing experimental studies. For example,
if a drug is suspected to have some value in preventing illness and death, it may be
difficult to perform a clinical trial, which may involve randomly not giving the drug
to some people.
Let’s pause now while you answer some questions on what you have just learned.

Exercise 2

Observational Studies
Now let’s look at the other main type of analytical study:
observational studies.
In observational studies, the researcher does not deter-
mine who receives the exposure. The researcher simply
observes or records the study participants and their
outcomes.
Observational studies come in four main types: Cohort,
case-control, cross-sectional, and ecologic.
In cohort studies, researchers determine the study popu-
lation’s exposure and observe over time who gets ill.
In case-control studies researchers identify people who
are ill and select or identify a comparison, or control group of people who aren’t ill.
Researchers then compare the prior exposures of the two groups.
In cross-sectional studies, researchers survey both the exposure and the condition
or disease among individuals at a single moment in time.
In ecologic studies researchers survey the community
level exposures and the condition or disease, but entire
populations are compared rather than individuals.
Let’s take a more in-depth look at cross-sectional and
ecologic studies.

Cohorts and Cohort Studies

Cohorts are groups of people who share similar characteris-
tics. Cohort studies may involve people in the same school,
church congregation, town, city, or occupation.
In cohort studies, people under investigation are divided
into two groups, depending on whether they were exposed

Northwest Center for Public Health Practice 8

Study Types in Epidemiology
Transcript

to something. The exposed and unexposed groups are observed to determine and
compare the proportion of each group that develops the disease.
Cohort studies may be more useful when the study population is well defined, if
the exposure is uncommon, or if there are several possible outcomes from the expo-
sure. Cohort studies can be prospective or retrospective.

Cohort Study Types

Prospective cohort studies are those in which the investiga-
tion is performed starting with exposures in the present and
looking forward to outcomes in the future. Retrospective
cohort studies, on the other hand, start with exposures in
the past, and then look forward to outcomes in the more
recent past or the present.
For example, when a new influenza vaccine is devel-
oped, scientists may study a cohort of individuals, some
of whom receive the vaccine, and some who do not. This
cohort, or group, of people can be followed prospectively
over the next influenza season to see what proportion of
those vaccinated or not vaccinated become ill with influenza.
On the other hand, if an influenza outbreak occurs in a nursing home, a retro-
spective cohort study might be done to determine if influenza vaccination was
protective. Immunization records at the nursing home would be reviewed to see
who had received influenza vaccinations in the past. To determine if the vaccine
provided protection against influenza, the nursing home records would also be
reviewed—to see what proportions of vaccinated and unvaccinated residents
developed influenza.
Similar cohort studies might be done following food borne outbreaks, such as
in people who become ill at a church supper or a wedding reception.

Case-Control Studies
The second type of observational study is the case-control
study. This involves selecting people with a particular
disease or condition. These people are called “cases.”
Another group of people are selected who don’t have
the disease. These are called “controls.” Ideally, the
controls are similar to the cases in every way, except
they do not have the disease or condition of concern.
Information is collected from the cases and the controls to
document and compare their exposures.
Let’s look at an example of an E. coli O157:H7 outbreak.

Northwest Center for Public Health Practice 9

Study Types in Epidemiology
Transcript

In this example, 16 people became ill with gastroenteritis. Of these, 12 had eaten at
Restaurant A prior to their illness.
A group of 16 well people were selected as a control group. These people had
the same age and lived in the same neighborhood as the ill people. They differed
only in that they were well. In comparing the possible exposures of the two groups,
researchers found that none of the control group had eaten at restaurant A. This
finding suggests that restaurant A was the outbreak source.
Case-control studies may be more useful when the study population is not well
defined, if the disease is relatively rare, or if there are many possible exposures for a
disease. Case-control studies are always retrospective.

Comparing Cohort and Case-Control Studies

Let’s review a bit about these two types of analytic studies.
Cohort studies start out by classifying a group of people
according to an exposure. Then, the investigator deter-
mines whether the exposed and unexposed people
develop a disease or not. Cohort studies can be either
prospective, following a group over time, or retrospective,
looking at what happened to the group in the past.
In case-control studies, the investigator identifies a group
of people with a disease. Then the investigator selects a
comparison or control group without the disease. The two
groups are compared and analyzed for exposures. Because
case-control studies begin with people who are sick or not sick, and interview
people about previous exposures, they can only be retrospective.

Cross-Sectional and Ecologic Studies

In contrast to an incidence study, which looks at new cases
over time, cross-sectional, or prevalence, studies describe
a population at a specific point in time. You can think of
them as a snapshot of the population. They’re called cross-
sectional because they look at both the exposure and the
condition or disease. They’re called prevalence studies
because they look at the amount of a disease or condition
at a moment in time. The investigator defines a population
and identifies the presence or absence of exposure and
disease for each individual. Each individual is categorized
into one of four possible groups: exposed and having the
disease, exposed and not having the disease, not exposed and having the disease,
and not exposed and not having the disease. A cross-sectional study can be used

Northwest Center for Public Health Practice 10

Study Types in Epidemiology
Transcript

to assess the prevalence of acute or chronic conditions in a population, and it may

suggest a possible risk factor for a disease.
Ecologic studies focus on groups, rather than on separate individuals. Exposure
and risk factors are known only at the group level. An example would be the correla-
tion between average air pollution concentrations in different cities in the Northwest
and chronic lung disease in these same cities. Ecologic studies may be used to gener-
ate or test hypotheses of an association between exposure and disease. One poten-
tial problem with ecologic studies is that just because an association exists between
exposure and disease at the group level, it does not mean that a similar association
exists at the individual level.
Let’s pause now while you answer some questions on what you have just learned.

Exercise 3

Summary
In summary, epidemiologic investigations involve two types
of studies: descriptive and analytical.
Descriptive studies are more useful if little is known
about a new disease or condition. They can also be used to
generate hypotheses on risk factors and causes of disease.
Descriptive study types include the case report, cases
series, and incidence studies.
Analytic studies, on the other hand, should be performed
if hypotheses exist for risk factors and diseases, and if these
hypotheses need to be tested. Analytic studies fall into two
categories: experimental and observational.
Experimental studies, which include clinical and community trials, may be used to
study the effects of new drugs or vaccines. However, observational analytic studies
of drugs are also performed, especially after drugs or vaccines are licensed.
Observational studies fall into four different types.
Cohort studies may be more useful when the study population is well defined, if
the exposure is uncommon, or if there are several possible outcomes from the expo-
sure. Cohort studies can be either prospective or retrospective.
Case-control studies may be more useful when the study population is not well
defined, if the disease is relatively rare, or if there are many possible exposures for a
disease. Case-control studies are always retrospective.
Cross-sectional, or prevalence, studies look at individual exposures and condi-
tions at the same time. Ecologic studies look at exposures and conditions at the same
time, but at the community level. Both of these study types look for the same kind of
relationships that might be shown by an analytic study, but often at less cost.

Northwest Center for Public Health Practice 11

Study Types in Epidemiology
Transcript

Resources
If you would like to learn more about the concepts in this
module, you might want to explore some of the resources
listed here.
Now, if you’re ready, please go on to the final
assessment.

Final Assessment

Northwest Center for Public Health Practice 12