2. 1. Definition
2. Objectives of screening
3. Uses of screening
4. Explanation of related terms
5. Screening vs diagnostic tests
6. Types of screening
7. Criteria of screening
8. Performance of a new diagnostic test
9. Application of Bayes’ theorem in screening
10. Choosing a screening test
11. Evaluation of a screening program
12. Study designs used to evaluate screening programs
Contents
3. Definition:
“The search for unrecognised disease or defect
by means of rapidly applied tests or other
procedures in apparently healthy individuals”.
“Specific examination of members of population
to identify any existing illness at a pre-
symptomatic stage or identify susceptibility to
a particular disease”
4. Objectives of screening:
To reduce the morbidity and mortality of the
screened disease in those persons screened;
To protect public health;
To carry out research.
5. Uses of screening:
1. Case detection (prescriptive screening):
It is the presumptive identification of
unrecognised disease, which does not arise
from a patients requests.
eg.- Neonatal screening
eg.- screening for breast cancer, cervical cancer,
deafness in children, DM, IDA, PKU etc.
6. 2. Control of disease (prospective screening):
It is screening of people for the benefit of
others.
eg.- screening of immigrants from infectious
diseases such as TB and syphilis to protect
home population.
eg.- screening for streptococcal infection to
prevent rheumatic fever.
7. 3. Research purposes:
Many chronic diseases whose natural history is not
fully known (cancer, HT etc) screening may aid in
obtaining more basic knowledge about the natural
history of disease.
Initial screening provides a prevalence estimate and
subsequent screening, an incidence figure.
8. 4. Educational opportunities:
Screening provides opportunities for creating
public awareness and for educating health
professionals.
eg.- Screening for DM.
9. Explanation of related terms:
SCREENING: Testing for infection or disease in populations
or in individuals who are not seeking health care;
eg.-Neonatal screening for CDH, CHD, Cerebral palsy etc.
CASE FINDING: Use of clinical and /or lab tests to detect
disease in individuals seeking health care for other reasons;
eg.- Use of VDRL test to detect syphilis in pregnant women
DIAGNOSTIC TESTS: Use of clinical and/or lab procedures
to confirm or refute the existence of disease or true
abnormality in patients with signs and symptoms presumed to
be caused by the disease;
eg.- VDRL testing of patients with lesions s/o sec syphilis.
10. Screening and diagnostic tests
contrasted
SCREENING TESTS
D/o apparently healthy
Applied to groups
Test results are arbitrary and final
Less accurate
Less expensive
Not a basis for Rx
The initiative comes from the
investigator or agency providing
care
Based on one criterion or cut off
point (eg. D M)
DIAGNOSTIC TESTS
D/o those with indications / sick
Applied to single pts
Diagnosis is not final but
modified in light of new evidence.
Diagnosis is sum of all evidence.
More accurate
More expensive
Used as a basis of Rx
The initiative comes from a pt
with complaints
Based on evaluation of a no. of
symptoms, signs and lab findings.
11. Types of screening:
1. MASS SCREENING
2. HIGH RISK SCREENING
3. MULTIPHASIC SCREENING:(MULTIPLE
TESTING) :-
a) Parallel Testing
b) Serial Testing
c) Multistage Screening
12. 1. MASS SCREENING: i.e.. Screening of a
whole population or a subgroup, as for eg..
all adults.
It is offered to all irrespective of the
particular risk an individual may run of
contracting the disease in question.
Disadvantage: Indiscriminate mass screening
is not a useful preventive measure unless it is
backed up by suitable Rx that will reduce the
duration of illness or alter its final outcome.
13. 2. HIGH RISK SCREENING: i.e.. Selective application of
screening to high risk groups.
eg.- Ca Cx – occurs less commonly in high socio -economic
groups, so screening for Ca Cx in lower SE groups could
increase the yield of new cases.
eg.- Screening of family members (close contacts) in cases of HT,
DM, breast cancer etc.
Lately screening for disease ---------screening for risk factors
as these apparently antedate the development of actual
disease.
eg.- Elevated serum cholesterol is associated with a high risk of --
-----coronary heart disease.
