2. Learning objectives
By the end of this session, you should be able to:
• Define and explain epidemiologic screening principles
• Identify criteria to be fulfilled before a screening program is
introduced.
• Understand & calculate sensitivity, specificity, PPV and NPV of a
screening test.
3. ‘The presumptive identification of unrecognized disease by
application of rapid tests or examinations, or other procedures
that can be applied rapidly.’
‘The examination of asymptomatic people in order to classify
them as likely or unlikely to have the disease of interest.’
Early detection of disease (earlier than would usually occur in
clinical practice) in the hope of modifying prognosis
b/c those who have a positive result from the screening test
are made to go further evaluation
What is screening?
4. Screening sort out apparently well persons
who probably have a disease from those
probably do not have a disease
Q: When should we do
screening?
Q: what are the aims of
screening?
7. To reverse, halt, or slow the progression of disease more effectively
than would probably normally happen
To alter the natural course of disease for a better outcome for
individuals affected.
To reduce morbidity & mortality through early detection & treatment
(prevention)
To protect the society from contagious d/s
• Assumptions
– Early detection before development of symptoms will lead
to a more favorable prognosis.
– Treatment begun before the disease becomes clinically
manifest will be more effective than later treatment …
Aims of screening
8. To protect an institution
• Rational allocation of resource
To make preliminary identification of cases in epidemiological studies
• Selection of healthy individuals: employment, military…
• study on natural history of disease…
aims…
9. Strategies of screening
1. Population approach
– Apply methods to total population eg HIV testing
Advantage- Inclusive of everyone and Ethically appropriate
Disadvantage- Costs
2. Risk group approach:
Advantages
– It could be appropriate to individuals
– Could produce subject motivation
– Could have favorable benefits to cost ratio
Disadvantages
– It may be difficult in identifying high risk individuals
– There may be misclassification of individuals
– Unwanted labeling of individuals could bring anxiety
9
10. How to establish a screening program?
Mechanism: CRITERIA called Modified Wilson criteria
The Wilson criteria for screening emphasize the important features of
any screening program, as follows:
the condition should be an important health problem
the natural history of the condition should be understood
there should be a recognizable latent or early symptomatic stage
there should be a test that is easy to perform and interpret,
acceptable, accurate, reliable, sensitive and specific
there should be an accepted treatment recognized for the disease
treatment should be more effective if started early
there should be a policy on who should be treated
diagnosis and treatment should be cost-effective
case-finding should be a continuous process
11. Criteria for introducing a screening
program
CONDITION
/problem
PROGRAM
TREATMENT
/intervention
TEST/Dx
12. 1. The condition, i.e. health problem
Does the significance of the disease warrant its consideration as a
community problem??
Importance (prevalent, serious, predispose to other d/s)
– Prevalence of pre-clinical disease should be high
– Disease must be serious: congenital hypothyroidism, breast cancer
Is the natural history of the disease known?
– It should have a recognizable early or latent stage (asymptomatic)
– Natural history should be known
– Long period between 1st sign and overt disease
13. Is the screening test
– Simple (not sophisticated)
– Cheap
– Precise, validated
– Safe, Suitable & acceptable to population (impose minimal
discomfort on patient)
– Agreed diagnostic test(s)
– Sensitive and specific
– Reliable
– Facility should be available
– There should be confirmatory test which is accurate,
relatively cheap and easy to perform
2. The test or diagnosis
14. Should be available, acceptable & Effective
Of the disease at its earlier stages more effective than treatment
begun after the development of symptoms?
• If treatment makes no difference there is no need for screening
– Cervical cancer (PAP smear): good prognosis
– Lung cancer: bad prognosis
• Presence of evidence-based policies on whom to treat (i.e.
when the intervention is appropriate)
Treatment during the detectable preclinical phase must provide the
patient better prognosis
3. The treatment
15. Is there a mechanism for follow-up of individuals with positive
screening results to ensure thorough diagnostic testing?
Is there a schedules for the interval to repeat a screening?
Intervention
– Is it acceptable to public & professionals
– Are facilities & staff available (ensure continuity-ongoing)
– Benefits outweigh risks: cost-benefit and cost-effectiveness
– Is the overall costs (cost of detection & treatment) reasonable
Is there a counselling program for people who got Psychological/
Physical Harm?
4. The program, i.e. testing
16. Examples of screening Programs
• Cancers: breast, cervix, colon, prostate, lung
• Congenital : congenital displacement of hip
• Cardiovascular d/s: aortic aneurysm
• Which screening program might be appropriate to
introduce to Ethiopia?
