Diagnostics tests Flashcards
Diagnosis
CLASSIFY animals as having disease or not
2 types of Diagnostic tests results
Dichotomous = 2 possible answer (- or +)
Continuous = continuum of possible answers
Hematology Diagnostic test measure
amount of different cell types in venous blood sample
CBC, Differential blood count, PCV
Biochemical Diagnostic test measure
enzymes, metabolites, proteins in venous blood
liver enzymes, BUN, CREA, amylase, lipase
Immunological Diagnostic test measure
antibodies (ELISA, IFA, Western blot, SNAP FIV/ E.cani)
use antibodies to detect antigen
SNAP FeLV / heartworm
FIV/FeLV snap tests detect
anti-FIV antibodies or FeLV antigens
Pathogen detection test
Detect pathogens itself (not antibodies against it)
present during infection with no time Lag
PCR detecting pathogen DNA/ RNA in clinical sample
How is decision of test made?
Cut off value is used which was determined experimentally as the test value that minimizes False +/ -
How to determine if test value is + / -
test a lot of individuals
proven to have disease or condition
proven to not have disease or condition
Define cut-off that best separates the two tested groups
False Positives
Non-disease animals that test positive (Sp issue)
False Negatives
Diseased animals that test negative (Se issue)
2 parameters of evaluating Diagnostic tests/
Sensitivity (Se) and Specificity (Sp)
Sensitivity (Se)
proportion of diseased animals that correctly test positive
Specificity (Sp)
Proportion of non-diseased animals that correctly test negative
Low specificity causes
contamination
mistake pathogen ID
animals with antibodies from vaccine (not infection)
SnOUT
highly sensitive test rules disease OUT
(NO false negatives)
negative result from highly sensitive test is most likely
from non-diseased animal
(RULES OUT DZ)
false positive are possible if Sp is low
T/F: Sensitivity tells you about how well the test performs on Non-diseases animals
False:
Doesn’t tell you a fucking thing about non-diseased
Another way of saying Specificity
how often the test detects the things specifically intended to detect - Not healthy animals
SpIN
highly specific test rules Disease IN
most positive results are from diseased
75% Sp
correctly classifies 75% non diseased animals as negative
25% non diseased as positive (false positives)
100% Sp
correctly classifies all non-diseased animals as negative
(if Se low than false negatives possible)
doesn’t tell you how well test performs on diseased animals
What do you never want as test results?
False negatives
How to test imported animals for infections?
test with High Se and only allow negative results to enter country
maximize sensitivity when?
you need to detect ALL disease or infected animals
maximize Specificity when?
testing for very serious diseases with serious consequences
High Sp
correctly classify ALL non-diseased animals as negative
To accuratly detect all diseased and eliminate non diseased animals?
1st : high Se
2nd: high Sp
T/F: Se and Sp doesn’t tell you the probability of + is truly disease
PPV and NPV do this
True
PPV
Positive predictive Value
answers
my patient tested positive and probability it has disease
only include true positives and false positives
NPV
Negative predictive value
answers
my patient tested negative and probability it doesn’t have disease
true negatives and false negatives
Se and Sp won’t change if prevalence of disease in population changes
True
PPV and NPV will change as prevalence of disease in population changes
True
