PUBH501, Fall 2015: Topics in

Community & Preventive Medicine

Principles of Screening
Page 6 onwards

Learning Objectives
Define the characteristics that distinguish
screening from diagnosis
List the factors that make a disease or
condition appropriate for screening
Describe the attributes of a good screening
test
Describe the expected and unexpected
consequences of screening
List at least two sources of screening
recommendations
2

CPM Screening

Screening: Definition
Identification of unrecognized disease, disease
precursors, or susceptibility to disease in
persons without evidence of disease, by rapidly
applied tests, examinations or other procedures
Focuses on healthy populations, looks for
asymptomatic disease
Positive or suspicious findings are referred for
follow-up diagnostic tests and treatment
3

CPM Screening

Screening: Goal, Objective

Goal
to identify disease or susceptibility when it is
more easily / more successfully treated, i.e.,
to reduce morbidity and mortality
Objective
to classify individuals as likely or unlikely to
have the disease (or susceptibility)
those classified as likely are referred for
diagnostic testing
4

CPM Screening

Screening vs. Diagnosis

Screening = testing among asymptomatic persons

Diagnosis = testing among individuals suspected to have the disease to

confirm presence or absence of disease

CPM Screening

Screen for What?

Disease or condition itself (usually, subclinical
evidence of)
Risk factor(s) for disease / condition

Behaviors
Exposures
Biologic factors
Genetic traits

CPM Screening

Examples of Screening Tests

Questions
Clinical Examinations
Laboratory Tests
Genetic Tests
Radiographic exam

Goel
CPM Screening

What conditions should we screen for?

Should we routinely screen for:
breast cancer?
colon cancer?
lung cancer?
testicular cancer?
cardiovascular disease?
stroke?
HIV / AIDS?
scoliosis?
other?
8

CPM Screening

Criteria for Adopting Screening Program

Characteristics of disease / condition
Characteristics of screening test
Characteristics of health care system

CPM Screening

W.H.O. Prerequisites for a

Successful Screening Program (I)

1. Condition is important public health problem

2. Understanding of natural history of condition
3. Recognizable latent or early symptomatic
stage
4. Accepted treatment
5. Early treatment influences course and
prognosis of the disease
Wilson and Jungner

http://whqlibdoc.who.int/php/WHO_PHP_34.pdf

CPM Screening

10

W.H.O. Prerequisites for a

Successful Screening Program (II)

6. Suitable screening test or examination

Relatively sensitive and specific

Detects disease at latent or early symptomatic stage
Simple and inexpensive
Safe
Acceptable to population and providers

Wilson and Jungner

CPM Screening

11

W.H.O. Prerequisites for a

Successful Screening Program (III)

7. Available facilities for diagnosis and treatment

8. Screening costs economically balanced with
possible expenditures on medical care
9. Diagnostic workup for positive test result that
has acceptable morbidity given number of
false-positive results
10. Ongoing case-finding, not one-time shot
Wilson and Jungner
CPM Screening

12

Intrinsic Properties of a Screening Test

Reliability (reproducibility)
= extent to which repeated measurements get similar results

Validity (accuracy)

= ability of test to detect who has the disease and who does not
Usually judged against another test of greater known accuracy (gold
standard)
Measured by sensitivity, specificity, predictive value of positive or
negative test