18. Accuracy of a screening test
Measurement error
Discrepancy b/n the numbers we use to represent the thing
we are measuring & the actual value of the thing we are
measuring
Ex: your true wt is 60 Kg, but a scale this morning read 62 kg; the
discrepancy (i.e. 2kg) is due to the measuring instrument & is called
measurement error (measurement error of 2)
• The two determining property of measure are
1. Validity
2. Reliability
19. Validity
• The extent to which a test measures what it is supposed to
measure.
• (does it truly measure what it sets out to measure?:
• Sensitivity and Specificity
Ex: a devise for measuring sperm motility that actually
measures sperm count is not valid
20. Reliability
• Does it give consistent results? i.e. does it give the same
measurement each time?:
• Repeatability or precision
• Refers to the degree to which results obtained can be
replicated.
• Lack of reliability can be due to
• The measurement & measuring instrument
• Instability of the attribute (characteristic) being measured
(Intra-subject variation)
• The observers (Inter-observer)
20
22. Gold standard”
• Confirmatory test
• Is the best available information, which is the basis for
evaluation of the performance of a screening (second
diagnostic) test which is usually is cheaper, easier, or safer.
23. What can be calculated to check the
accuracy of a screening test against Gold
standard”
1. SENSITIVITY of the screening test (validity)
2. SPECIFICITY of the screening test (validity)
3. POSITIVE PREDICTIVE VALUE of the screening test
(Yield)
4. NEGATIVE PREDICTIVE VALUE of the screening test
24. Validity of a categorical test: four possible relationships
b/t a diagnostic test & actual presence of disease
Screening
Test result
Disease status or Diagnostic test
or Reference Test
(Gold standard)
Positive Negative
Positive True + (a) False + (b)
Negative False - (c) True - (d)
25. Sensitivity
Ability of a test to correctly identify those with the disease
Sensitivity =
a
a + c
Probability of a positive
test in people with the
disease
Disease +ve
a + c
Test Posit.
Negat.
a
c
Screening
Pos
Neg
Positive Negative
Confirmatory test
Gold standard
a + b
c + d
a + c b + d
a b
c d
26. Specificity
Ability of a test to correctly identify those without the disease
Specificity =
d
b + d
Probability of a negative
test in people without the
disease
Disease -ve
b + d
Test Posit.
Negat.
b
d
Screening
Pos
Neg
Positive Negative
Confirmatory test
Gold standard
a + b
c + d
a + c b + d
a b
c d
28. Specificity and sensitivity
• Increased sensitivity decreases the specificity
–Leads to more false positives
– people wrongly diagnosed as having the disease
– unnecessary panic among people & Health workers
–Unnecessary exposure for treatment
–Unnecessary wastage of resource (human, time,
drug, etc)
–Unnecessary stress on the health system
–Wrong negative implication external relation
28
29. Specificity and sensitivity
• Increased specificity decreases the sensitivity
–Leads to more false negatives
– people wrongly diagnosed as not having the disease
– delay in diagnosis
–Negative implication on prognosis of disease
–Unnecessary wastage of resource (human, time,
drug, etc) for sever disease
–Unnecessary stress on the health system
29
30. Specificity and sensitivity
There is a need to weigh consequences of leaving cases
undetected (false negatives) against erroneously classifying
healthy persons as diseased (false positives)
30
31. Issues..
• Sensitivity should be increased at the expense of specificity
when:
– Consequence of missing a case is a problem
– Disease is serious and definitive treatment exists
– Disease easily spreads (eg. syphilis, HIV)
– Subsequent diagnosis and treatment have minimal costs and
risk (eg further series of blood pressure reading for BP)
32
32. Cont…
• Specificity should be increased over sensitivity
when
–having the disease results in stigmatization/
discrimination
34
33. POSITIVE predictive value (Yield)
= is the probability of having a disease in a person with a positive
(abnormal) test result or when the test reads positive.