Advantage:
Thus preventive measures can be applied before the disease
occurs.
Also economical use of resources will occur if screening
tests are applied selectively to individuals in high risk
groups.
14. 3. MULTIPHASIC SCREENING:(MULTIPLE TESTING)
Because most diagnostic tests are less than perfect, a single
test is frequently insufficient for making an unequivocal
diagnosis.
We may require to employ 2 or more screening tests in order
to improve validity (sensitivity and specificity) as a single test
may be inadequate in making preliminary diagnosis.
Therefore clinicians often use multiple diagnostic tests
administered either in parallel or serially.
15. a) PARALLEL TESTING:
In parallel testing, a battery of tests is given concurrently.
eg. 1- ECG (T1) and Echocardiography(T2) used as diagnosis of
VSD
When done in parallel manner,
T1+ T2+
T1+ T2- any of these will be considered positive results
T1- T2+
A 2x2 table can then be constructed to find sensitivity and
specificity.
eg. 2 - For diagnosis of Glaucoma-
Tonometry, ophthalmoscopy and perimetry can be
administered together.
eg. 3 - For diagnosis of DVT- Impedence plethysmography and
leg scanning after injecting I fibrinogen are performed
together.
16. Uses of Parallel Testing:-
Parallel tests should be ordered when rapid
assessment is needed ( eg. in emergencies) or
for routine physical examinations.
Result is gain in sensitivity as compared to
individual testing as the disease is less likely to
be missed, but less specific as FP are more and
PVN is increased (FNR is decreased).
17. b) Serial Testing
In this a battery of tests is performed sequentially
until an unequivocal diagnosis can be made.
eg. 1- For glaucoma screening initially all persons to be
screened are subjected to tonometery and then to
perimetry.
eg. 2- For TB screening – initially miniature
radiography – large X-ray films – CT scan followed
by sputum examination.
All tests must have positive results (T1+, T2+…) for
the sequence to be considered positive for the disease.
18. Uses Of Serial Testing
Serial tests should be used when rapid
assessment is not necessary.
Serial testing results in increased specificity,
lower sensitivity ( higher FNR) & increased
PVP ( lower FPR).
19. c. Multistage Testing
eg.1- Screening for diabetes mellitus –
First stage – FBS
Second stage – OGTT is performed only in positive
patients.
eg.2- Mass screening of syphilis
First stage – RPR/VDRL (more sensitive and less
specific, so large number of FP).
Second stage – Positive persons are exposed to FTA –
ABS test (more specific & sensitive + costly). This
test helps to weed out the FP cases.
20. Criteria For Screening
These are based on 2 considerations –
1. The disease –
a) The condition sought should be an important
health problem.
b) Recognisable latent or early asymptomatic stage.
c) Natural history of the condition should be
adequately understood.
d) Accepted & effective treatment.
e) Facilities for diagnosis & treatment available.
f) Suitable test or examination that can detect the
disease prior to the onset of S/S.
21. g) The test or examination acceptable to public &
professionals.
h) Agreed policy on whom to consider as patients
including management of borderline cases.
i) Good evidence that early detection & Rx reduces
morbidity & mortality.
j) Cost of early diagnosis & treatment
economically balanced in relation to total
expenditure on medical care.
When the above criteria are satisfied then only it is
appropriate to consider a suitable screening test.
22. 2. Assessment of screening test-
a) Acceptability:
• Test should be acceptable to the people at
whom it is aimed.
• Painful, discomforting or embarassing tests
(rectal/vaginal exam) are less likely to be
accepted
23. b) Repeatability:
• Also called Reliability, precision, Reproducibility.
• Test must give consistent results when repeated more than
once on the same individual or material, under the same
conditions.
• It depends on 3 factors-
Observer variations-
Intra-observer variations ( within ) eg. BP
readings.
Inter-observer variations (between) eg. PBS may show
MP+ by one tech and MP- by another tech.
Intra observer variations can be minimized by taking average of
several replicate observations at the same time.
Inter observer variations can be minimized by –standardization of
procedures, intensive training of all observers and making
use of 2 or more observers for making independent
observations.