13

CPM Screening

Accuracy / Validity

Screening Test Value

100

75

50

25

Disease +

Disease

For an ideal screening test, the test values of those with

and without the condition do not overlap
CPM Screening

14

Accuracy / Validity

Screening Test Value

100

75

50

25

Disease +

Disease

In the real world, test values of healthy and ill subjects

DO overlap
CPM Screening

15

Accuracy / Validity

Screening Test Value

100

75

50

25

Disease +

Disease

Need to set threshold (cut-off) value to distinguish abnormal

from normal values. Where should the cut-off be?
CPM Screening

16

Accuracy / Validity

Screening Test Value

100

positive or
abnormal

75

50

negative or
normal

25

Disease +

Disease

Arbitrary threshold (cut-off) value distinguishes what we call

abnormal from normal values.
CPM Screening

17

Accuracy / Validity
Four possible combinations of disease and test

True positive: ill with abnormal test

True negative: healthy with normal
(negative) test

False positive: healthy with abnormal test

False negative: ill with normal test
18

CPM Screening

Accuracy / Validity
100
Screening Test Value

True positives

75

False positives

50

False negatives

25

True negatives
Disease +

Disease

Any value above the red line will be considered a

positive screening test for the disease; below the red
line will be considered negative
CPM Screening

19

Sensitivity
100
Screening Test Value

True positives

75

50

25

False negatives
0

Disease +

Disease

Sensitivity = Among those with disease, probability of a

positive test (= TP / TP+FN)
CPM Screening

20

Specificity

Screening Test Value

100

75

False positives

50

25

True negatives
Disease +

Disease

Specificity = Among those without disease, probability of a

negative test (= TN / TN+FP)
CPM Screening

21

Predictive Value of a Positive Test

100
Screening Test Value

True positives

75

False positives

50

25

Disease +

Disease

Predictive Value Positive = Among those who test positive,

probability of actually have the disease (= TP / TP+FP)
CPM Screening

22

Predictive Value of a Negative Test

Screening Test Value

100

75

50

False negatives

25

True negatives
Disease +

Disease

Predictive Value Negative = Among those who test negative,

probability of actually not having the disease (= TN / TN+FN)
CPM Screening

23

Example: Sensitivity, Specificity

Population = 100,000
Prevalence = 10%

Sensitivity = 95%
Specificity = 98%

Disease Present

Disease
Absent

Total

Test positive

True Positive (a)

False Positive (b)

a+b

Test
negative

False Negative
(c)

True Negative
(d)

c+d

a+c

b+d

a+b+c+d
24

CPM Screening

Example: Sensitivity, Specificity

Population = 100,000
Prevalence = 10%
Disease Present

Test positive
Test
negative

Sensitivity = 95%
Specificity = 98%
Disease
Absent

Total

9,500

1,800

11,300

500

88,200

88,700

10,000

90,000

100,000
25

CPM Screening

Predictive Value of Positive Test

Sens = 95%, Spec = 98%, Prev = 10%
Disease Present

Test positive
Test
negative

Disease
Absent

Total

9,500

1,800

11,300

500

88,200

88,700

10,000

90,000

100,000

Predictive value positive =

9,500 / 11,300 = _____
84.1%
26

CPM Screening

Predictive Value of Positive Test

Sens = 95%, Spec = 98%, Prev = 1%
Disease Present

Test positive
Test
negative

Disease
Absent

Total

950

1,980

2,930

50

97,020

97,020

90,000

100,000

1,000

Predictive value positive =

950 / 2,930 = _____
32.4%
27

CPM Screening

Expected Consequences of Screening

True-positive result
early diagnosis
early (possibly less radical) treatment
reduced morbidity, mortality, disability
reduced cost
labeling effect
True-negative result
reassurance

28

CPM Screening

Law of Unexpected Consequences

False-positive
Positive screening test even though person does not really
have the disease / condition
False-negative
Negative screening test even though person does really have
the disease / condition

29

CPM Screening

False-Positive Screening Test Result

False-positive: positive screening test even
though person does not really have the disease
/ condition
Consequences of false-positive test result
Unnecessary follow-up tests
(inconvenience, morbidity, expense)
Labeling, anxiety
Over-treatment of questionable
abnormalities
Fear of future tests
30

CPM Screening

False-Negative Screening Test Result

False-negative: negative screening test even
though person does really have the disease /
condition
Consequences of false-negative test result
Delayed diagnosis, more advanced disease,
premature death / disability
Disregard of early signs / symptoms
False reassurance
Exposure of others to infection
31

