Sensitivity, Specificity
what does the variable "b" represent
# of people who do not have the disease and the test is positive
what does the variable "c" represent
# of people who have the disease and the test is negative
what does the variable "a" represent
# of people who have the disease and the test is positive
positive LR (likelihood ratio) definition and calculation
ratio of the true positive results to false positive results Pos LR = sensitivity / (1-specificity)
negative LR definition and calculation
ratio of true negative results to false negative results neg LR = (1-sensitivity) / specificity
what an LR of less than 1.0 means
represents a decrease in the likelihood of the disease
what an LR greater than 1.0 means
represents an increase in the likelihood of the disease
purpose of likelihood ratios
to determine the likelihood that a positive test result is a true positive and a negative result is a true negative
false pos calculation
B / (B+D)
what does the variable "d" represent
# of people who do not have the disease and the test is negative
characteristics of a highly specific test
- good at identifying patients without a disease - low percentage of false positives
characteristics of a highly sensitive test
- good at identifying the patient with the disease - low percentage of false negatives
characteristics of a low specificity test
- limited in identifying patients without a disease - high percentage of false positives
characteristics of a low sensitivity test
- limited in identifying the patient with a disease - high percentage of false negatives
ROC curve values
.5-.6: fail (F) .6-.7: poor (D) .7-.8: fair (C) .8-.9: good (B) .9-1: excellent (A)
false negative calculation
C / (C+A)
when are likelihood ratios calculated?
after pretest probability (disease prevalence) and diagnostic testing (sensitivity and specificity), but before posttest probability (PPV, NPV)
ROC curve: C statistic
area under the curve. probability that the test result from a randomly selected person with the disease will be positive
ROC curve: what the blue line and black line represent
blue: actual data black: reference line that represents a 50/50 chance of accurately predicting the disease
receiver operating characteristics curves (ROC curves)
descriptive graph that plots the true positive rate against the false positive rate
how is the accuracy of a screening test determined?
evaluated in terms of its ability to correctly assess the presence or absence of a disease or condition as compared to the gold standard
how to calculate specificity
going down columns, not across (D) / (B+D) (true neg) / (true neg + false pos)
how to calculate sensitivity
going down the columns. (A) / (A+C) --> (true positive) / (true pos + false neg)
false positive
indicates a disease is present when it is not
false negative
indicates that a disease is not present when it is
how is accuracy measured in ROC curve
measured by the area under the curve (labeled as "Area" in the SPSS output)
relationship between specificity and sensitivity
more sensitive = less specific
gold standard
most accurate means of currently diagnosing a particular disease and serves as a basis for comparison with newly developed diagnostic or screening tests
what an LR less than 0.1 means
patient does not have the disease
what negative results of a sensitive test mean
patient is less likely to have the disease
what does it mean if a specific test has positive results
patient is more likely to have the disease
false negative rate
probability of having disease but having negative result
sensitivity
probability of having the disease; true positive
false positive rate
probability of no disease but having a false positive test
specificity
probability of the absence of the disease; true negative
test variable
screening variable
true positive
sensitivity - accurately identifies the presence of a disease
true negative
specificity - accurately indicates that the disease is not present
state variable
the disease state or the gold standard
ROC curve: what the distance between the lines means
the greater the distance the blue line is from the black line, the more accurate the test
what is the importance of determining the sensitivity of a screening test?
the higher the sensitivity of a screening test, the more likely the test is to be positive when a person has the disease (true positive)
what is the importance of determining the specificity of a screening test?
the higher the specificity the more likely the test is to be negative when a person does not have a disease (true negative)
what does the false positive rate mean?
the number of people who do not have the disease and the test is positive for the disease
what an LR value greater than 10 means
the patient has the disease
four possible outcomes of a screening test for a disease
true positive, false positive, true negative, false negative
ROC curve: x-axis and y-axis
x: false positive rate (1-specificity) y: true positive rate (1-specificity)