24. Biological variations- Like
Changes in parameters observed.
Eg.- Cx smears taken from the same woman may be abnormal
one day and normal the other day.
Eg.- Myocardial infarction may present with or without pain.
Variations in the way pts perceive their symptoms and answer.
Regression to the mean.- ie. Tendency for values at extremes of a
distribution ,either very high or low to regress towards the
mean on repeat measurements.
Eg.- Rheumatoid arthritis pain
Stool frequency in ulcerative colitis
BP in hypertension.
Errors related to technical methods-
Eg.- defective instruments
Erroneous calibration
Faulty reagents.
25. c) Validity of a test-
• Refers to what extent the test measures what
it purports to measure.
Eg. - Glycosuria is a useful screening test for
DM but a more valid/accurate test is glucose
tolerance test.
• Validity has 2 components:
1. Sensitivity
2. Specificity
26. Performance Of A New Diagnostic
Test
Physicians are often faced with the task of evaluating the merit of
a new diagnostic test
A. The Gold Standard –
Assessment of a new DT begins with the identification of a group
of patients known to have the disorder of interest, using an
accepted reference test called the gold standard.
Limitations
1. The GS is often most risky, technically difficult, expensive
or impractical of available diagnostic options. Eg.- PM brain
biopsy ----- a GS for diagnosis of Alzheimer's disease
2. For some conditions, eg.- angina pectoris no GS is available.
3. Comparison with imperfect GS may lead to erroneous
conclusions (inc FP or inc FN).
27. B. Some Definitions
1. Index Test : Test whose accuracy is to be evaluated.
2. Study Population : Population on whom results are
applicable,
3. Gold Standard
4. True Positive (TP) : These are individuals with the disease
who are correctly identified as diseased by the new test.
5. False Positives (FP) : These are disease - free individuals
falsely labeled as diseased by the new test.
6. True Negatives (TN) : Disease – free individuals who are
correctly identified as disease – free by the new test.
7. False Negatives (FN) : Individuals with disease falsely
labeled as disease – free by the new test.
28. C. Test Performance Characteristics
Screening
Test
Results
Diagnosis Total
Diseased
(D+)
Non Diseased
(D-)
Test +
(T+)
a (TP) b ( FP) a + b
Test –
(T -)
c (FN) d (TN) c + d
Total a + c b + d a+b+c+d
29. 1. Sensitivity :
It is the probability that a diseased individual will have
a positive test result.
It is the true positive rate (TPR) of the test.
Sensitivity = TPR
= diseased with positive test
all diseased
= P (T +/ D +)
= a / a + c x 100%
Eg. – A 90 % sensitivity means that 90% of the diseased
people screened by the test will give a TP result and
remaining 10% will be diagnosed as non diseased (FN).
30. 2. Specificity :
It is the probability that a disease – free individual will have
a negative test result.
Specificity is the true negative rate (TNR).
Specificity = disease – free with negative test
all disease – free
= P (T- / D- )
= d / b + d x 100%
Eg. – A 90 % specificity means that 90% of non diseased
people screened by the test will give a TN result and
remaining 10% of non diseased will be wrongly
diagnosed as diseased (FP) when they are not.
31. 3. False Negative Rate (FNR) :
It is the probability that a diseased individual will have a negative
test result.
FNR = diseased with negative test
all diseased
= P (T- / D+)
False positive Rate (FPR) :
It is the probability that a disease – free individual will have a
positive test result.
FPR = disease - free with positive test
all disease - free
= P (T+ / D-)
32. 5. Prevalence :
It is the proportion of individuals in a population
who have the disease.
Also called as prior probability or pre test
probability
Prevalence = No. with disease
Total No. of individuals in study
= P (D+)
33. 6. Predictive Value Positive (PVP / PV+) :
It is the probability that an individual with a
positive test result has the disease. Also called
as posterior probability or post test probability
of disease.
PVP = Diseased with positive test
all with positive test
= P (D+ / T+)
= a / a + b x 100%
34. 7. Predictive Value Negative (PVN / PV-) :
It is the probability that an individual with a
negative test result does not have the disease.
