Validity and reliability expressesions means as to how measurements and diagnostic approaches can more efficiently and maintaning the accuracy with many repeated tests. In validity we basically speak of specificity and sensitivity of tests, which can be affected by prevalence.
2. Introduction
To understand how a disease is transmitted
and develops
To provide appropriate and effective health
care
The quality of screening and diagnostic test is
a critical issue
Validity and reliability
3. Validity of a test
The validity of a test is ability to do what is
supposed to do
distinguish between who has a disease and who
does not
The validity has 2 components sensitivity
and specificity
Sensitivity is the probability of testing positive if
the disease is truly present
Specificity is the probability of testing negative if
the disease is truly absent
4. Quality of a test
Test Results
Disease Status
Diseased Non diseased
Positive True positive
(a)
False positive
(b)
Negative False negative
(c)
True negative
(d)
5. Sensitivity
The formula for sensitivity is
The limits of a test’s sensitivity are presented by false
negative results
The false negative error rate is
c
a
a
y
sensitivit
c
a
c
FNrate
6. Specificity
It is the ability of a test to indicate non
diseased when disease is absent
The formula is
The limits of a test’s specificity are false positive
results
False positive error rate is given by
d
b
d
specificy
d
b
b
ate
F
Pr
7. Example One
Suppose we have a population of 1000
people of whom 100 have a certain
disease, 80 were correctly identified as
positive by the test. Of the 900 who
didn’t have the disease, the 800 were
correctly identified as negative. How
good was the test?
8. Prevalence
The proportion of people who have a
given disease or condition of interest at
a specified point in time.
Sensitivity and specificity are
independent of disease prevalence
Sensitivity includes only subjects with the
disease
Specificity includes only subjects without
the disease
9. Why are FP and FN important
False positives
False positive results pose burden on
the health care system (retest)
Worry is induced in persons
False negative
If the disease is a serious one for which
effective intervention is available in its
early stage (cancer) a FN result could be
a virtual death sentence
10. Predictive value of a test
If we screen the disease free population what
is the proportion of people who have the
disease who will be correctly identified
If the result is +ve, what is the probability that
s/he has the disease being tested (ie positive
predictive value)
If the result is –ve, what is the probability that
the patient doesn’t have the disease (ie negative
predictive value)
11. Relationship between predictive
value & disease prevalence
The higher the prevalence in the
screened population tend to increase
value of positive predictive value
Screening program is most productive
and efficient if its directed to a high risk
target population
12. Question One
A physical examination was used to screen for
breast cancer in 2500 women with biopsy-
proven adeno-carcinoma of the breast and in
5000 age and race matched control women.
The results of the physical examination were
positive in 1800 cases and in 800 control
women, all of whom showed no evidence of
cancer at biopsy.
Determine the validity of the physical examination and
predictive values.
13. THIS IS A 2ND OPINION, AT
1ST I THOUGHT U HAD
SOMETHING ELSE
RELIABILITY OF TESTS
14. Introduction
Aspect of assessing diagnostic and screening
tests
It asks whether a test is reliable or repeatable
Regardless of the sensitivity & specificity
if the test results cant be reproduced, the value and
usefulness of the test are minimal
Intra-subject variation and inter-observer
variation do cause the variation between test
results
15. Intra-subject variation
The values obtained in measuring many
human characteristics often vary over
time, even during a short time
Blood pressure, weight? Height?
In evaluating any test result, its
important to consider conditions under
which the test was performed including
the time of the day
16. Intra-observer variation
Sometimes variation occurs between
two observations made by the same
observer
Tests & examinations differ in the
degree to which subjective factors
enter into the observer’s conclusions
The greater the subjective element in
the reading, the greater the intra-
observer variation in readings is likely
to be
17. Inter-observer variation
This is variation between observers
Two examiners often don’t derive the same
results
The extent to which observers agree or disagree
is an important issue
We need to be able to express the extent of
agreement in quantitative terms
Overall % agreement is used to measure extent
of variation
18. Overall percent agreement
In 2by2 table
High chance for observers to agree on “d”
Tends to inflate the agreement percent
Disregard the subjects who were labelled
negative by both observers (ie cell d)
Therefore the overall % agreement will be
100
% x
c
b
a
a
agreement
19. KAPPA statistic
Reading the results can agree by chance
alone
To what extent the readings agree
beyond what we would expect by chance
alone
Kappa statistics will provide the answer
Numerator: % agreement observed minus
% agreement expected by chance alone
Denominator: 100%-%agreement by
chance
20. Historical classification of non
small cell carcinoma by two
pathologists
Grade II Grade III Total by B
Grade II 41 3 44
Grade III 4 27 31
Total by A 45 30 75
Pathologist
B
Pathologist A
21. Interpretation of kappa
statistics
Above 0.75 represents Excellent agreement
beyond chance
Kappa of 0.40 to 0.75 represents
intermediate to good agreement
Below 0.40 represents poor agreement
22. Relationship btn validity &
reliability
Curve for test results that are reliable
Curve for test results that are valid
Curve for both validity and reliability