EPIDEMIOLOGY AND

BIOSTATISTICS
REVIEW, PART II
Danielle Tsingine Chang MSII
Topics to be covered today:
Incidence vs. prevalence
Sensitivity
Specificity
Negative and Positive Predictive Value
Case fatality
Quantifying risk
T-test
ANOVA
Chi-square
 Incidence vs. Prevalence
Incidence Prevalence

The number of new cases that The number of affected persons

occur during a specified period present in the population at a
specific time divided by the
of time in a population at risk number of persons in the
for developing the disease population at that time

Incidence is a rate because it Useful measure of burden of

includes time disease in a community

= # of existing cases / Population

= # of new cases of a disease at risk
in a specified time period /
Population at risk of developing Prevalence Incidence x
disease during same time disease duration
period
Incidence and Prevalence
How can we increase prevalence?

Add new cases i.e. increase the incidence

of a disease

How can we decrease prevalence?

Death or cure = reduce the number of

diseased persons in the population
Disease X has an incidence of 5 per 1,000 per year and 80% of the
people with the disease will die from it. Before 1985, lab test A was
used to detect the disease. In 1985, a screening test B was developed
that could detect disease X two years earlier than test A. However,
early detection of the disease did not improve prognosis for disease X.
Assume after 1985, all people were screened for disease X using test
B and that test B has a higher sensitivity and specificity than test A.

Compare the incidence and prevalence of disease X in 1984 to 1985,

the year test B was first used.

In 1985, it is true that:

a. Incidence is higher and prevalence is higher than in 1984
b. Incidence is higher in 1984 and prevalence remains the same
c. Incidence is the same and prevalence is higher than in 1984
d. Both incidence and prevalence are the same as in 1984
e. Incidence is the same as in 1984 and prevalence is lower than in
1984
Answer
A. Incidence is higher and prevalence is higher than in
1984

Test B catches disease X two years earlier than test A, thus

the duration of the disease is increased. It is also more
sensitive and specific, so you are catching more new cases
i.e. increasing the incidence

Prevalence = incidence x duration of disease

The incidence of a chronic disease in a population may be
decreased by:

a. Prolonging the lives of persons with the disease

b. Decreasing the case-fatality rate for the disease
c. Improving the treatment of the disease once it has been
diagnosed
d. Primary prevention
e. Secondary prevention
Answer
d. Primary prevention

Primary prevention = prevent disease occurrence (i.e.

vaccine), thus prevent the numbers of new cases

Secondary prevention = early detection of a disease (i.e.

screening)

Tertiary prevention = reduce disability from disease (i.e.

Sensitivity and Specificity
Used more often in a public health setting

In effect, we are asking, If we screen a population, what

proportion of people who have the disease will be
correctly identified?
Sensitivity
How good is the test in correctly identifying those who
had the disease?

Definition: Proportion of diseased people who were

correctly identified as positive by the test

Sensitivity = True Positives/(True Positives + False

Negatives) OR = 1 false-negative rate

Value close to 100% is desirable for ruling out disease

Used for screening in diseases with low prevalence

Specificity
How good is the test in correctly identifying those who did
not have the disease?

Definition: proportion of non-diseased people who are

correctly identified as negative by the test

Specificity = True Negatives / True Negatives + False

Positives OR = 1 false-positive rate

Value close to 100% is desirable for ruling in disease

Used as a confirmatory test after a positive screening test

Positive and Negative Predictive value
Used more often in a clinical setting

We are asking, If the test result is positive in this patient,

what is the probability that this patient truly has the
disease?
Positive Predictive Vale
What proportion of patients who test positive actually have the
disease in question?

Definition: proportion of positive test results that are true positive

PPV = True positives / Total number who tested positive (TP + FP)

PPV varies directly with prevalence high prevalence = high PPV

Negative Predictive Value
If the test is negative, what is the probability that this patient does not
have the disease?

Definition: Proportion of negative test results that are true negative

NPV = True Negatives / All people who tested negative (TN + FN)

NPV varies inversely with prevalence high prevalence = low NPV

Pulling it all together
A physician examined a population of 1,000 patients in an
attempt to detect heart disease. The prevalence of heart
disease in this population is known to be 15%. The
sensitivity of the physicians exam is 60% and the
specificity is 80%. Patients who test positive by the
physician are sent for examination by a cardiologist.

