Epidemiological Methods

Epidemiology is derived from the Greek word.

 Epi: On or upon.

 Demos: people.

 Logos : the study of

 Epidemiology is the basic science of Public Health

“Epidemiology is the study of the incidence, prevalence, and distribution of illness or disease in

a given population.”

Incidence refers to the rate of new cases of illness that develop within a given period of time.

 Incidence gives us some sense of whether the rate of new cases of the illness or disorder

is on the increase (e.g., is the rate of newly diagnosed AIDS cases increasing this year

compared to last year?).

Prevalence refers to the overall rate of cases (old or new) within a given period.

 Prevalence rates estimate what percentage of the target population is affected by the

illness or disorder. For example, the lifetime prevalence rate of schizophrenia is

estimated at 1%, suggesting that a member of the general population has 1 chance in 100

of developing this disorder in his or her lifetime.

Historically, epidemiology has been most closely associated with medical research designed to

help understand and control the major epidemic diseases, such as cholera and yellow fever.

 Epidemiological methods can also be quite important in identifying groups of individuals

who are at risk.

A well-known example of epidemiological research is the study of Smoking and Health

(Surgeon General, 1964). This study linked cigarette smoking with lung cancer by the simple

methods of counting and correlating. Although there was great debate as to whether smoking

caused lung cancer, there were definite relationships and associations between smoking and lung

cancer (e.g., about 90% of lung cancer in males was associated with cigarette smoking, and the

amount and duration of smoking were positively correlated with the probability of cancer). It is

true that epidemiological research frequently suggests the possibility of multiple causation

(several factors must be present before the disease occurs, or the greater the number of associated

factors that are present, the greater the risk of the disease). It is equally true that obtained

correlations suggest causes rather than definitively prove causation. However, the entire story of

causation need not be known before preventive steps can be taken. Thus, we may not be sure that

smoking causes lung cancer, or we may believe that some inherited predisposition interacts with

smoking to produce cancer. Nevertheless, we know that groups of males who quit smoking

reduce their risk of lung cancer.

Epidemiological research is based on surveys or interviews. However, survey and interview

data present a number of issues and potential problems. For example, how do we define a mental

health problem, and having done so, where do we locate cases for counting? Checking only

clinics and hospitals means ignoring other possible locales. These difficulties are magnified

when we become interested in milder forms of dysfunction. In effect, we need objective methods

of defining and measuring a problem. Further, we need survey procedures that will enable us to

estimate the problem’s true incidence or prevalence and not just to locate those cases that are

already under treatment or that have identified themselves by seeking treatment.

Co-relational Methods

 Epidemiology often relies on co-relational method.

 It determines relationship between two variables and also its directions (increases,

decreases or no relation).

 These techniques enable us to determine whether variable X is related to variable Y. For

example, is a certain pattern of scores on an intelligence test related to specific

psychiatric diagnostic categories? Are particular patient characteristics related to therapy

outcomes? Is depression related to gender?

The Technique:

To correlate two variables, we first obtain two sets of observations. For example, suppose we

administer two tests to 10 study participants. One test measures anxiety and the other measures a

belief in external (rather than internal) control.

 The data are correlated, the result is a correlation coefficient; in this case, it is +.76,

indicating a strong positive relationship.

 The Pearson product-moment correlation coefficient is a commonly used index to

determine the degree of relationship between two variables. This is symbolized by r,

which may vary anywhere from –1.00 to +1.00. An r of +1.00 denotes that the two

variables are perfectly and positively related. An r of –1.00 indicates a perfect negative

relationship. The r of +.76 from the data of Table 4-2 signifies a high but less than perfect

relationship.

The Question of Causality:

As noted previously in the case of epidemiological research, correlational methods cannot

answer the question of cause and effect. For example, suppose that an investigator discovers a

correlation between being diagnosed with schizophrenia and indices suggesting high levels of the

neurotransmitter dopamine in the central nervous system. Does this mean that schizophrenia is

caused by excessive levels of dopamine or, alternatively, that the experience of an episode of

schizophrenia results in changes in dopamine levels? Maybe the real culprit is a third variable.

For example, many patients with schizophrenia have a long history of taking psychoactive drugs

(e.g., amphetamines); such long-term use could conceivably affect neurotransmitter levels.

Therefore, the investigator must avoid assuming that one variable causes the other because there

is always the possibility that a third (unmeasured) variable is involved.

Correlational methods can demonstrate that a cause–effect hypothesis is not valid. If the

expectation is that variable A causes variable B, we should at least be able to show they are

correlated. Failure to find a significant relationship most certainly contradicts the hypothesis.

Also, although causal inferences are not possible from correlation coefficients, this is not to say

that cause–effect relationships do not in fact exist.

Factor Analysis:

Derived from the correlational strategy, factor analysis is a way of examining the

interrelationships among a number of variables at the same time.

Example: Suppose we are trying to identify the basic elements of what is called “clinical skill.”

First, we ask a panel of judges to select 100 clinicians who are known to have excellent records

in providing skilled services. Then, we administer a large number of tests that are believed to

assess a variety of clinical skills and achievements. Let us suppose that we use the following

seven tests:
A IQ test

B Mathematical achievement test

C Test of spatial reasoning

D Test of analytical reasoning

E Measure of empathy

F Measure of personal adjustment

G Measure of altruism

Next, we correlate each of these tests with every other test. This gives us a correlation matrix in

which the correlations between all possible pairs of tests are displayed.

 When we look at the correlation matrix, an interesting pattern emerges. Measures A, B,

C, and D all show a strong positive relationship (correlations range from .70 to .80). At

the same time, E, F, and G also correlate highly with one another (correlations range from

.75 to .85).

 It would appear that two factors or dimensions are involved. Let us call them Factor 1

(derived from the correlations among A, B, C, and D) and Factor 2 (derived from E, F,

and G). Together, these two factors account for the significant relationships in the matrix.

The Experimental Method

The experimental method involves manipulating one variable to determine if changes in one

variable cause changes in another variable. This method relies on controlled methods, random

assignment and the manipulation of variables to test a hypothesis.

To determine cause–effect relationships among events, we must use experimental methods.

Consider the following hypothetical study, an expansion of a pilot study by Naylor et al. (2007),
that examined whether the addition of bibliotherapy, reading educational books about

psychological conditions, would improve outcome in patients receiving medication for

depression. The design of the research illustrates important features of the experimental method.

The study was carried out in a primary care clinic with participants ranging in age from 21 to 50.

The experimental group contained 50 participants who within the last month had started taking

an antidepressant medication. In addition to taking medication, the experimental group was

instructed to read a popular self-help book on depression, Feeling Good (Burns, 1999). To ensure

compliance, those in the experimental group were instructed to call in once per week to report

which chapters of the book they finished that week as well as their medication compliance. The

control group consisted of 50 participants who, like the experimental group, had started taking

antidepressant medication within the last month. Unlike the experimental group, these

participants did not receive a reading assignment, but simply called in once a week to report

medication compliance.

The two groups were matched initially on several variables, including level of depression before

medication, gender, age, socioeconomic level, and overall level of emotional and physical

distress. However, their assignment into the bibliotherapy or medication-only conditions was

determined randomly. The primary measure of interest was the level of depression reported at

baseline, after 1 month on medication, 2 months after starting medication(and following

bibliotherapy for the experimental group), and at 3-month follow-up. Figure 4-4 shows the

findings for self-reports of depression for the two groups. As can be seen, depression scores for

the experimental group were clearly lower following the bibliotherapy intervention as well as at

follow-up. These results suggest that bibliotherapy may help alleviate depression above and
beyond the effects of antidepressant medication.