= proportion of screening test positives who are truly positive
= +PV= a/(a+b)
Test result
DISEASE STATUS (Gold standard)
PRESENT ABSENT
+ TP (a) FP (b)
_ FN (c) TN (d)
34. Positive Predictive Value (PVP)
37
a+b
Ill Not ill
Test
positive
PVP = a / (a+b)
a b
Positive Negative
Screening
Pos
Neg
Confirmatory test
Gold standard
a + b
c + d
a + c b + d
a b
c d
35. NEGATIVE predictive value
The probability that an individual free of disease will test negative by
the test
= is the probability of not having the disease in a person with a
negative result or when the result is negative (normal)
= proportion of screening test negatives who are truly negative
= -PV= d/(c+d)
Test result
DISEASE STATUS (Gold standard)
PRESENT ABSENT
+ TP (a) FP (b)
_ FN (c) TN (d)
36. Negative Predictive Value (PVN)
39
c+d
Ill Not ill
Test
Negative
PVN = d / (c+d)
c d
Positive Negative
Screening
Pos
Neg
Confirmatory test
Gold standard
a + b
c + d
a + c b + d
a b
c d
37. Relationship between PPV, NPV
Sensitivity & Specificity
• Lower PPV: more likely to be the result of poor
specificity than poor sensitivity
• Lower NPV: more likely to be the result of poor
sensitivity than poor specificity
40
38. Issues….
• The more specific the test the greater the PPV
• PPV can also be increased if prevalence of preclinical disease is
high in screened population.
– Can accomplish this by targeting high risk group
• The more sensitive the test the greater the NPV
41
39. • The usefulness of a test depends not only on its sensitivity and
specificity, but also on the prevalence of the condition tested for
in the population.
• As prevalence of disease increases…..the predictive value of a
positive test also increases
• As prevalence of disease decreases….the predictive value of a
negative test increases
Issues….
40. Example 1
Comparison of routine
diagnosis of malaria and
double read research slide
Reference test:
100 fields read by 2
technicians
Positive Negative
Screening test:
routine smear
at 10 hospitals
Positive 1555 988
Negative 514 1394
42. Strategies to Increase Yield
• Select a test with high specificity
–High sensitivity >> Low false negative - >> High NPV
–High specificity >> Low false positive >> High PPV
• Select disease with high prevalence of preclinical stage
• Target high risk group for screening
43. Which test to choose
• A sensitive test is preferable when
– there is an important penalty for failing to detect a disease (e.g., when
trying to detect a dangerous but treatable condition).
– the probability of disease is relatively low and the purpose of the test is to
discover possible cases.
A sensitive test is, therefore, most helpful when the test result is negative.
• A specific test is preferable when
– the test result is positive, and are often used to confirm a diagnosis, which
has been suggested by other data.
A highly specific is preferable when false positive results might have negative
(physical, emotional, or financial) consequences.
44. Sensitivity
Sensitivity = True positives / Affected persons
Persons testing positive
(True positives)
Persons testing negative
(False negatives)
Affected persons
(Positive by gold standard)
The sensitivity of a test in the ability of the test to identify
correctly affected individuals
Proportion of persons testing positive among affected individuals
45. What factors influence
the sensitivity of a test?
Characteristics of the affected persons?
YES: Antigenic characteristics of the pathogen in the area
(e.g., if the test was not prepared with antigens reflecting the
population of pathogens in the area, it will not pick up infected
persons in the area)
Characteristics of the non-affected persons?
NO: The sensitivity is estimated on a population of affected
persons
Prevalence of the disease?
NO: The sensitivity is estimated on a population of affected
persons
Sensitivity is an INTRINSIC characteristic of the test
46. Specificity
Specificity = True negatives / Non-affected persons
Persons testing negative
(True negatives)
Persons testing positive
(False positives)
Non-affected persons
(Negative by gold standard)
The specificity of a test in the ability of the test to identify
correctly non-affected individuals
Proportion of person testing negative among non affected individuals
47. What factors influence
the specificity of a test?
Characteristics of the affected persons?
NO: The specificity is estimated on a population of non affected
Characteristics of the non-affected persons?
YES: The diversity of antibodies to various other antigens in the
population may affect cross reactivity (e.g., If malaria is endemic,
polyclonal hyper gammaglobulinemia may increase the proportion
of false positives)
Prevalence of the disease?
NO: The specificity is estimated on a population of non affected
Specificity is an INTRINSIC characteristic of the test
48. Predictive value of a positive test
Persons affected
(True positives)
Persons not affected
(False positives)
Persons testing positive
(Positive by test)
Predictive value of a positive test =
True positives / Persons testing positive
The predictive value of a positive test is the probability that
an individual testing positive is truly affected
Proportion of affected persons among those testing
positive
49. What factors influence
the predictive value positive of a test?
• Sensitivity?
YES: To some extend.
• Specificity?
YES: The more the test is specific, the more it will be negative for
non affected persons. Thus, when the test is positive, it is probably
truly positive (All non affected were correctly identified as testing
negative).
• Prevalence of the disease?
