Diagnostic Tests Flashcards
when is a Diagnostic test used
A diagnostic test may be used:
by a clinician, together with a clinical examination, to diagnose or exclude a particular disorder in a patient;
as a screening test to ascertain which individuals in an apparently healthy population are likely to have (or sometimes, not have) the disease of interest.
Individuals flagged in this way will then usually be subjected to more rigorous investigations in order to have their diagnosis confirmed.
what is the equation for sensitivity
a/(a+c)
remember 2x2 table
what is sensitivity
Sensitivity = Given the patient has the disease, what is the proportion of times the test is positive
are the number of false negatives (e.g. test is negative when they have disease).
So sensitivity is also 1 minus proportion of false negatives.
Sensitivity = the probability that someone with disease has a positive test result.
what is the equation for specificity
d/(b+d)
what is specificity
Given a subject is free of disease, specificity is the proportion of times the test will be negative
false positives, i.e. test is positive when person does not have disease.
So Specificity is 1 minus proportion of false positives
Specificity = the probability that a person without disease has a negative test result
Do sensitivity and specificity give us the probability of the test giving the correct diagnosis, whether it is positive or negative
Sensitivity & specificity do not give us this
what is the PPV
Positive predictive value (PPV) is the probability that someone has the disease when the test is positive, OR the proportion of individuals with disease when the test is positive.
Depends on prevalence
Equation for PPV
a/(a+b)
what is NPV
Negative predictive value (NPV) is the probability that someone is without disease when the test is negative, OR the proportion of individuals without disease when the test is negative
depends ON prevalence
equation for NPV
D/(C+D)
is PPV the proportion of patients with positive test results who are correctly diagnosed
yes
is NPV the proportion of patients with negative test results who are correctly diagnosed
yes
what test gives a direct assessment of the usefulness of the test in practice.
PPV and NPV
Do sensitivity and specificity depend on prevalence of abnormality
NO the PPV and NPV do
what is the disadvantage of sensitivity & specificity
The disadvantage of sensitivity & specificity is that they do not assess the accuracy of the test in a clinically useful way
what tests depend on prevalence of abnormality
PPV and NPV
Are both sensitivity and specificity conditional on having or not having the disease
Both sensitivity and specificity are conditional on having or not having the disease. They are properties of the test, not the disease.
what are properties of the test, not the disease
sensitivity and specificity
equation for accuracy
(a+d)/N
Equation for prevalence
(a+c)/N
FUNCTION OF A SCREENING
Screening – on (apparently) healthy people. Need high sensitivity to rule out those without disease
Function of diagnostic
Diagnostic – on people with a high suspicion of disease . Need high specificity to include those with disease
How is the Receiver Operating Characteristic (ROC) curve obtained
The Receiver Operating Characteristic (ROC) curve is obtained by plotting sensitivity (true positive rate) vs 1 – specificity (false positive rate) for every distinct cut-off value.
what are some assumptions about the ROC curve
Plot sensitivity vs 1-specificity for different cut-offs.
An ROC curve that is no better than chance will lie along 45o line.
The best cut-point is the one nearest the upper left-hand corner.
When comparing two curves the one with the largest area under the ROC is considered more accurate.
what is a Likelihood ratio
For any test result we can compare the probability of getting that result, if the patient truly had the condition of interest, with the corresponding probability if they were healthy.
The ratio of these probabilities is called the likelihood ratio (LR).
equation for LR+
Sensitivity/ (1-specificity)
equation for LR-
(1-sensitivity)/ specificity
what is Bayes Theorem
A very simple but important result in statistical theory.
This property enables you to calculate (conditional) probabilities when it is not possible to observe the proportions directly.
Bayes theorem equation
(sensitivity x prevalence) / probability of a positive result
what is the use of a large LR+
A large LR(+), e.g. > 10, suggests test may be useful in ruling in a diagnosis
what does a small LR- suggest
A small LR(-), close to zero (e.g. <0.01) suggests the test may be useful in ruling out a diagnosis
is the PNZ is close to 1-prevalence is the test useful
no
are sensitivity and specificity independent of the disease prevalence
yes
