Screening, Testing And Clinical Decision Making Flashcards
Primary prevention
Prevention of disease in those who do not yet have it
-immunization, diet, exercise, sun protection, public policies, purifying water supply, health fairs, counseling
Secondary prevention
Identification of those who have the disease but not yet developed signs and symptoms
- shorten its life span, or if no curs, increase quality of life
- cholesterol, prostate, breast exams, SCREENING PROGRAMS
Tertiary prevention
Prevention of complications in those with the signs and symptoms
- reduce disability
- rehabilitations, PT, respiratory therapy
Screening programs are considered _____ prevention
Secondary
Characteristics of a good screening test
Easy to perform, quick, inexpensive, safe
The probability of cases of a condition accurately identified by the screening test
Yield
-E.g. to detect 1 case of glaucoma, 100 must be screened
Ability of a test to distinguish between those who have the disease and those who dont
Internal validity (accuracy)
Generalizability
External validity
Repeatability
Reliability
-OHTS study-86% of first time abnormal fields were normal on second tests
The abiltiy of a test to correctly identify people with a disease
Sensitivity
Sensitivity proportions
# of people with the disease who testpositive/# of people with the disease who are tested -positive/all diseased who are tested
TP/TP+FN
Ability of a test to correctly identify people without a disease
Specificity
Specificity proportions
TN/TN+FP
Number of people without the disease who test negative/number of people without the disease who are tested
4 possible outcomes for specificity and sensitivity
True positive
False positive
True negative
False negative
Screening for glaucoma at a health fair
- screen 100 persons with icare tonometer and FDT matrix VF
- criteria for failure-IOP >21 and VF defect
- all 100 will later have complete eye exams at the clinic determine whether or not they truly have glaucoma-“gold standard”
What is the sensitivity and specificity: test glaucoma-20 true glaucoma, 10 true normal; test normal-5 true glaucoma, 65 true normal.
Sensitivity
20/25=80%
Specificity
65/75=87%
False negative
5/(5+20)=1-0.8 (sensitivity)
20%
False positive
10/(10+65)=1-0.87 (specificity)
13%
Tests with both high sensitivity and specificity
It is difficult to find a test with both-there is often a trade off between the two
A test with poor sensitivity
Many false negatives
-many people with the disease will pass the test
A test with poor specificity
Many false positives
-many people without the disease will fail the test
Emphasize sensitivity to minimize
False negatives
- when there is a big penalty for missing the diagnosis
- dangerous but tredaqtable conditions-HIV, syphilis, TB, brain tumors
- when you are more suspicious of the disease (e.g. higher prevalence)
Emphasize specificity to minimize ______
False positive
- when treatment involves risk and costs. Before subjecting patients to chemo, tissue diagnosis (highly specific test) is required.
- when you are less suspicious of the disease. E.g. lower prevalence
Choosing between sensitivity and specificity: confrontation visual fields
- higher specificity (if you’re normal, tour more likely to pass this test than a Humphrey visual field)
- low sensitivity (patients with subtle field defects will not be identified because targets are easily seen)
- good for general population
Choosing between sensitivity and specificity: automated threshold visual fields
- lower specificity-many normals will give abnormal results t first, due to learning curve
- much higher sensitivity
- use when there is a higher suspicion for glaucoma
- do not use indiscriminately because of the time and costs of repeating “abnormal” fields on normal eyes
How do we decide whether a test result is normal or abnormal
Cutoff points are chosen on a continuum between normal and abnormal
-want the test to be sensitive enough to diagnose disease but specific enough not to subject patients to unnecessary treatment
How would sensitivity and specificity change if I used a cutoff point other than 21 for IOP in a glaucoma screening?
Abnormal=18
- increase sensitivity
- decrease specificity
Abnormal=30
- decrease sensitivity
- increase specificity
PSA levels in black men to detect prostate cancer: sensitivity and specificity
The higher the PSA level, the more specific it is, the less sensitive it is
Receiver operating curve (ROC)
- plots sensitivity vs specificity
- shows trade off between sensitivity and specificity across a range of cutoff values
- S and S range from 0-100%
- values that give the max combinations of S and S are located closest to the upper left corner
ROC curves for multiple tests
- better tests have curves that crowd toward upper left corner
- larger area below the curve=more accurate tests
Testing for glaucoma using different tests: ROC
On the ROC, OCT had the line closest to the upper left corner and more area under the curve
-GDX was considered the worst one
Limitation of sensitivity and specificity
- they are properties of the test, BUT DO NOT TAKE INTO ACCOUNT THE PREVALENCE of the disease
- you perform a test for a very rare condition, even with a specific test, abnormal results are apt to be false positives
- you perform a test for a common condition, even with a sensitive test, normal results are apt to be false negatives
Predictive value
Takes into account sensitivity, specificity and prevalence of the disease in the population tested
-give a positive or negative result, what is the probability of having or not having the disease
Positive predictive value
The probability of the patient HAVING the disease given a positive (abnormal) test result
-like S/S, also a proportion
TP/TP+FP
Negative predictive value
The probability of the patient NOT HAVING the disease given a negative (normal) result
TN/TN+FN
What is the PPV and NPV: test glaucoma-20 true glaucoma, 10 true normal; test normal-5 true glaucoma, 65 true normal.
PPV=20/30=67%
- there is a 67% probability that those who test positive will actually have glaucoma
NPV=65/70=92%
-there is a 92% probability that those who test negative actually not have glaucoma
THE TEST HAS MORE VALUE IN IDENTIFYING NORMALS THAN IDENTIFYING GLAUCOMA PATIENTS WHEN THE PREVALENCE IS LOW
Prevalence=25/100=25%
The PPV of a test increases as the prevalence _____
increases
The NPV of a test increases as the prevalence ____
Decreases
Clinical decision making with snesitivety/specificity/PPV/NPV
- you cannot perform or emphasize every test on every patient
- you must consider the probability of a disease in your clinical setting before ordering a test for it
- consider the predictive value of a test before performing or emphasizing it
The prevalence of narrow angles (and angle closure glaucoma) is greater in older, female, Chinese, hyperopic, cataract patients. You measure and estimate the angles by Van Hericl as grade 1 (abnormal) in a myope and hyperopia, in a Caucasian and Chinese patients. Narrow angles are confirmed on gonio. For which patients is the test better at predicting of tru narrow angles?
The patient with more risk factors. If you see a grade 1 angle in a myope and in a hyperope, that grade 1 angle is going to be more a true narrow angle in the hyperope. Better PPV in patient with more risk factors for narrow angles.
DIFFERENT TESTS ARE MORE VALUABLE IN DIFFERENT POPULATIONS
The prevalence of retinal tears is greater in myopes. You find a suspected retinal tear during a peripheral retinal exam in a myope and emmetropia. You are more likely to be correct if the patient is a ____
Myope
- the PPV of a peripheral retinal exam is greater in myopes
- binocular indirect ophthalmoscope should be emphasized more in myopes
- FACTORS THAT INFLUENCE TEST SELECTION INCLUDE RISL FACTORS, SUSPICION, PREVALENCE, AND SEVERITY OF DISEASE