1. What is the total number of people who test positive for

heart disease based on the physicians exam?
a. 90
b. 260
c. 60
d. 680
e. 740
Answer
1. b. 260
Step 1: Prevalence is 15%. If the population is 1,000, then 150 total have heart
disease (1,000 X 0.15 = 150)
Step 2: Set up 2x2 table using the given values for sensitivity (60%) and
specificity (80%)
A physician examined a population of 1,000 patients in an
attempt to detect heart disease. The prevalence of heart
disease in this population is known to be 15%. The sensitivity
of the physicians exam is 60% and the specificity is 80%.
Patients who test positive by the physician are sent for
examination by a cardiologist.

2. What is the positive predictive value of the physicians

exam?
a. 34.6%
b. 78.8%
c. 85.0%
d. 91.9%
Answer
2. a. 34.6%
Quantifying risk
How do we determine whether a certain disease is
associated with a certain exposure?

To determine whether an association exists, we can use

data from case-control and cohort studies
Odds Ratio
Most often used in case-control studies

Defined as the ratio of the odds that the cases were exposed to the
risk factor to the odds that the controls were exposed
Relative Risk
Used most often in cohort studies

Defined as the risk (i.e. incidence) of developing disease

in the exposed group divided by the risk in the unexposed
group
Attributable risk
Defined as the difference in risk between exposed and
unexposed groups, or the proportion of disease
occurrences that are attributable to the exposure
Absolute risk reduction (ARR) and
Number needed to treat and harm
ARR is defined as the absolute reduction in risk associated with a
treatment as compared to a control

Number needed to treat is defined as the number of patients who

need to be treated for 1 patient to benefit

Number needed to harm is defined as the number of patients who

need to be exposed to a risk factor for 1 patient to be harmed
A study was performed to determine if an association exists between
smoking and lung cancer. In this study, 100 people with a history of
smoking tobacco for 10 years and 100 people with no smoking history
were followed for 20 years, and the incidence rates for lung cancer
were compared in the two groups. The results are below.
Developed lung Did not develop
cancer lung cancer

(+) Smoking 40 60
( - ) Smoking 10 90

1. What is the relative risk for lung cancer in the exposed group?
a. 5
b. 10
c. 4
d. 3
Answer
1. c. 4

RR = Incidence rates in the exposed / incidence rates in

the unexposed

RR = (40/40+60) / (10/10+90) = 0.4/0.1 = 4

A study was performed to determine if an association exists between
smoking and lung cancer. In this study, 100 people with a history of
smoking tobacco for 10 years and 100 people with no smoking history
were followed for 20 years, and the incidence rates for lung cancer
were compared in the two groups. The results are below.

Calculate the attributable risk and the absolute risk reduction

 Or the incidence in the exposed the incidence
in the unexposed

AR = 0.4 0.1 = 0.3 means that 0.3 (or 30%) of people who smoke
develop lung cancer as a result of smoking (i.e. the exposer)

Absolute risk reduction percent = (40% of those who smoke develop lung
cancer) (10% of those who do not smoke develop lung cancer) = 30%

Describes the difference in risk of developing lung cancer between smokers

and nonsmokers.
Mortality rate vs. case-fatality rate
Mortality rate = a rate calculated Case-fatality rate = # of
by dividing the # of deaths individuals dying during a
occurring in the population during specified period of time after
a stated time period / # of persons disease onset or diagnosis / # of
at risk of dying during the period individuals with the specified
disease
Can limit the population by age,
gender, and disease i.e. Key: denominator is limited to
annual mortality rate from all causes for those who already have the
children younger than 10 years disease
annual mortality rate from lung cancer in
one year
Measure of the severity of the
Key: the denominator represents disease
the entire population at risk of
dying from the disease, including
those who have the disease and
those who do not (but are at risk
of developing disease)
t-test vs. ANOVA vs. Chi-squared
t-test = checks the difference between the means of 2
groups

ANOVA = checks the difference between the means of 3

or more groups

Chi-square = checks the difference between 2 or more

percentages or proportions of categorical outcomes (NOT
mean values); used for frequency data rather than for
comparison of means
A physician is studying the effects of drug A and drug B on
cognitive performance in Alzheimer patients. She
administers a memory test to two groups of subjects (those
taking drug A and those taking drug B) and compares their
mean scores. Which of the following statistical tests would
be most appropriate for this purpose?

a) ANOVA
b) Chi-square test
c) Linear regression analysis
d) t-test
e) Multiple linear regression
Answer
d. t-test

t-test is used to compare two means derived from two

samples
Resources
Gordis, Leon. Epidemiology. Philadelphia: Saunders
Elsevier, 2009.

Le, T. and V. Bhushan. 2013. First aid for the USMLE step
1 2013. New York: McGraw-Hill Medical.

USMLE Step 1 Qbook, Fifth edition