renter Flashcards
what are the 4 diagnostic strategies?
- pattern recognition
- method of algorithm
- method of exhaustion
- hypothetico-deductive method
what is a test?
any device or process used to detect or elicit a sign, substance, tissue, change or response
diagnostic test
application for a test to clinically ill individuals in order to assess disease status
screening test
application of a test to apparently healthy individuals to detect infection, exposure or subclinical disease
define accurate
both valid and reliable
define reliability
will give the same results if administered more than once to same patient
define valid
correctly identifies the true status of the animal (identifies what you want to identify)
false positive test
tests positive for disease but the animal is healthy
false negative test
tests negative for disease but the animal has the disease
define sensitivity
-ability of the test to detect disease
-diseased animals test positive
A/(A+C)
define specificity
-ability of the test to detect non diseased
-healthy animals that test negative
D/(B+D)
when the sensitivity is high, a negative test:
rules out the disease
when the specificity is high, a positive test:
confirms the disease
we would want to max out our ability to rule out disease by maximizing:
sensitivity
we would want to maximize our ability to diagnose disease by maximizing:
specificity
what is true prevalence?
-proportion or percentage of the total population that are truly diseased (N)
Apparent prevalence- (1-sp)/
1-[ (1-sp) + (1-sn) ]
apparent prevalence
proportion or percentage of the total population that test POSITIVE
positive predictive value
proportion of test-positives that are infected/diseased
negative predictive value
proportion of test-negatives that are free of disease
expected prevalence
pre-test probability
-(probability the patient has disease prior to next test)
predictive value
post-test probabililty
define cut point
it defines positive or negative results, used for continuous values
True/False: concerning specificity and sensitivity, you often have to sacrifice one to get the other.
true
low cut points make a lot more people test:
positive
high cut points make a lot more people test:
negative
What is the ROC curve?
graphical approach to choosing a cut point
What happens with a TG-ROC curve?
the cut point is where two lines cross
define reference range
the set values 95% of the normal population falls within
-(that is, 95% prediction interval)
what does sequential testing improve?
post-test probability
-followup testing only on just the positive or negative animals
what is surveillance?
continuous and systematic process of collection, analysis and interpretation of data and dissemination of info on disease occurence in a population to initiate control measures or further investigation
what is monitoring?
routine observation and analysis of measurements
if threshold is lowered, you increase what?
sensitivity
what is passive surveillance?
collects voluntary data
what active surveillance?
data solicited or actively collected
fewer false negatives =
higher sensitivity
fewer false positives =
higher specificity
lower cut-point =
no false negatives (100% sensitivity, 100% NPV)
higher cut-point =
no false positives (100% specificity, 100% PPV)
in an ROC curve, a dot where on the curve optimizes both Sn and Sp?
top left corner
on a TG-ROC curve, a dot where on the graph optimizes Sn and Sp?
where the lines intersect
____ increases as prevalence or pre-test probability of dz increases
PPV
what are the Sn and Sp of the answers to your questions that you ask the owner?
90% Sn and 60% Sp
summary of tests in series:
- all tests positive = positive
- increases Sp and PPV
- decreases Sn and NPV
summary of tests in parallel:
- one or more tests positive = positive
- increases Sn and NPV
- decreases Sp and PPV
