Screening
Definition
Presumptive identification of unrecognized disease by
application of a procedure to sort out apparently
well persons who probably have a disease from
those who probably do not.
Screening tests are not intended to be diagnostic...
Screening
Issues relating to the screening tests
Validity
Sensitivity
Specificity
Predictive values
Positive
Negative
Purpose of screening
It is carried out in a hope that earlier diagnosis and
subsequent treatment favorably alters the natural
history of the disease in a significant proportion of those
who are identified as positives
Types of screening
Mass screening, no selection of population (e.g.,
checking all infants for hearing problems)
Selective screening (e.g., by age and sex:
mammograms for women aged over 40)
Multiphase screening (a series of tests, as family
doctors do at annual health exams)
When to screen??
The Principles of Screening
The choice of disease for which to screen;
The nature of the screening test or tests to be used;
The availability of a treatment for those found to have
the disease;
The relative costs of the screening.
Screening Test Criteria
Simplicity
Safety
Rapidity
Ease of administration
Cost
Repeatability
Validity of a screening test (Accuracy)
Sensitivity
Probability that a person who has a disease will test
positive for that disease by the screening test.
Specificity
Probability that a person who is free of a disease will test
negative for that disease by the screening test.
Validity of a screening test
Sensitivity of a test is the ability of the test to
detect disease in all those who actually have
the disease(i.e., correctly identify all those who
have the disease)
Specificity of a test is the ability of the test to
rule out disease in all those in whom the
disease is actually absent (i.e., correctly identify
all those who do not have the disease).
Validity
An ideal screening test is one that is 100% sensitive and
100% specific.
In practice this does not occur.
2*2 table
The simplest cross tabulation is a 2 x2 table
A 2 x 2 table is one which has only two rows
and two columns.
Validity
....... an example
Disease status
Screening
test's result
Present
Absent
Total
Positive
TP
FP
TP + FP
Negative
FN
TN
Total
TP + FN
FP + TN
(All Diseased)
(All Non-diseased)
(All test positives)
FN + TN
(All test negatives)
1000
interpreting the cells
a+c = All those who actually have the
disease.
b+d = All those who actually do not
have the disease.
a+b = All those who test positive on
the screening test.
c+d = All those who test negative on
the screening test
Validity
....... an example
Disease status
Screening
test's result
Present
Absent
Total
Positive
80
100
180
Negative
20
800
820
Total
100
900
1000
Validity
....... an example
Sensitivity
Specificity
TP / TP + FN
TN / FP + TN
80/100 = 0.8 or 80%
800/900 = 0.888 or 88.8%
Probability (chance) that a
diseased individual will be
correctly identified by a
screening test.
Probability (chance) that a
disease free individual will be
identified as such by a
screening test.
Predictive values (Yield)
Positive
Probability that a person who tests positive for a disease
indeed has the disease.
Negative
Probability that a person who tests negative for a disease
is indeed disease free.
Yield of screening test
Positive predictive value (PV+) is the proportion of
disease positive in test positives.
Or
is the probability that an individual with a positive screening
result has the disease
Negative predictive value (PV-) is the proportion of
disease negative in test negatives.
Or
is the probability that an individual with a negative screening
result do not have the disease
Predictive values
Disease status
Screening
test's result
Present
Absent
Total
Positive
80
100
180
Negative
20
800
820
Total
100
900
1000
Predictive values
Positive
TP / TP + FP
80/180 = 0.444 or 44.4%
Probability (chance) of
having the disease given a
positive test.
Negative
TN / TN + FN
800/820 = 0.975 or 97.5%
Probability (chance) of not
having the disease given a
negative test.
Predictive values
Disease status
Screening
test's result
Present
Absent
Total
Positive
400
56
456
Negative
100
444
544
Total
500
500
1000
Predictive values
.
Positive
Negative
TP / TP + FP
400/456 = 0.877 or 87.7%
TN / TN + FN
444/544 = 0.816 or 81.6%
Probability (chance) of
having the disease given a
positive test.
Probability (chance) of not
having the disease given a
negative test.
Prevalence
P= Total # of Diseases Individuals
________________________
Total Population
P =
a+c
________________
a+b+c+d
X 100
Logic of Screening
Apparently well population
Screening test
Negative results
Disease
FN
No disease
TN
Positive results
Disease No disease
TP
FP
Summary
Screening ????
Validity
Sensitivity
Specificity
Predictive values
Positive
Negative