PVN = Disease – free with negative test
all with negative test
= P (D- / T-)
= d / c + d x 100%
35. Bayes’ rule is a mathematical formula that may be
used as an alternative method for obtaining
unknown conditional probabilities such as PVP or
PVN from known conditional probabilities such as
sensitivity & specificity.
Formula : P ( A/B ) = P (B/A) P (A)
P (B/A) P (A) + P(B/Ā)P(Ā)
Application Of Bayes’ Theorem In
Screening
36. Using Bayes’ rule :
PVP = P(D+/T+)
= P(T+/D+)P(D+)
P(T+/D+)P(D+)+P(T+/D-)P(D-)
(using the definition of conditional probability in
terms of joint probability)
This formula can be rewritten as –
PVP = (sensitivity) (prevalence)
(sensitivity) (prevalence) + (1- specificity) (1-prevalence)
37. Similarly :
PVN = P(D-/T-)
= P(T-/D-)P(D-)
P(T-/D-)P(D-) + P(T-/D+)P(D+)
= (specificity) (1-prevalence)
(specificity) (1-prevalence) + (1-sensitivity)(prevalence)
Therefore :
The more sensitive a test, the better its PVN.
The more specific a test the better its PVP.
38. Choosing A Screening Test
1. Situations requiring a more sensitive test –
i.e. ↑TPR and ↓FNR
Lethal diseases like cancer (crucial early detection),
neonatal PKU screening where a FP result is far
less deleterious than a FN result. While most FP are
identified during follow up testing, a FN result may
postpone essential dietary intervention until mental
retardation is evident.
High specificity tests are expensive.
39. 2. Situations requiring high specificity tests :
i.e. ↑TNR and ↓FPR
When a FP diagnosis may lead to a risky treatment
(eg. Chemotherapy, open heart surgery ) or follow
up procedure (eg. Cardiac catheterisation ).
When a FP diagnosis may have deleterious effects
on a patients lifestyle, self image, financial situation
( eg. AIDS, mental illness, learning disorders)
40. Evaluation Of The Screening
Program
Ideal screening program –
Cheap
Safe
Identifies a large number of people in whom
prognosis is bad and treatment is effective.
Eg. – LBW babies (<2000 gm)----- account for half of
all neonatal deaths----- information that neonatal
mortality can be decreased by special care & facilities
------ can help in identifying high risk groups
41. Factors that need to be considered –
a) Cost of the test :
Needs to be measured as it affects patients in
terms of time lost, discomfort, anxiety as
well as the resources of health services.
Eg. – for Ca breast screening it was found that
identification of one new case required the
examination by palpation and mammography
of 170 women and taking of 9 biopsies.
42. b. Yield:
The summarising figure for yield is the no. of tests
that must be done in order to identify one case
whose prognosis will be improved as a result of
early detection.
Many conditions forwhich screening might be
considered have too low a prevalence to justify
routine exam of the entire population.
43. Eg -Among inborn errors ----PKU has a prevalence of
about 1 in 10,000 and this is considered to justify
routine screening.
Eg. -Likewise Wilsons disease also calls for early
recognition and Rx but unfortunately with a
prevalence of 1 in 100000 the yield may be
unacceptably low.
Sometimes it may be possible to improve yield by
restricting screening to identifiable risk groups.
Eg. -Incidence of Chorionic Ca following normal
pregnancy is 1 in 50,000 (too low yield to justify
screening).
Following H- mole the incidence is about 25%,so that
screening is justified despite the cost of the test.
44. c. Policy decisions:
Eg.-From a Swedish trial of Ca breast screening the
following balance sheet was derived-
GAINS LOSSES
Gain in life
(2.5+ years)
Costs to medical services and
women of screening 1850 women
1800
reassured
75 FP reports, 50 correctly
identified, but without benefits.
Ref- Epidemiology in medical practice- Barker, Cooper ,Rose.
Thus these results show that policy decisions require
judgments.
45. Sources of bias in evaluation of a
screening program-
Patient self selection bias (volunteer bias)
Lead time bias
Length bias
46. Volunteer bias:
• Better health, lower mortality, and more likely to
adhere to prescribed regimes.