YES: Low prevalence: The test will pick up more false positives
YES: High prevalence: The test will pick up more true positives
50. Predictive value of a negative test
Persons non affected
(True negatives)
Persons affected
(False negatives)
Persons testing negative
(Negative by test)
Predictive value of a negative test =
True negatives / Persons testing negative
The predictive value of a negative test is the probability that
an individual testing negative is truly non-affected
Proportion of non-affected persons among those testing
negative
51. What factors influence
the predictive value negative of a
test?
• Sensitivity?
YES: The more the test is sensitive, the more it captures affected
persons. Thus, when the test is negative, it is probably truly
negative (all affected were captured among the positive).
• Specificity?
YES: But to a lesser extend.
• Prevalence of the disease?
YES: Low prevalence: The test will pick up more true negatives
YES: High prevalence: The test will pick up more false negatives
52. To whom sensitivity, specificity and
predictive values matters most?
Sensitivity and specificity matter to laboratory specialists
– Studied on panels of positives and negatives
– Look into the intrinsic characteristics of the test:
• Capacity to pick affected
• Capacity to pick non affected
Predictive values matter to clinicians and epidemiologists
– Studied on homogeneous populations
– Look into the performance of the test in real life:
• What to make of a positive test
• What to make of a negative test
53. Q: How do you evaluate a
given screening program?
54. Evaluation of screening programs
• Feasibility of screening program is determined by
– Acceptability of program by potential screens (quick and easily
administered without much discomfort)
– Cost effectiveness
– Yield of cases: number of cases detected by screening program
• Yield measured by predictive value positive (PVP) and negative
(PVN)
– Predictive value measures whether or not an individual actually has the
disease, given the results of the screening test
57
55. Evaluation of a screening program
What proportion of tests are positive?
What type of disease is being detected?
What is the predictive value of the test?
What is the cost per test?
Do screen-detected cases have a better short term prognosis?
The only proper evidence of effectiveness of a
screening program is a reduction of total age-specific
mortality or morbidity, ideally demonstrated by
randomized trial.
56. Increasing Sensitivity and Specificity
• Clinicians can improve the sensitivity or specificity of the
diagnostic approach by using multiple tests
• Using several tests in combination, i.e. in parallel, will
significantly increase the sensitivity (while decreasing the
specificity)
• Conversely, using multiple tests in sequence, i.e. in serial
significantly increases the specificity (while decreasing the
sensitivity)
57. Different Strategies to Improve Screening by
Combining 2 Independent Tests
• IF, You want to be sure of your diagnosis
– Prioritise a test with higher PPV and specificity
– Reduce false positive
• IF, You do not want to miss a potential case
– Prioritise a test with higher NPV and sensitivity
– Reduce false negative
58. Combination Testing
• Serial Testing
– A test is first applied to a group. All those with a
positive result are retested.
– E.g., Serological testing for syphilis
• Parallel Testing
– Two tests are applied together. All those with either
or both tests are considered to be positive.
61
59. Two Stage Screening
• Those who initially tested positive are called again for further
testing
– a more expensive, more invasive and less comfortable
test
• Designed to reduce the problem of false positives
• Where as net sensitivity will decline in a two stage screening,
net specificity will increase.
• Why? Because we reduced false positives.
60. Volunteer bias
Volunteers & non-volunteers are different in the basis of:
1) Volunteers tend to have better health and lower mortality rates
than general population
they are more likely to adhere to intervention (screening,
medication, etc)
2) Those who volunteer are the “worried well”
At higher risk of developing disease because of family history
or life style characteristics
May have an increased risk of mortality regardless of efficacy
of screening program
61. Lead time bias
Lead time is the interval between the diagnosis of a disease at
screening and when it would have been detected due to
development of symptoms
• It represents the amount of time by which the diagnosis has been
advanced as a result of screening
• Depends on how soon the screening is performed
• If not taken into account, screened groups may appear to survive
longer than unscreened simply because diagnosis was made
earlier in the course of disease (lead time bias)
63. Length bias
• Refers to the over representation among screen detected cases of those with
a long preclinical phase of disease and thus a more favorable prognosis
• Those with long preclinical phase are more readily detectable by screening
than more rapidly progressing cases with a short preclinical phase.
• Thus length bias could lead to erroneous conclusion that screening was
beneficial
• Yet observed difference is as a result of detection of less rapidly fatal cases
through screening, while those more rapidly fatal are diagnosed after the
development of symptoms
64. o o
o
o
o
o
x
x
x
x
x
X onset of disease process
O time of death
screening
Rapidly
Progressive
Disease
Slowly
Progressive
Disease
68
x
