Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
39 Cards in this Set
- Front
- Back
What is the difference between serum chemistry results and infectious disease test results?
|
serum chemistry results are on a continuous scale, while infectious disease test results are either positive or negative based on cutoff points
|
|
What is considered the normal range for a biochemical profile?
|
mean +/- two standard deviations
|
|
What values are considered normal for a biochem profile if the data isn't normally distributed?
|
2.5-97.5 percentile
|
|
What is a titer?
|
highest dilution that produces a test reaction
|
|
What scale of dilution is usually used for titers?
|
2 fold
(can see 10 fold, or 2 then 10 fold) |
|
define seropositive
|
animal that has a detectable titer at or above the cutoff
|
|
define seronegative
|
animal that does not have a detectable titer at the cutoff
|
|
What is a cutoff value?
|
threshold for what is considered a positive or negative test
|
|
What do circulating antibodies indicate?
|
infection
exposure recovered from a previous infection maternal antibodies in neonates |
|
What do circulating antigens indicate?
|
current infection only
|
|
define seroprevalence
|
proportion of seropositive animals in a population
|
|
What does seroprevalence depend on?
|
rate of seroconversion
half-life of the antibodies length of time after seroconversion *doesn't necessarily mean there is a high rate of infection |
|
What change in titer is considered significant?
|
four fold change, may indicate infection (but not always)
|
|
define analytical sensitivity
|
lowest level of substance that can be detected by an assay
|
|
define analytical specificity
|
how well the test detects only one substance (lack of cross reactivity)
|
|
define prevalence
|
probability that a randomly selected animal has the disease
|
|
define apparent prevalence
|
probability that a randomly selected animal has a positive test result
|
|
define diagnostic sensitivity
|
probability that an animal has a positive test result given it has the disease
|
|
define diagnostic specificity
|
probability that an animal has a negative test result, given it doesn't have the disease
|
|
define positive predictive value
|
probability that the animal tests positive when it has the disease
|
|
define negative predictive value
|
probability that an animal that tests negative will not have the disease
|
|
Why are predictive values more useful clinically than sensitivity and specificity?
|
they use the test result to make an inference about the true disease status which is more useful in the clinic
|
|
What do predictive values depend on?
|
prevalence
|
|
What is a gold standard?
|
perfectly predicts which animals do and do not have disease
*in real life, the test with the highest sensitivity and specificity |
|
What is the relationship between true prevalence and apparent prevalence?
|
as true prevalence increases, so does apparent prevalence
|
|
How do cutoff values affect sensitivity and specificity?
|
if the cutoff value is increased, specificity increases and sensitivity decreases (and vice versa)
|
|
What effect does prevalence have on PPV and NPV?
|
increased prevalence increases the PPV and decreases the NPV (and vice versa)
|
|
What assumptions are generally made about sensitivity and specificity?
|
constant, do not change based on the population in which they are used (this may not be true!)
|
|
What is a receiver-operating characteristic curve?
|
graphical representation of all possible outcomes of Se and Sp for a test (based on the possible cutoff values)
|
|
How are ROC curves interpreted?
|
generally the point closest to the northwest corner of the graph is the Se and Sp with the greatest accuracy for a given test
|
|
When should you NOT select the most accurate point on an ROC curve as the cutoff for a test?
|
when the sensitivity or specificity needs to be very high (ie: the risk of having a positive animal test negative is unacceptable or vice versa)
|
|
How can ROC curves be used to compare different tests?
|
plot them all on the same graph, see which one has the greatest area under the curve (closest to the NW corner). That one is the most accurate.
|
|
What is a kappa statistic?
|
measure of agreement
|
|
how is kappa interpreted?
|
less than 0 = poor agreement
1 = perfect agreement |
|
When is kappa used?
|
to compare two or more observers or tests
|
|
What assumptions are made when kappa is used?
|
observers/tests are independent conditional on disease status
observers/tests are blinded to the other's results |
|
What is conditional independence?
|
second test sensitivity is not dependent on the results of the first
|
|
What does it mean to have serial test interpretation?
|
one test applied first, then the positive animals are tested with another test
*animal must test positive on both to be considered positive *increases specificity, decreases sensitivity |
|
What is parallel testing?
|
multiple tests are applied at the same time, animal is considered positive if either test is positive
*increases sensitivity, decreases specificity |