• On the other hand ,volunteers for a screening prog
represent ‘worried well’ i.e. asymptomatic individuals
who have increased risk due to medical /family
history /lifestyle characteristics---------such people
have increased risk of mortality regardless of efficacy
of screening prog.
47. Lead time bias:
Lead time is defined as 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,
49. Age in Years
Biological onset
of disease
Disease
Detectable
By screening
Woman A
Diagnosed
At Screen
Symptoms
developed
Woman B
diagnosed
Woman A and B
Both die from
Breast cancer
35 41 43 46
40
Apparent increase in duration of survival due to lead time in
hypothetical screened & symptom diagnosed cases of breast
cancer.
From the time of diagnosis women A survived a total of 5 years / 2 years longer then
woman B. however though an exam of survival following diagnosis would appear to
indicate that screening was beneficial, infact it only pushed forward the time of diagnosis
; both woman died at exactly the same age.
50. Since screening is applied to asymptomatic
individuals, by definition every case detected by
screening will have its diagnosis advanced by some
amount of time.
Cases that progress rapidly from preclinical phase ---
clinical phase will gain less lead time and vice versa.
If an estimate of lead time is not taken into account,
when comparing mortality among screened and
unscreened groups, survival may erroneously appear
to be increased among screen detected cases simply
b’cos diagnosis was made earlier in the course of
disease.
51. Length Bias :
Refers to over representation among screen
detected cases of those with a long pre-clinical
phase of disease and thus a more favourable
prognosis.
Thus those with a long pre-clinical phase are
more readily detectable by screening than
those with shorter pre-clinical phases that
progress more rapidly to symptoms, regardless
of any effect of screening prog itself.
52. Study designs used to evaluate
screening programs:
Correlational studies (CS):
These have been used to examine trends in disease rates in
relation to screening frequencies within a population.
Eg.- CS have shown inverse relationship between frequencies of
use of pap smear and rate of Ca Cx mortality.
Though some findings were difficult to interpret-
• Decline in Ca Cx mortality occurred in developed countries
before pap smear was widely used.
• Effect of potential c.f such as SE status
Thus it is difficult to know if women who experienced decline in
Ca Cx mortality rates were actually those who received
screening.
53. Limitations of CS-
• Information from these studies concerns
populations rather than individuals. It is not
possible to establish that those experiencing
the decreased morbidity and mortality are
infact the same persons who received
screening.
• Does not allow for control of potential c.f.
54. Observational analytic studies:
Case control studies-
• Individuals with and without disease are compared
with respect to their past exposure (screening).
• Eg.-H/o pap smear was examined in newly diagnosed
cases of invasive Ca cx and controls.
Cohort studies-
Case fatality rate of those who chose to be screened is
compared with the comparable rate among those
whose diagnosis was symptom related.
• Most commonly used.
• Limitation- self selection bias, lead time and length
bias must be taken into account.
55. Randomised control trials:
Optimal assessment of the efficacy of a
screening program.
• Large sample size-----confounding controlled
by randomisation.
• Randomisation also takes care of self
selection/volunteer bias.
• Lead time bias------By comparing age specific
mortality rates for screened and unscreened
groups, OR By adjusting for average lead
time when comparing survival among screened
and symptom detected groups.
56. Eg.- HIP breast Ca screening project
conducted in USA was a randomised trial
designed to evaluate whether periodic breast
Ca screening with mammography and physical
exam would result in reduced Ca breast
mortality among women aged 40 - 64 yrs.
After 9 yrs of follow up , there was statistically
significant reduction in ca breast mortality
among screened women as c.t unscreened
women.
Limitations- cost, ethics, feasability.
57. Conclusion:
The primary objective of screening is to reduce
morbidity and mortality from disease through early
detection and Rx.
Given a disease with characteristics appropriate for
screening and a valid screening test, both the
feasibility and effectiveness of screening must be
evaluated.
The same epidemiological principles and methods that
apply to primary prevention of disease apply to the
evaluation of screening program as a means of
disease control.