CPM Screening

Sources of Screening Recommendations

Government agencies
Panels convened by government agencies
Medical specialty associations
Special interest organizations
Individual experts
Others
National Guideline Clearinghouse:
www.guideline.gov
32

CPM Screening

U.S. Preventive Services Task Force

Independent panel of private-sector experts in
primary care and prevention
conducts rigorous, impartial assessments of the
scientific evidence of effectiveness of clinical
preventive services, including screening,
counseling, and preventive medications
first convened in 1984
recommendations are the gold standard
www.ahrq.gov/clinic/uspstfix.htm
33

CPM Screening

Evidence-Based Medicine and

Recommendations
The conscientious, explicit, and judicious use of current
best evidence in making decisions about the care of
individual patients. The practice of evidence-based
medicine means integrating individual clinical expertise with
the best available external clinical evidence from
systematic research.
Evidence-based recommendations require
1. test or procedure is medically effective in reducing
morbidity or mortality
2. medical benefits must outweigh the risks
3. Cost of test or procedure must be reasonable
4. Recommended actions must be practical and feasible
Eggert RW, adapted from Eddy DM

CPM Screening

34

Strength of Recommendations
A. USPSTF strongly recommends that clinicians routinely
provide [the service] to eligible patients
B. USPSTF recommends that clinicians routinely provide
[the service] to eligible patients
C. USPSTF makes no recommendation for or against
routine provision of [the service].
D. USPSTF recommends against routinely providing [the
service] to asymptomatic patients.
I. USPSTF concludes that the evidence is insufficient to
recommend for or against routinely providing [the
service].
Harris 2001

CPM Screening

35

Influence of Evidence and Benefit

on Grade
Net Benefit
Quality of
evidence

Substantial

Moderate

Small

Zero /
negative

Good

Fair

Poor = I
36

CPM Screening

Selected Recommended Screening Tests (USPSTF)

Infants
Phenylketonuria
Congenital hypothyroidism

Sexually active adolescents (girls)

Chlamydia ( 25 yrs)
Cervical cancer (PAP test)
Rubella

Adults

Cervical cancer (PAP test)

Colorectal cancer (men & women 50 yrs)
Breast cancer ( 40 yrs)
Lipid disorders (men 35 yrs, women 45 yrs)
Osteoporosis (women 65 yrs)

CPM Screening

37

Example of Positive Recommendation

USPSTF recommends that women aged 65 and older be
screened routinely for osteoporosis. The USPSTF
recommends that routine screening begin at age 60 for
women at increased risk for osteoporotic fractures.
Rating: B Recommendation
Rationale: USPSTF found good evidence that the risk for
osteoporosis and fracture increases with age and other factors,
that bone density measurements accurately predict the risk for
fractures in the short-term, and that treating asymptomatic women
with osteoporosis reduces their risk for fracture. The USPSTF
concludes that the benefits of screening and treatment are of at
least moderate magnitude for women at increased risk by virtue of
age or presence of other risk factors.
38

CPM Screening

Example of Negative Recommendation

USPSTF recommends against the routine screening of
asymptomatic adolescents for idiopathic scoliosis.
Rating: D Recommendation.
Rationale: USPSTF did not find good evidence that screening
asymptomatic adolescents detects idiopathic scoliosis at an earlier
age than detection without screening. The accuracy of the
forwarding bending testis variable, and there ispoor follow-up of
adolescentsidentified in community screening programs.
USPSTF found fair evidence that treatment of idiopathic scoliosis
during adolescence leads to health benefits in only a small proportion
of people. Most cases detected through screening will not progress
to a clinically significant form of scoliosis.
USPSTF found fair evidenceunnecessary brace wear and
unnecessary referral for specialty care.
CPM Screening

39

Summary
Screening = healthy population, asymptomatic
disease, non-diagnostic test
No test is perfect, but high sensitivity and
specificity are desirable consider
consequences of false-positives and falsenegatives
Screen if important health problem, decent
acceptable test, and available effective early
treatment, based on the best available data
and unbiased review
40

CPM Screening